journal of economics, management & agricultural development

ARTICLES IN THIS ISSUE

Price Cointegration and Leadership in Regional Tilapia and

Roundscad Markets in the Philippines, 1990-2007

Socio-Economic and Environmental Assessment of a

Microcontroller-Based Coffee Roasting Machine: Implications

for Market Potential and Technology Commercialization

Farmers’ Willingness to Pay for an Alternative Irrigation Policy

to Reduce Greenhouse Gas Emissions from Rice Farming in

the Upper Pampanga River Integrated Irrigation System

Building Disaster Resilience to Address Household Food

Security: The Case of Sta. Rosa-Silang Subwatershed

Operations and Profitability of Snail Dredging in Laguna,

Rizal and Pasig City

Effects of Extension Services on the Technical Efficiency of

Rice Farmers in Albay, 2014-2015

Decoy Effect and Student Preference with regard to

USB Flash Drives

Economic Impacts of Smallholder Oil Palm

(Elaeis guineensis Jacq.) Plantations on Peatlands in

Indonesia

Lyndon A. Peña and

Bates M. Bathan

Ma. Eden S. Piadozo,

Roberto F. Rañola Jr.,

Ma. Joy N. Malabayabas

and Dominic M. Hamada

Ruel M. Mojica and

Marilyn M. Elauria

Fezoil Luz C. Decena and

Isabelita M. Pabuayon


Michael A. Cuesta,

Bam Razafindrabe and

Ryohei Kada

Mark Angelo R. Alcaide and

Jefferson A. Arapoc

Volume 1 Number 2

December 2015 ISSN: 2449 - 4585

Journal of Economics, Management

& Agricultural Development

in economics, management and agricultural development

University of the Philippines Los Baños

College of Economics and Management

http://cem.uplb.edu.ph/index.php

Leadership

Relevance

Excellence

Yolanda T. Garcia,

Maria Esperanza T. Gracia

and Flordeliza A. Lantican

Muhammad Akmal Agustira,


Asa Jose U. Sajise and

Leonardo M.Florece

Editor

Isabelita M. Pabuayon, Department of Agricultural and Applied Economics,


Associate Editors

Nanette A. Aquino, Department of Agribusiness Management and Entrepreneurship,


Amelia L. Bello, Department of Economics, University of the Philippines Los Baños

Ma. Eden S. Piadozo, Department of Agricultural and Applied Economics,


Dinah Pura T. Depositario, Department of Agribusiness Management and Entrepreneurship,


Marilyn M. Elauria, Department of Agricultural and Applied Economics,


Agnes T. Banzon, Department of Agribusiness Management and Entrepreneurship,


Zenaida M. Sumalde, Department of Economics, University of the Philippines Los Baños

Rodger M. Valientes, Department of Economics, University of the Philippines Los Baños

Editorial Advisory Board

Narciso R. Deomampo, Former Food and Agriculture Organization/ Regional Office for Asia and the

Pacific Senior Officer

Gregmar Galinato, Associate Professor, Washington State University

Flordeliza A. Lantican, Retired Professor, College of Economics and Management,


Rodolfo M. Nayga, Jr., Professor, Food Policy Economics and Agribusiness, University of Arkansas

Nguyen Thi Duong Nga, Chair, Department of Quantitative Analysis, Vietnam National University of

Agriculture

V. Bruce J. Tolentino, Deputy Director General (Communication and Partnerships), International Rice

Research Institute

Jimmy B. Williams, Adjunct Professor, Department of Agribusiness Management and

Entrepreneurship, University of the Philippines Los Baños

Managing Editor

GIideon P. Carnaje, Department of Economics, University of the Philippines Los Baños

Cover and Layout

Marion M. Bueno

Journal of Economics, Management & Agricultural Development

College of Economics and Management University of the Philippines Los Baños

College, Laguna 4031 Philippines

Articles in this Issue

Price Cointegration and Leadership in Regional Tilapia and Roundscad

Markets in the Philippines, 1990-2007

Yolanda T. Garcia, Maria Esperanza T. Garcia and Flordeliza A.

Lantican


Microcontroller-Based Coffee Roasting Machine: Implications for

Market Potential and Technology Commercialization

Ruel M. Mojica and Marilyn M. Elauria

Farmers’ Willingness to Pay for an Alternative Irrigation Policy to

Reduce Greenhouse Gas Emissions from Rice Farming in the Upper

Pampanga River Integrated Irrigation System

Fezoil Luz C. Decena and Isabelita M. Pabuayon

Building Disaster Resilience to Address Household Food Security:

The Case of Sta. Rosa-Silang Subwatershed

Roberto F. Rañola Jr., Michael A. Cuesta, Bam Razafindrabe and

Ryohei Kada

Operations and Profitability of Snail Dredging in Laguna,

Rizal and Pasig City

Ma. Eden S. Piadozo, Roberto F. Rañola Jr., Ma. Joy N. Malabayabas

and Dominic M. Hamada

Effects of Extension Services on the Technical Efficiency of Rice

Farmers in Albay, 2014-2015

Lyndon A. Peña and Bates M. Bathan

Decoy Effect and Student Preference with regard to USB Flash Drives

Mark Angelo R. Alcaide and Jefferson A. Arapoc

Economic Impacts of Smallholder Oil Palm (Elaeis guineensis Jacq.)

Plantations on Peatlands in Indonesia

Muhammad Akmal Agustira , Roberto F. Rañola Jr., Asa Jose U. Sajise

and Leonardo M. Florece

Journal of Economics, Management & Agricultural Development

1

21

35

53

69

83

Volume 1 Number 2 December 2015

95

105

Journal of Economics, Management & Agricultural Development Vol. 1, No. 2 1

1 College of Economics and Management, University of the Philippines Los Baños

Email: [email protected] 2 BS Agricultural Economics Graduate, 2007

College of Economics and Management, University of the Philippines Los Baños 3 Retired Professor, College of Economics and Management, University of the Philippines Los Baños

Abstract

This study aims to investigate the price dynamics between the wholesale and retail

prices of roundscad and tilapia in the regional markets of the Philippines. It also seeks to

analyze how the observed dynamics in these prices could be linked to the development in

market-related infrastructures that can promote trade, e.g., telecommunication and

transportation facilities. Results of the study reveal that there is a general decrease in the

number of regions that show cointegrated wholesale and retail prices over time in both fish

species. Moreover, price leadership under the ―Granger-causality‖ sense (of either

wholesale or retail price) seems to diminish in the recent years for both markets. These

results suggest that the development in telecommunication and transportation facilities that

generally facilitate the movements of fish from one supply point to another could increase

price competition in these markets leading to more uniform prices, thus reducing the

influence of a dominant price.

Keywords: price cointegration, price leadership, market infrastructure, tilapia,

roundscad, Philippines

Introduction

Roundscad and tilapia are two of the most popular fish species that are

commonly consumed by Filipino households. The share in total fish expenditure of households for roundscad and tilapia in 2000 is estimated to be nearly the same, i.e.,

15% and 13%, respectively (Garcia, Dey and Navarez 2005). Furthermore, national

prices for these two species on the same year are competitively close, registering PhP 55 per kg and PhP 58 per kg, respectively (PSA 2000). These similarities in shares of

consumption and prices have largely contributed to their being substitute foodfish for

the Filipino consumers. Notably, the substitutability of roundscad and tilapia is observed to be higher among low income households as compared with the high

income households with cross-price elasticity of 0.34 for the poorest segment of the

population and 0.21 for the more affluent group (Garcia, Dey and Navarez 2005).

For the longest time, roundscad is popularly known as the ―poor man‘s fish‖ due

to its relatively cheaper price. More recently, however, roundscad is rapidly losing that title to tilapia due to the decreasing real price of the latter over time. Tilapia,

unlike roundscad, is a fish that can be farmed under freshwater aquasystem, while

roundscad is a marine fish that can only be caught in the wild.

Roundscad and tilapia have served as major sources of animal protein especially

among the poorer households in the country due to their affordable prices. Thus, a study on how the prices of these two fish commodities have moved over time could

help producers (i.e., fish farmers for tilapia and coastal fishermen for roundscad) in

matching consumer demand with their production flows.

Price Cointegration and Leadership in Regional Tilapia and Roundscad

Markets in the Philippines, 1990-2007

Yolanda T. Garcia1, Maria Esperanza T. Garcia2 and Flordeliza A. Lantican3

2 Yolanda T. Garcia, Maria Esperanza T. Garcia and Flordeliza A. Lantican

Given the archipelagic nature of the Philippines, markets are geographically separated which makes it difficult and costly to move products from one market to

another. However, with the current developments in market infrastructures (such as telecommunication, transport services, construction of modern fish ports/landing sites

and more farm-to-market roads), dynamics in fish prices can be facilitated. Moreover,

these infrastructure developments have the potential to increase trade between spatially

separated markets, thus promoting price integration (Garcia and Salayo 2009b).

This study, therefore, highlights the importance of linking production points to

consumption centers through the shortest possible route to help minimize margins between wholesale and retail prices, thus lowering fish prices and making it more

affordable to consumers (Ling 2003, Petersen and Muldoon 2007, Salayo 1989).

Generally, prices serve as signals to producers, traders and consumers which

guide product flows within and across markets. Prices, therefore, play an important

role in integrating markets through efficient transmission of relevant information. However, as markets become disconnected and independent, trade flows, in turn, tend

to become inefficient and costly. Hence, factors that help increase the efficiency of trade flows across markets are also the same factors that enhance market integration.

Specifically, infrastructure development that aid market related functions - i.e.,

transport and communication - could play an important role in the speedy and more effective integration and transmission of wholesale to retail prices and vice versa

(Garcia and Salayo 2009a).

Considering that this study is one among the very few studies in the country that deal with the issue of price dynamics in fish (i.e., tilapia and roundscad) markets, it

can serve as a benchmark for future researches on the dynamics of prices for other fish species. Since this study considers both the monthly trend and regional differences in

tilapia and roundscad prices, other researchers can therefore use the present results as

basis for comparison of temporal and spatial dimensions in fish price integration and leadership. At the same time, the analytical techniques that are employed in this study

can be applied to the analysis of similar issues for other fish species, whether supply comes from aquaculture or marine fish catch. Incidentally, the same analytical

procedures could also be used in the analysis of price dynamics in agricultural crops or

manufactured goods, provided that price data series are available. Lastly, this study could help policymakers understand how development in market-related infrastruc-

tures can promote efficiency in price formation especially in areas where marketing

facilities are poor.

Objectives of the Study

This study aims to establish the long-term price relationship between monthly wholesale and retail prices in regional markets of roundscad and tilapia, covering the

period from 1990 to 2007. Moreover, the study aims to relate how developments in

market-related infrastructures that promote trade and marketing affect the integration and leadership of tilapia and roundscad prices at the national and regional levels.

Specifically, the objectives of the study are as follows:


a. to determine the existence of long-term relationship between monthly regional wholesale and retail prices of roundscad and tilapia in the Philippines, covering the

period 1990-2007;

b. to investigate the leader-follower relationship between the monthly wholesale

and retail prices under the ―Granger-causality‖ sense across the regional markets of

these two fish species;

c. to establish whether the same leader-follower relationship between the prices of

roundscad and tilapia exists at the wholesale and retail levels;

d. to relate the observed dynamics in regional wholesale and retail prices of tilapia and roundscad to the developments in road infrastructure, telecommunication and

transportation facilities; and

e. to recommend policy directions that could promote more efficient price

formation for the benefits of the consuming public.

This study, therefore, seeks to investigate whether wholesale and retail prices of roundscad and tilapia in their respective markets show long-term relationships. If they

do, it is worth asking further which between the two prices serves as the leader/follower under the ―Granger-causality‖ sense. The answers to these questions have

wide ranging implications for the fishery sector, which impinge on the economic

welfare of the producers, traders and consumers. It also has environmental implication in terms of efficiency in the use of fishery resources in the country‘s

coastal waters and the use of production inputs in the aquaculture sector. For

example, it is often claimed that the biggest component of production cost in aquaculture is the cost of feeds because of the practice of ad libitum feeding among

fish farmers. Often this practice leads to the pollution of fishponds and other water

bodies where fish pens and cages are constructed.

However, if price signals could be efficiently transmitted by consumers and

traders to fish farmers - through transmission from retail to wholesale price - prices could guide aquaculture producers regarding the proper timing of harvesting their fish

stocks. This in turn could lead to a reduction in the cost of feeds, which often

increases unnecessarily due to excessive feed use.

Methodology

This study makes use of cointegration analysis to establish the existence of long-term price relationship between the wholesale and retail prices of roundscad and

tilapia. At the same time, price leadership between these prices is investigated using

the Granger causality test. This test sought to answer the question of whether the wholesale price or the retail price is the leader (under the ―Granger-causality‖ sense)

in the tilapia and roundscad markets. Finally, the existence of price cointegration and leadership between the wholesale and retail prices of roundscad and tilapia for a given

region is related to the presence or absence of market-related infrastructure

developments - e.g., road density, transport and telecommunication facilities - that are crucial to the movement of fishery products from production points to consumption

centers.


Cointegration of Wholesale and Retail Prices

Application of regression analysis to several price series can lead to either a

spurious relationship or real long-term relationship depending on the behavior of the prices that are being analyzed. Spurious relationship occurs when two non-stationary

prices are regressed and yields a non-stationary error term. In such a case, the

standard t-test on the regression coefficients and the F-test on the price model are not valid (Gujarati 1995). However, there is a possibility that the regression results from

these two non-stationary prices can be meaningful when the residual term of the

regression model can be proven to be stationary. Under this situation, the two prices

are said to be cointegrated.

In this study, the wholesale and retail prices of tilapia and roundscad in their

respective regional markets are tested for cointegration to determine whether these

prices are temporally independent or cointegrated. To conduct the cointegration test,

the Dickey-Fuller (DF) test is used to establish the stationarity or randomness of the

error term µ1 of the cointegrating regressions shown below for species-specific whole-

sale and retail prices:

Wholesale Priceroundscad = α1 + β1 Retail Priceroundscad + µ1

Wholesale Pricetilapia = α2 + β2 Retail Pricetilapia + µ2

Note that the reverse specification of the cointegrating regressions will simply yield the same result, hence only one version of the test is used in the analysis. To

establish whether the above regressions yield spurious or meaningful results, it is important to test whether the error terms µ1 and µ2 of the two functions above are

stationary or not. This is done by applying the DF tests on both µ1 and µ2. The DF test

is implemented by running the following auxiliary regressions:

∆µ1,t = δ1 µ1,t-1 + Є1t

∆µ2,t = δ2 µ2,t-1 + Є2t

If δi is found to be statistically significant (i.e., accept Ha: δi≠0 at α=5% level of

significance), then the error term of the cointegrating regression is deemed random and stationary. Specifically, this means that the wholesale and retail prices of

roundscad (or tilapia) in a given regional market have a legitimate long-term relationship. Furthermore, this means that the two prices are cointegrated and they

move in a common trend. On the other hand, if the δi is found to be statistically non-

significant (i.e., accept Ho: δi=0), then the error term of the cointegrating regression is

non-stationary which renders the price regression to be spurious.

It is important to note that before the cointegration test can be applied, the respective prices of tilapia and roundscad must be tested for the occurrence of

integration of order one or I(1) in their respective series. This condition requires that

the level form of the prices should be non-stationary while their first-difference must be stationary. The same DF test can be used to establish the stationarity of the level

form of prices and their first differences, i.e.,


Non-stationarity in the prices at level form:

∆WPt = Ɵ1 WPt-1 + Є1t

∆RPt = Ɵ2 RPt-1 + Є2t

Stationarity of the first-difference of prices:

∆2WPt = Ɵ3 ∆WPt-1 + Є3t

∆2RPt = Ɵ4 ∆RPt-1 + Є4t

To satisfy the I(1) condition of the prices, Ɵ1 and Ɵ2 should be statistically non-significant in the DF test while Ɵ3 and Ɵ4 should be statistically significant. Note that

cointegration analysis can only be conducted if the prices in the cointegrating

regression models are both integrated of order 1 or I(1). In cases where this condition is not satisfied, the fitted regression for the price series can not imply co-movement in

their long-term relationship.

Granger Causality Tests for Tilapia and Roundscad Markets

Price leadership in this study is defined under the ―Granger-causality sense‖. It is

characterized as the precedence of a price series over another price series. For example, the wholesale price causes the movements of the retail price if the changes in the

wholesale price precede the changes in the retail price. Under the ―Granger causality‖ sense, the wholesale price is considered as the price leader while the retail price is the

price follower. Note that all references for price leadership in this study follows this

definition.

Price leadership is established by conducting the test of Granger causality. The

models for the causality tests of wholesale and retail prices for a given fish species are

specified as follows:

WPi,t = α1 WPi,t-1 + α2 WPi,t-2 +…+ αn WPi,t-j + β1 RPi,t-1 + β2 RPi,t-2 +…+ βn RPi,t-j + Є1it

RPi,t = δ1 RPi,t-1 + δ2 RPi,t-2 + …+ δn RPi,t-j + θ1 WPi,t-1 + θ2 WPi,t-2 +…+ θn WPi,t-j + Є2it

where: WPi,t is the wholesale price of species i at time t

WPi,t-j is the lagged wholesale price of species i up to t-j periods

RPi,t is the retail price of species i at time t

RPi,t-j is the lagged retail price of species i up to t-j periods

i pertains to individual species, i.e., tilapia and roundscad

j pertains to the number of parameter estimates α, β, δ and θ in the equations

There are four possible relationships that could exist between the wholesale and retail

prices of a given fish species in a market:

a) Unidirectional relationship between the wholesale and retail prices:

i) Retail price is Granger causing the wholesale price ( Σαj = 0 and Σβj ≠ 0);

ii) Wholesale price is Granger causing the retail price ( Σδj = 0 and Σθj ≠ 0);


b) Bilateral relationship between the wholesale and retail prices: ( Σβj ≠ 0 and

Σθj ≠ 0);

c) Independence between the wholesale and retail prices: ( Σβj = 0 and Σθj = 0).

Price leadership occurs under cases a.i and a.ii. Under case a.i, the retail price is

said to be the leader, while under case a.ii, the wholesale price is the leader. Also, it

is important to note that Granger causality test between the wholesale and retail prices can only be pursued if these prices are found to be cointegrated. Under cases of

spurious relationship between these market prices, Granger causality test cannot be

undertaken. Note that the same Granger test can be conducted to determine whether price leadership between roundscad and tilapia prices exists at the wholesale and

retail levels.

Results of the Granger causality tests can identify which particular price triggers

the direction of the price movement: a) whether the wholesale price leads or follows

the retail price of a given fish species or vice versa; and b) whether the roundscad price leads or follows the tilapia price at the wholesale or retail level and vice versa. It

is expected that the two fish markets will show differential price leaders since roundscad production is dependent on wild catch while tilapia production, being an

aquaculture species can be controlled by the fishfarmers. Hence, studying the

movements in their prices would help tilapia fish farmers and fishers in their production decisions especially during lean seasons of wild species like the

roundscad.

Relationship of Price Dynamics to Infrastructure Developments

Relationships of price dynamics to market-related infrastructure developments

are analyzed in the study by linking the price dynamics (i.e., price cointegration and leadership) with various developments in the following: a) road density; b)

telecommunication facilities and c) transport facilities. To track the changes in the

dynamics of the fish prices, the analyses for price cointegration and leadership are divided into two consecutive periods: a) Period 1 covering the months from 1990-

1999; and b) Period 2 which covered the months from 2000 to 2007.

The changes in the price dynamics are then associated with the changes in the

levels of market-related infrastructures over the two periods. The key to the analysis

is to trace the changes in the price dynamics from Period 1 to Period 2 as they relate

to the changes in the levels of infrastructure facilities that take place over the periods.

Selection of Study Area

The selection of regions that are included in the study is based on the availability of price series for a given region. Specifically, the price data were collected from the

Bureau of Agricultural Statistics (BAS) on a monthly basis for the period 1990-2007.

All the price series are expressed in real terms to remove the effect of inflation.

On the other hand, data on country side infrastructure developments such as road

density, telecommunication and transport facilities were gathered from the National Statistical Coordination Board (NSCB), Department of Public Works and Highway

(DPWH) and the Bureau of Fisheries and Aquatic Resources (BFAR). Specifically,

the length of national roads and bridges is used to represent the state of road infra-

structure in the regions.


In the case of access to telecommunication facilities, the percent of households with subscription to landline and mobile phones within a region is used as a proxy

variable. Finally, transport facilities are represented by the number of registered utility

vehicles and trucks that are used in marketing farm produce in the particular region.

Results and Discussion

Cointegration Analysis (Roundscad)

The results of the regional tests for cointegration of wholesale and retail prices

for roundscad are presented in Table 1. There are six regions that are tested for

cointegration in period 1 and only two regions in period 2. Note that prices in some of the regions are not tested for cointegration since either one of the prices exhibited

stationarity which violates the I(1) condition for cointegration of prices. All the

regions that are tested for price cointegration show cointegrated wholesale and retail

prices for both periods. However, there are more regions with cointegrated wholesale

and retail prices in the roundscad markets in period 1; (i.e., NCR, Regions 4B, 6, 9, 10 and 11) compared to Period 2; (i.e., only Regions 4-B and 6). However, when the

price data are pooled, i.e., covering the period 1990-2007, only Regions 4-B, 9 and 11 show cointegrated prices. Similarly, the national data for wholesale and retail prices

of roundscad also show cointegrated series. These results imply that the behavior of

the prices in the two periods varies. Only the price series for Region 4-B shows consistent co-movements in the three periods that are considered in the study.

Additionally, for some regions like Region 6, where prices show common trend in

both periods 1 and 2, the co-movement of their prices disappears when the price series are connected for longer price movements. Presumably, the common trend in

the two periods is different despite their co-movement such that a different pattern of

price behavior emerges when the price series are connected for the entire period.

Cointegration Analysis (Tilapia)

Table 2 presents the results of cointegration tests for the wholesale and retail prices of tilapia in the regional markets. Note that the tests are conducted only in the

Luzon regions since tilapia is not a popular fish species in the Visayan and Mindanao regions, hence prices in these areas are scarce. Results show that tilapia prices register

cointegrated behavior only at the national level and Region 3 in period 1. For period

2, prices at the national level and Region 4B are also found to be cointegrated. However, when the price series are connected for the entire period, none of the

regions including the national level prices show cointegrated behavior. These results

suggest that the long run trends in the wholesale and retail prices of tilapia in the

entire period are generally independent of each other in most of its regional markets,

especially at the national level.

On the whole, these results suggest that while there is no co-movement that

existed between the wholesale and retail tilapia prices over the span of 18 years, there

are some common trends in these prices when the time series are divided into shorter periods. This could be explained by the possibility of opposite price movements in the

two periods such that the common trend tend to cancel out when the two periods are

combined. Interestingly, there may be reasons to believe that there are factors at play that change the behavior of the price series when they are broken down into shorter

series.


Table 1. Tests for cointegration of wholesale and retail prices of roundscad by

region, Philippines, 1990-2007

Critical value of t at α =5% is 2.88

*- Significant at α = 5%

Table 2. Tests for cointegration of wholesale and retail prices of tilapia by

region, Philippines, 1990-2007

Period 1

1990-1999

Period 2

1990-1999

All

1990-2007 Region

ADF

statistic

Cointegrated? ADF

statistic

Cointegrated? ADF

statistic

Cointegrated?

Philippines -7.237* Yes - - -8.638* Yes

NCR -5.738* Yes - - -2.277 No

Region 1 - - - - -2.776 No

Region 2 - - - - - -

Region 3 - - - - -2.741 No

Region 4-A - - - - - -

Region 4-B -7.754* Yes -4.715* Yes -8.825* Yes

Region 5 - - - - -2.617 No

Region 6 -5.687* Yes -5.671* Yes -2.640 No

Region 7 - - - - -2.826 No

Region 8 - - - - -

Region 9 -6.470* Yes - - -8.744* Yes

Region 10 -7.362* Yes - - -2.092 No

Region 11 -7.041* Yes - - -7.930* Yes

Period 1 Period 2 All

Region ADF

statistic

Cointe-

grated?

ADF

statistic

Cointe-

grated?

ADF

statistic

Cointe-

grated?

Philippines -6.259* Yes -6.101* Yes -2.699 No

NCR - - - - -2.709 No

Region 1 -2.167 No - - -2.521 No

Region 2 -2.782 No -2.823 No -2.489 No

Region 3 -4.714* Yes - - -2.570 No

Region 4-A -2.542 No - - -2.768 No

Region 4-B -2.482 No

-5.621* Yes

-2.592 No


This could be associated with the nature of tilapia production in the country. Specifically, tilapia is often harvested partially, i.e., supply in the market is often

controlled not by production season but by the demand in the market. When wild fish species are scarce, tilapia fish farmers can decide to harvest the bulk for their fish

stock to increase fish supply. However, when wild fish is abundant in the market,

fish farmers can hold on to their fish stock to protect the price of tilapia (Garcia and Sumalde, 2011). In such a case, the price movements in the market do not follow a

natural trend. Instead tilapia prices fluctuate depending on the demand and supply

situations in the market of competing wild species like the roundscad.

Granger Causality Analysis (Roundscad Wholesale and Retail Prices)

Granger causality analysis seeks to establish the existence of price dependence

in the wholesale and retail markets. When prices are dependent with each other, their

relationship could be unilateral, which can be characterized by a price leader–follower

type or bilateral which indicates simultaneity in price relationship.

Table 3 presents the Granger causality tests between the wholesale and retail

prices of roundscad by regional markets. Note that Granger causality can only be implemented for cointegrated prices, hence the Granger tests are only conducted for

regions that show co-movement in prices (see Table 1). For the entire period, the

Granger tests show that the retail price is the leader in the roundscad markets of NCR and Regions 2, 3, 5, 7, 9 and 10. This means that the wholesale price follows the

movements of the retail price in these markets. This result suggests that the retail

price generally leads the price movements in the roundscad market.

At the national level, the reverse is observed where the wholesale price emerges

as the price leader. When the price series are divided into 2 periods, the wholesale price also appears to be the price leader for Region 6 in Period 1. However, the

reverse is observed in Region 10, where the retail price serves as the leader. On the

other hand, for Period 2, there is no price leadership that is established in the two regions that shows cointegrated wholesale and retail prices. These results suggest that

there is no pattern that could be established on which a particular price leads the

movements in roundscad prices for this period.

Granger Causality Analysis (Tilapia Wholesale and Retail Prices)

Table 4 presents the Granger causality tests for the wholesale and retail prices of tilapia by region. At the national level, the retail price emerges as the leader in period

1 but there is no price leader found in period 2 nor in the entire period. At the

regional level, the retail price also emerges as the price leader in Region 3 for period 1 and Region 4-B in period 2. On the whole, these results suggest that the retail price

generally leads the price movements in the tilapia markets.


Per

iod

1

Per

iod

2

All

Reg

ion

R

P=

f(la

g

WP

t-i)

WP

=f(

lag

RP

t-i)

Pri

ce

Lea

der

RP

=f(

lag

WP

t-i)

WP

=f(

lag

RP

t-i)

Pri

ce

Lea

der

RP

=f(

lag

WP

t-i)

WP

=f(

lag

RP

t-i)

Pri

ce

Lea

der

Phil

ipp

ines

2

0.9

2 *

7.5

6 n

s W

ho

lesa

le

a a

b

24

.53 *

1

4.1

4 n

s W

ho

lesa

le

NC

R

7

.03

ns

11

.16 n

s -

a a

b

a a

b

Reg

ion 4

-B

20

.97 *

2

7.9

0 *

-

14

.80 n

s 1

7.5

9 n

s -

23

.28 *

3

5.4

4 *

-

Reg

ion 6

1

8.4

8 *

1

3.6

3 n

s W

ho

lesa

le

13

.32 n

s 1

9.6

0 n

s -

a a

b

Reg

ion 9

9.0

1 n

s 1

6.8

7 n

s -

a a

b

19

.77 n

s 4

5.0

5 *

R

etai

l

Reg

ion 1

0

8

.95

ns

19

.13 *

R

etai

l a

a b

a

a b

Reg

ion 1

1

4

.30

ns

8

.66

ns

- a

a b

8.6

9 n

s

16

.17

ns

-

Tab

le 3

. P

rice

lea

der

ship

in

wh

ole

sale

an

d r

etail

pri

ces

of

rou

nd

scad

, b

y r

egio

n, P

hil

ipp

ines

, 1990 -

2007

*

-

Sig

nif

ican

t at

α =

5%

n

s -

S

tati

stic

ally

non

-sig

nif

ican

t

a

-

G

ran

ger

cau

sali

ty t

est

cann

ot

be

imp

lem

ente

d s

ince

th

e p

rice

s ar

e n

ot

coin

tegra

ted

b

-

Pri

ce l

ead

er c

ann

ot

be

esta

bli

shed

- n

on

e of

the

pri

ces

show

ed l

ead

ersh

ip


Tab

le 4

. P

rice

lea

der

ship

in

wh

ole

sale

an

d r

etail

pri

ces

of

tila

pia

, b

y r

egio

n, P

hil

ipp

ines

, 1990 -

2007

Per

iod

1

Per

iod

2

All

Reg

ion

R

P=

f(la

g

WP

t-i)

WP

=f(

lag

RP

t-i)

Pri

ce

Lea

der

RP

=f(

lag

WP

t-i)

WP

=f(

lag

RP

t-i)

Pri

ce

Lea

der

RP

=f(

lag

WP

t-i)

WP

=f(

lag

RP

t-i)

Pri

ce

Lea

der

Phil

ipp

ines

6

.75

ns

31

.28 *

R

etai

l 4

7.0

1 *

3

6.1

7 *

-

a a

b

NC

R

a a

b

a a

b

a a

b

Reg

ion 1

a

a b

a

a b

a

a b

Reg

ion 2

a

a b

a

a b

a

a b

Reg

ion 3

2.0

4 n

s 9

.54

*

Ret

ail

a a

b

a a

b

Reg

ion 4

-A

a a

b

a a

b

a a

b

Reg

ion 4

-B

a a

b

14

.59 n

s 1

7.7

2 *

R

etai

l a

a b

*

-

Sig

nif

ican

t at

α =

5%

n

s -

S

tati

stic

ally

non-s

ignif

ican

t

a

-

Gra

nger

cau

sali

ty t

est

can

not

be

imp

lem

ente

d s

ince

th

e p

rice

s ar

e not

coin

tegra

ted

b

-

P

rice

lea

der

can

not

be

esta

bli

shed

- n

on

e of

the

pri

ces

show

ed l

ead

ersh

ip


Granger Causality Analysis (Rounscad vs Tilapia Prices)

Table 5 and 6 present the Granger causality tests on the wholesale and retail

prices of roundscad and tilapia, respectively. These cross-causality tests are done to determine whether the prices (at the wholesale and retail levels) of these two fish

species are independent of each other or a price leader exists between them. Note that

the number of regions where the tests are carried out is limited to those regions where tilapia is consumed by the general public, i.e., only in the Luzon regional markets due

to the non-popularity of the said fish species in the Visayan and Mindanao regions.

In general, the wholesale price of roundscad generally leads the wholesale price of tilapia in both periods 1 and 2. Similarly, for the entire period, the tilapia wholesale

price generally followed the movements of the roundscad wholesale price.

At the retail level, a different picture emerges. In period 1, the tilapia retail price

generally leads the roundscad retail price, while none of the prices serves as a leader

in period 2. Specifically, their price relationship is generally found to be that of independence. When the price series are connected, there are more regions that show

leadership of the roundscad retail price compared to tilapia, especially in regions that

are close to coastal waters, i.e., Regions 3, 4A and 4B.

Price Dynamics and Market-related Infrastructure Development

Indicative relationships of price cointegration and leadership with market infrastructure developments are examined by relating the price dynamics that are

uncovered in the earlier analyses with the changes in the identified infrastructure

development indicators (Table 7). Generally, over the study period 1990 to 2007, there is increased subscription in mobile phones, with the exception of Region 6 and

ARMM. In the case of length of national roads and number of registered utility trucks, both infrastructure indicators show improved status over the study period. These

developments could help explain the progress or regress that takes place in the price

dynamics between the wholesale and retail prices for the two species that are observed

in the respective regional markets.


Per

iod

1

Per

iod

2

All

Reg

ion

W

PT=

f(la

g

WP

RS

,t-i)

WP

RS=

f(la

g

WP

T,t

-i)

Pri

ce

Lea

der

WP

T=

f

(lag

WP

RS

,t-i)

WP

RS=

f(la

g

WP

T,t

-i)

Pri

ce

Lea

der

WP

T=

f(la

g

WP

RS

,t-i)

WP

RS=

f(la

g

WP

T,t

-i)

Pri

ce

Lea

der

Phil

ipp

ines

2

7.0

1 *

6.0

6 n

s R

ou

nd

-

scad

9

.44

ns

12

.09 n

s -

22

.71 *

1

7.6

6 n

s R

ou

nd

scad

NC

R

13

.38 n

s 3

2.2

3 *

T

ilap

ia

21

.07 *

1

5.5

1 n

s R

ou

nd

-

scad

5.9

5 n

s 2

9.1

9 *

T

ilap

ia

Reg

ion 1

1

4.7

6 *

3.8

2 n

s R

ou

nd

-

scad

2

8.7

6 *

1

3.9

3 n

s R

ou

nd

-

scad

2

7.9

5 *

1

3.3

5 n

s R

ou

nd

scad

Reg

ion 2

1

6.9

7 n

s

9.4

2 n

s -

9

.27

ns

9

.53 n

s -

19

.21 n

s 2

5.6

3 *

T

ilap

ia

Reg

ion 3

1

8.7

3 n

s 1

4.8

1 n

s -

5

.31

ns

9.8

6 *

T

ilap

ia

24

.62 *

5.8

3 n

s R

ou

nd

scad

Reg

ion 4

-A

13

.37 n

s 1

1.9

2 n

s -

11

.97 n

s 7

.37

ns

- 6

1.1

1 n

s 1

1.5

5 n

s -

Reg

ion 4

-B

29

.07 *

1

4.9

9 n

s R

ou

nd

-

scad

1

9.9

7 *

1

3.9

1 n

s R

ou

nd

-

scad

2

8.0

8 *

2

7.2

5 *

-

T

ab

le 5

. P

rice

lea

der

ship

in

wh

ole

sale

pri

ces

of

rou

nd

scad

an

d t

ilap

ia, b

y r

egio

n, P

hil

ipp

ines

, 19

90 -

2007

*

-

Sig

nif

ican

t at

α

= 5

%

ns

-

Sta

tist

ical

ly n

on-s

ignif

ican

t

-

- n

on

e of

the

pri

ces

show

ed l

ead

ersh

ip


Tab

le 6

. P

rice

lea

der

ship

in

ret

ail

pri

ces

of

rou

nd

scad

an

d t

ilap

ia, b

y r

egio

n, P

hil

ipp

ines

, 1990 -

2007

Per

iod

1

Per

iod

2

All

Reg

ion

R

PT=

f(la

g

RP

RS

,t-i)

RP

RS=

f(la

g

RP

T,t

-i)

Pri

ce

Lea

der

RP

T=

f(la

g

RP

RS

,t-i)

RP

RS=

f(la

g

RP

T,t

-i)

Pri

ce

Lea

der

RP

T=

f(la

g

RP

RS

,t-i)

RP

RS=

f(la

g

RP

T,t

-i)

Pri

ce

Lea

der

Phil

ipp

ines

2

7.0

1 *

6.0

6 n

s -

9.4

4 n

s 1

2.0

9 n

s -

22

.71 *

1

7.6

6 n

s -

NC

R

13

.38 n

s 3

2.2

3 *

T

ilap

ia

21

.07 *

1

5.5

1 n

s -

5

.95 n

s 2

9.1

9 *

T

ilap

ia

Reg

ion 1

1

4.7

6 *

3.8

2 n

s T

ilap

ia

28

.76 *

1

3.9

3 n

s -

27

.95 *

1

3.3

5 n

s T

ilap

ia

Reg

ion 2

1

6.9

7 n

s

9.4

2 n

s -

9

.27

ns

9

.53 n

s -

19

.21 n

s 2

5.6

3 *

-

Reg

ion 3

1

8.7

3 n

s 1

4.8

1 n

s -

5

.31

ns

9.8

6 *

-

24

.62 *

5.8

3 n

s R

ou

nd

-

scad

Reg

ion 4

-A

13

.37 n

s 1

1.9

2 n

s -

11

.97 n

s 7

.37

ns

- 6

1.1

1 n

s 1

1.5

5 n

s R

ou

nd

-

scad

Reg

ion 4

-B

29

.07 *

1

4.9

9 n

s R

ou

nd

-

scad

1

9.9

7 *

1

3.9

1 n

s -

28

.08 *

2

7.2

5 *

R

ou

nd

-

scad

*

-

Sig

nif

ican

t at

a =

5%

n

s -

Sta

tist

ical

ly n

on

-sig

nif

ican

t

-

-

non

e of

the

pri

ces

show

ed l

ead

ersh

ip


Tab

le 7

. C

han

ges

in

th

e le

vel

of

mo

bil

e p

hon

es a

nd

lan

dli

ne

ph

on

es s

ub

scri

pti

on

s p

er 1

00 p

op

ula

tion

in

th

e P

hil

ipp

ines

, 1999

an

d

2007

Mo

bil

e P

ho

ne

Su

bsc

rib

ers

(%)

L

eng

th o

f N

ati

on

al

Ro

ad

(k

m)

Nu

mb

er o

f R

egis

tere

d

Uti

lity

Veh

icle

s R

egio

n

19

99

20

07

Dir

ecti

on

of

Chan

ge

19

99

20

07

Dir

ecti

on

of

Chan

ge

19

99

20

07

Dir

ecti

on

of

Chan

ge

Phil

ipp

ines

3

.87

4.5

3

+

28

,523

29

,370

+

1,2

06,6

26

1,6

76,5

89

+

NC

R

14

.51

18

.15

+

1,0

02

1,0

32

+

45

7,0

40

58

5,4

63

+

Reg

ion 1

2

.13

2.8

2

+

1,5

63

1,6

10

+

53

,397

76

,921

+

Reg

ion 2

1

.03

1.0

5

+

1,7

14

1,7

65

+

30

,424

45

,087

+

Reg

ion 3

2

.61

3.1

7

+

1,9

73

2,0

32

+

16

7,7

66

23

4,3

45

+

Reg

ion 4

-A

2.4

1

2.6

8

+

2,3

35

2,4

04

+

12

7,9

26

16

9,2

38

+

Reg

ion 4

-B

1.3

1

.45

+

2,1

22

2,1

85

+

68

,884

91

,128

+

Reg

ion 5

1

.17

1.4

2

+

2,1

34

2,1

97

+

25

,587

35

,170

+

Reg

ion 6

2

.43

2.0

7

- 2

,79

7

2,8

80

+

58

,291

85

,051

+

Reg

ion 7

3

.35

3.4

7

+

1,9

78

2,0

36

+

64

,664

11

0,6

84

+

Reg

ion 8

0

.58

0.9

2

+

2,3

04

2,3

73

+

17

,203

28

,768

+

Reg

ion 9

0

.86

1

+

1,1

83

1,2

18

+

19

,384

29

,839

+

Reg

ion 1

0

1.4

1

.56

+

1,6

34

1,6

82

+

23

,329

43

,835

+

Reg

ion 1

1

2.1

9

5.5

2

+

1,4

06

1,4

47

+

44

,086

52

,307

+

Reg

ion 1

2

0.8

5

0.9

9

+

1,2

66

1,3

04

+

17

,364

42

,385

+

AR

MM

0

.4

0.2

-

*

*

*

*

*

*

S

ou

rces

: N

atio

nal

Sta

tist

ical

Coord

inat

ion B

oar

d (

NS

CB

)

Dep

artm

ent

of

Pub

lic

Work

s an

d H

igh

way

s (D

PW

H)

Bu

reau

of

Fis

her

ies

and

Aq

uat

ic R

esou

rces

(B

FA

R)

L

egen

d f

or

dir

ecti

on

of

chan

ge

in i

nfr

astr

uct

ure

fac

ilit

ies:

+

in

crea

se

-

d

ecre

ase

*

d

ata

not

avai

lab

le


Based on the trends in the price dynamics and infrastructure development, there is an observed decrease in the number of regions that show cointegrated wholesale

and retail prices in the roundscad market but no discernable pattern is observed in the case of the tilapia market. Concurrent with the improvements in telecommunication

and transport facilities, the retail and wholesale prices of roundscad and tilapia are

observed to become stationary at their level forms in period 2, hence test for cointegration is no longer valid. Therefore, there are less cointegrated wholesale and

retail prices that are found in both the roundscad and tilapia markets in the second

period. These results imply that increased communication among stakeholders in the roundscad and tilapia markets may partly be responsible in increasing competition

among fishers and traders in the market leading to the existence of uniform prices.

Also, the increase in the density of roads and bridges in the countryside must

have supported the improvements in the communication of traders in both markets

which reinforce the existence of uniform prices. In the case of the transport sector, the disappearance of cointegrated prices in period 2 coincides with the regions where

high degrees of improvements are manifested.

Based on the results of the Granger causality tests, leadership of the wholesale

and retail prices in the roundscad and tilapia markets generally diminishes from

period 1 to period 2. These results partly suggest that the development in communication and transportation facilities that generally helps in facilitating the

movements of fish from one supply point to another could possibly reduce the

influence of price leaders in the market. The increased communication between wholesalers and retailers possibly eliminates the dominance of a given price in the

market. This may be partly due to the speedy transactions among producers/fishers and traders due to improved telecommunication facilities. With more access to

production areas through increased communication and transport facilities, price

competition could be enhanced thus eliminating the chances of any price dominating

in the market.

Conclusions and Policy Recommendations

This study aims to establish the long-term price relationship between monthly

wholesale and retail prices of roundscad and tilapia in the regional markets of the

Philippines, covering the period 1990 to 2007. Specifically, it seeks to determine the existence of price integration using the cointegration analysis and the occurrence of

price leaders in these markets using Granger causality test. Moreover, the study aims

to relate how infrastructure developments that promote trade affect the integration and

leadership of tilapia and roundscad prices in the regional markets of the country.

The results of the cointegration tests show that there are more regions that exhibit long-term relationships between the wholesale and retail prices in the

roundscad market than in the tilapia market. In fact, the price cointegration that are

observed in the shorter terms, i.e., periods 1 and 2, totally disappears when the prices series are analyzed in the longer run, i.e., for the entire period 1990-2007. This

suggests that the wholesale and retail prices of tilapia become uniform and

independent from each other as market-related infrastructures develop.


On the other hand, the test for Granger causality is used in this study to determine whether prices have unilateral dependence, which could be further

characterized into price leader-follower type or bilateral/simultaneous price relationships. Results show that the wholesale price generally leads the retail price in

the roundscad markets. However, the reverse is true in the tilapia markets.

A cross-causality test is also conducted to determine if leadership-follower phenomenon exists between the prices of the two fish species. Overall, the tests show

that the wholesale price of roundscad generally leads the wholesale price of tilapia. At

the retail level, the same is observed especially in the regions that are close to coastal

waters where the supply of the former is abundant.

The disappearance of cointegrated wholesale and retail prices as market-related

infrastructures develop is indicative of increased efficiency in price formation as the

markets become more competitive. This result is supported by the reduction in the

occurrence of price leaders in both tilapia and roundscad markets as prices become more stable (or stationary in the language of time series analysis) due to increased

competition leading to uniform prices.

This study shows that understanding the price movements in the markets of

roundscad and tilapia is an important consideration in achieving efficiency in their

price formation. This, in turn, has significant implications on the production of fishers and fish farmers, and trade between wholesalers and retailers. Similarly, the

results of the study show crucial indications that improvements in market-related

infrastructures have some influence on the dynamics of the wholesale and retail prices

of roundscad and tilapia in the regional markets of the country.

The efficiency in pricing that could be deduced from the results of this study is partly attributed to the initiatives of the government in promoting communication and

transportation infrastructures in the countryside, which are vital in facilitating decision

making among fishers, fish farmers and traders. Overall, the present study suggests that recent policies in the country that promote trade of roundscad and tilapia, either in

the physical landing sites or through the cyberspace (i.e., electronic communication system) have proven helpful in improving pricing efficiency of the said commodities

in their respective markets.

Market-related infrastructures that could speed up physical movements of fish products throughout the country are still wanting. Farm-to-market roads are needed to

link fish productions areas to landing sites and finally to retail markets in fish deficit

regions. For example, the low preference of fish consumers for tilapia in the Visayas

and Mindanao may be addressed by promoting consumption of this species in the said

areas. This could be done if producers and traders are motivated by price bids that are conducive to attain normal profit margins. One way to do this is through the

reduction of transport cost and the enhancement of transparency in trading of the fish

commodities through more efficient communication.

This study, therefore, recommends that government infrastructure projects that

promote communication and transportation development should be focused on regions

that are lacking in these facilities in order to enhance movements of farm produce

from production points to consumption centers.


Given the improvements in telecommunication and transportation facilities, especially in depressed areas, the consumers, buyers and sellers could be expected to

receive benefits from increased price efficiency in the tilapia and roundscad markets. Considering that roundscad and tilapia are deemed ―poor man‘s fish‖, efforts in

making them more affordable especially for the poorer segment of the society could

make a lot of difference in addressing food security and poverty in the country.

References

Box, G.P., G.M. Jenkins and G.C. Reinsel. 1994. Time Series Analysis Forecasting

and Control. 3rd ed. Prentice-Hall Inc. NJ, USA, p. 598.

Bureau of Agricultural Statistics. 2002. Costs and Return of Tilapia Production, p. 68.

Bureau of Agricultural Statistics. 2007. Fisheries Statistics of the Philippines 2004-

2006, p. 435.

Bureau of Agricultural Statistics (BAS). 2008. ―Fisheries Situationer January-

December 2008‖.

Engel, F.R. 1982. ―Autoregressive Conditional Heteroskedasticity with Estimates of

the Variance of United Kingdom Inflation.‖ Econometrica. 50(4): 987-1007

Engel, R.F. and B.S. Yoo. 1987. ―Forecasting and Testing in Cointegrated Systems.‖

Journal of Econometrics, 35, pp. 143-159.

Engel, R.F. and C.W.J. Granger. 1987. ―Cointegration and Error Correction:

Representation, Estimation and Testing.‖ Economertica, 55, pp. 251-276.

Garcia, Y. T., M. Dey and M.S. Navarez. 2005. ―Demand Analysis for Fish in the Philippines: A Disaggregated Analysis.‖ Aquaculture Economics and

Management Journal, Vol. 9, Nos. 1-2.

Garcia, Y. T. and N. D. Salayo. 2009a. ―Price Dynamics and Market Integration of

Major Aquaculture Species in the Philippines.‖ Asian Journal of Agricultural De-

velopment, Vol. 6, No.1, June 2009.

Garcia, Y. T. and N. D. Salayo. 2009b. ―The Role of Infrastructure Developments on

Price Formation and Markets of Major Aquaculture Species in the Philippines‖, 21st DA-BAR National Research Symposium (Awarded Best AFMA R&D Paper

under the Socio-Economics Category), RDMIC Building, Diliman, Quezon City,

October 8-9, 2009.

Garcia, Y. T. and Z. M. Sumalde. 2011. ―The Role of Genetic Research and Public-

Private Partnership in the Dynamic Growth of the Tilapia Industry, 10th APAP

Forum, August 29-30, 2011, Olympic Parktel Hotel, Seoul, Korea.

Goodwin, B. and T. Schroeder. 1991. ―Cointegration Test and Spatial Price Linkages

in Regional Cattle Markets.‖ American Journal of Agricultural Economics, Vol.

79.

Goodwin, B.K. and N.E. Piggot. 2001. ―Spatial Market Integration in the Presence of

Threshold Effects‖. American Journal of Agricultural Economics, 83(2): 302-

317.


Granger, C.W.J. 1988. ―Some Recent Developments in the Concept of Causality.‖

Journal of Econometrics, 39, pp 199-211.

Granger, C.W.J. 1969. ―Investigating Causal Relationships by Econometric Models

and Cross Spectral Methods.‖ Econometrica, 37, pp. 424-438.

Greene, W.H. 2003. Econometric Analysis 5th Ed. Pearson Education Inc.,

New Jersey.

Ling B-H. 2003. ―Price Cointegration in Spatial Markets: An Application to Milkfish

Markets in Taiwan.‖ Aquaculture Economics & Management, 7, pp 85-94.

Petersen E.H. and G. Muldoon. 2007. ―Wholesale and Retail Price Integration in the Live Reef- fish Trade.‖ Economics and Market Analysis of the Live Reef-fish

Trade in the Asia Pacific Region, ACIAR Working Paper No. 63, Canberra.

Philippine Statistics Authority, 2000.

(http://countrystat.psa.gov.ph/?cont=10&pageid=1&ma=N10PRRPC)

Salayo, N. D. 2006. ―Price Relationships in Philippine Milkfish Markets: Univariate and Causality Analysis.‖ Aquaculture Economics and Management, Vol. 10, pp.

59-80.

Salayo, N. D. 1989. ―Market Integration in the Philippine Milkfish Industry‖, M.S.

Thesis, Universiti Pertanian Malaysia.

Sims, C.A. 1972. Money, Income and Causality. American Economic Review,

62: 540-552

Tubale, M. C. C. 2007. ―Chicken Market Integration in the Context of Developments

in Information and the Communication Technology (ICT) and Transportation Infrastructure, Philippines, 1995-2005.‖ BS Thesis, University of the Philippines

Los Banos.


1 Department of Agricultural Engineering and Food Engineering, Cavite State University, Indang, Cavite

[email protected] 2 College of Economics and Management, University of the Philippines Los Baños

[email protected] and [email protected]


Microcontroller-Based Coffee Roasting Machine: Implications for

Market Potential and Technology Commercialization

Ruel M. Mojica1 and Marilyn M. Elauria2

Abstract

The study assesses the socio-economic and environmental performance leading to

the commercialization of the micro-controller-based coffee roasting machine for small-

scale roasting operations. Key informant interview and coffee farmers‘ survey were

conducted to assess the farmer- respondents‘ perceptions of coffee roasting machines and

availability of resources. Results show that the designed coffee roaster almost fits the

farmers‘ criteria in selecting coffee roasting machine. Coffee farmers perceived the

machine to be a very good technology as evidenced by its highly acceptable rating.

Investment analysis reveals that using the roaster for custom work would be a profitable

business with an IRR of 76% and benefit-cost ratio of 1.44. Moreover, even with the

additional costs due to roasting, an average coffee farmer with one-hectare farm will get

an additional income of PhP 28,240.00 from coffee roasting or an equivalent of PhP 70.6

per kilo of dried berries roasted instead of selling green berries.

Keywords: coffee roasting machine, microcontroller-based, socio-economic and

environmental assessment

Introduction

Coffee ranks second to oil among the world's legally traded commodities.

Around the world there are an estimated 25 million coffee producers and workers in over 50 countries who are mostly small-scale farmers. It is estimated that around

300,000 Filipinos depend on the coffee industry that contributes about 3% in the

country‘s GDP. The national average yield is 0.3 metric tons per hectare, a very low production compared to leading coffee producing countries such as Brazil where total

production is 2,720,520,000 kilograms of coffee beans and 2,000 kg/ha in 2014. In

the case of Vietnam, the leading producer in Asia and second in the world, production is 3.1 metric tons per hectare, almost 11 times bigger than the Philippines‘

production. The largest plantation in the country can be found in Mindanao with Sultan Kudarat as the top coffee producing province with 22,613.06 metric tons

produced in 2014.

The current problem of low prices of dried coffee berries has caused shrinkage

in agricultural land planted to coffee. The town of Amadeo in the province of Cavite

for instance, where as much as 4,500 hectares used to be planted with coffee, now has a remaining 3,400 hectares of land planted and an average of 4,080 metric tons of

coffee produced annually. The total hectarage planted to coffee continues to decrease,

from 120,000 hectares in 2012 to 116,460 hectares in 2013 with an average 5%

reduction in area per year (Philippine Statistics Authority 2014).

Farm-level coffee processing is seldom practiced because of high capital

requirement for equipment and machineries which farmers normally lack. Farm owners are not keen on value-adding processes because of their need for immediate

cash, hence, the practice of selling the crops in their raw form even at a very low

price.

22 Ruel M. Mojica and Marilyn M. Elauria

The development of equipment suited to the needs of small-scale farmers has been a challenge for a long time to most engineers. Recently, a microcontroller-based

roasting machine intended for village-level operation was designed and developed. However, socio-economic and environmental assessment of the developed machine

must be done to augment findings of the study. This would provide clear information

on economic and environmental impacts that are beneficial to farmers and helpful to

decision makers.

Objectives

The main objective of the study is to assess the socio-economic and environmental performance of the developed microcontroller-based coffee roasting

machine developed for small-scale roasting operation. Specifically, it aims to: (1)

identify the coffee farmers‘ criteria for selecting roasting machines; (2) assess their

level of awareness and willingness to adopt roasting technology; (3) assess the

economic viability of the machine and its effects on the income of farmers and processors; (4) determine the environmental impact of the machine; and (5) assess the

market potential of the developed technology.

Methodology

Description of the Coffee Roasting Machine

The microcontroller-based coffee roasting machine shown in Figure 1 was de-veloped by Mojica (2010). With the exception of the roasting chamber, the auger and

the pulley driving the auger, most parts of the machine are made of G.I. sheet. Power

requirement of the roaster is 220 V, 2-kW, single–phase motor. There is no need for special skills in running the controller which is built-in or incorporated in the

machine. The user could simply enter the temperature and time of roasting and the

operation is automatic.

The outstanding feature of the roasting machine is its automatic operation which

produces evenly roasted beans. Other key features that distinguish it from other roasters include versatility, cost effectiveness, innovativeness and efficiency. With a

well-designed auger, the mechanical roaster is versatile and can also be used to roast other crops such as cacao and peanuts. The auger is designed in such a way that the

movement of beans inside the roasting drum is made uniform, thereby producing

beans of even roasts. One can vary the degree of roast (light, medium and dark) by

simply setting the time and temperature in the microcontroller.

Figure 1. The microcontroller-based mechanical coffee roaster


The machine is cost effective since it is made from locally available materials, thus the farmers and cooperatives could own it at a very reasonable price. It has a 10-

kg capacity that is appropriate to the needs of the small-scale processors. The machine has a microcontroller (Figure 2) that uses an easy-to-read display to give the user all

the information needed to successfully complete the roast. This innovativeness allows

the operator of the machine to set and see both time and temperature simultaneously for a better understanding of the relationship between the two factors. Roasting time

and temperature can be changed by 1 minute and 10°C increments, respectively.

The machine is efficient as evidenced by the shorter time to complete one cycle of roast. Sufficient amount of heat can be provided by the installed electric heaters. It

took less than ten minutes on the average to achieve the required temperature on the

next cycle of roasting process. Excessive roasting can be avoided as the desired level

of roast can be set.

Figure 2. The microcontroller circuit showing the individual components: (a) display

circuit, (b) liquid crystal display (LCD), and (c) motherboard.

Field Testing of the Coffee Roasting Machine

Thirty coffee farmer households were selected at random from the total

population of 212 farmer-households in Indang, Cavite. Using the interview schedule, data that describe the farmers‘ socio-economic condition, their coffee-production

practices, their perception/level of awareness and willingness to adopt the roasting technology and their selection and/or design criteria for a coffee roaster were collected

and analyzed. Cost effectiveness, portability, output capacity, innovativeness and ease

of operation are some of the criteria that were used.

(c)

(a (b


Sufficient amount of green coffee beans with the 12% moisture content were prepared. Prior to the test trials initial weight of the samples was measured using a

portable weighing scale. Initial moisture content of each sample was also determined using a moisture meter. Roasting temperature and time were set on the

microcontroller. Ten kilograms of green beans were loaded into the hopper when

the required temperature for each test run was reached. The roasting process was

ended when the set time has elapsed.

Farmers’ Perception and Social Acceptability

The 30 farmer-respondents were invited for a meeting and an actual field demonstration of the coffee roaster‘s performance. During the meeting, the new

technology on coffee roasting was introduced to farmers. After the demonstration,

the farmer-respondents were asked to rate the prototype coffee roaster using the

following criteria:

1. Relative advantage – the farmers‘ rating on the superiority of the coffee roaster

compared with his existing practice.

2. Complexity – the farmers‘ rating on the ease of operation of the coffee roaster.

3. Compatibility – the compatibility of the coffee roaster with the needs, values and

experience of the farmer.

4. Suitability – the adaptability/suitability of the coffee roaster to the local physical

condition of the farmer‘s farm.

5. Cost effectiveness (affordability) – the ability of the farmer to buy the machine.

In each of the five criteria, the response codes were the following: (1) poor technology, (2) average/same as the existing technology and (3) good technology.

The potential social acceptability was determined from the farmers‘ rating in the five enumerated evaluation criteria. Farmers‘ perception of the roaster and their

evaluation of the acceptability of the roaster were analyzed. The impacts of the

developed coffee roasting technology on the farmers and coffee industry were also

assessed.

Economic Assessment

The socio-economic impact of the technology, i.e., the coffee roasting

machine, was assessed in terms of the benefits to the stakeholders in the coffee

industry, namely, the coffee farmers and the owners/operators of the roasting

machine. To determine the economic viability of the coffee roaster, financial

analysis was done. The analysis focused on how the benefits obtained from the

technology (coffee roaster) measure against the investment (cost of the roasting machine). The initial investment cost and the operating cost of using the roaster as

well as the potential income were considered. From the data gathered, the different profitability measures such as annual net income, payback period, break-even point

and return on investment were computed. The added income arising from the

adoption of the technology was also measured and analyzed.


Environmental Impact Assessment

The environmental impact of the roaster was measured in terms of the reduction

in the greenhouse gas emission from the use of the electricity for the mechanical roaster as compared with the other types of roasters using fossil fuel such as kerosene

and LPG. Energy emission saved was computed as the difference between the total

emission from displaced kerosene and from electricity used in coffee roasting.

Assessment of Market Potential of the Machine

The market potential of the developed technology was assessed in terms of the

number of coffee roaster units required to process the available coffee supply. Potential adopters of the technology as well as commercialization, utilization and

replication of the machine were also considered in the assessment. The list of

prospective users of the machine was gathered through the use of internet and

personal communication. The potential markets for roasted beans include local coffee

shops and hotels. In some areas (e.g., Amadeo), the cooperative runs their own coffee

shop


Farmers’ Perception of the Roasting Machine

One of the objectives of the survey is to examine the perception of coffee

farmers regarding the coffee roaster. Table 1 reveals that majority of the coffee farmers in the area (90%) were not familiar with the mechanical roaster for coffee and

other crops. Ninety-six percent indicated that they had no experience of operating a

mechanical roaster. Regarding the ownership of coffee roasting equipment and machineries, 23 respondents (76.67%) indicated that they never owned a single

machine or equipment for coffee. However, the same number of respondents felt that having a coffee roasting machine is a necessity in the area. Moreover, they indicated

that it is important that the machine could be used either in roasting coffee or other

crops.

Table 1. Coffee farmers’ perception of the roasting machine, Indang, Cavite,

2012

Characteristic Frequency Relative

Frequency (%)

Awareness about mechanical roaster for coffee and other crops

Not aware 27 90.00

Slightly aware 3 10.00

Very much aware 0 0.00

Previous experiences with roaster for coffee and other crops

None 29 96.67

Little experience 1 3.33

Much experience 0 0.00

Ownership of coffee machineries and equipment

Coffee maker 7 23.33

Machine not available 23 76.67


A large number of survey respondents (86.67%) indicated that they prefer a small-scale roasting machine, with at least 10 kg output capacity, over a medium or

large scale machines. Twenty-three respondents (76.67%) revealed that a

microcontroller is an important feature of the machine. With regards to power source, 46.7% of the respondents prefer electricity (46.67%) while 33.33% and 20.0% of

them prefer LPG and biomass, respectively.

Characteristic Frequency Relative

Frequency (%)

Necessity of having a roasting machine for coffee

Not needed 2 6.67

Necessary 23 76.67

Very necessary 5 16.67

Can be used for other crops aside from coffee

Not important 0 0.00

Important 25 83.33

Very important 5 16.67

Scale of roasting

Small 26 86.67

Medium 4 13.33

Large 0 0.00

With microcontroller

Not important 7 23.33

Important 23 176.67

Very important 0 0.00

Output capacity

5 kgs and below 6 20.00

6 – 10 kgs 15 50.00

11 – 15 kgs 5 16.67

16 kgs and above 4 13.33

Power source

Biomass 6 20.00

LPG 10 33.33

Electricity 14 46.67

Cost of machine (pesos)

5,000 and below 11 36.67

5,001 – 10,000 8 26.67

10,001 – 15,000 7 23.33

15,001 – 20,000 2 6.67

20,001 and above 2 6.67

Materials

Stainless steel 18 60.00

G.I. sheet 7 23.33

Stainless/G.I. sheet 5 16.67


In reference to the present economic situation, the farmers perceived that owning a coffee roasting machine is an expensive business. Eleven respondents (36.67%)

indicated that the machine should not cost more than PhP 5,000.00. Only two respondents (6.67%) signified their intention of buying the machine at the price of

PhP 20,000.00 or more. Sixty percent of the respondents stated that they preferred

stainless steel over other materials such as G.I. sheet.

Social Acceptability of the Roasting Machine

Out of the total 30 original coffee farmer-respondents, only 15 of them joined in

the evaluation of the acceptability of the roasting machine― whether it is a poor technology, same as the existing technology, or a good one in terms of the different

criteria used. Table 2 shows that a high percentage of the respondents described the

roasting machine as easy to operate and compatible with their needs; this shown also

by the very high rating of 2.9 for ease of operation and 3 for compatibility with needs.

The technology was also rated good in terms of compatibility with field conditions and showed very visible difference with their own practice (selling of raw coffee

beans). The respondents were confident that they could afford to buy one at a price of PhP 30,000 per unit (assuming year 2009 selling price). However, making their own

would be difficult if no model were available. The overall mean rating is very high;

this means that the farmers perceived the roasting machine as very acceptable since it

is a very good technology.

A number of coffee farmers, cooperatives, coffee shop owners and entrepreneurs

have already signified their intention of buying a unit of micro-controlled coffee roasting machine. They have shown interest in buying at least one unit of coffee

roasting machine. The potential adopters/buyers of the machine are distributed among the different provinces in the country ― from Apayao in the north to Sulu in the south

and also include foreign buyers from the United States of America.

Table 2. Coffee farmers’ assessment of the acceptability of the roasting machine,

Indang, Cavite 2012

Economic Benefits

The economic benefit of the roasting machine was focused on the financial

profitability of the investment and additional income to coffee farmers and processors

from using the roasting machine.

Criteria Farmers’ Assessment Mean

Rating Poor (1) Average (2) Good (3)

Ease of operation

Compatibility with farmers‘ needs

Compatibility with farmers‘ field conditions

Visibility of result

Cost of roaster (PhP 30,000/unit)

7

0

40

17

13

93

100

60

83

87

2.93

3.00

2.60

2.80

2.87

Average 2.84

(percent of respondents)


The initial investment cost for the manufacture of the roasting machine (composed of the roaster and the microcontroller) is PhP485,000. Annual operating

costs amounted to PhP 222,585. Labor and electricity accounted for 93% of the total operating cost and the rest accounts for depreciation, interest on capital, repair and

maintenance and taxes and insurance. Annual gross income of PhP 432,000 was

obtained based on the existing rate of coffee roasting (PhP10/kg). With an annual net income of PhP 209,415 the return on investment shows that there is an annual net

income on PhP 0.42 for every peso invested on the roasting machine. This reflects the

high profitability of the use of the roasting machine. The income from the use of the

machine can be increased by increasing the number of hours of machine operation.

The payback period determines the number of years in which the investment capital can be recovered. Based on the analysis, the initial cost of the machine can be

recovered in 2.37 years. The level of production where the total income is equal to the

total expenses is the break-even point. The break-even point determines the volume or quantity of coffee beans that must be roasted to cover the total operating (variable and

fixed) costs. The result shows that an owner of a mechanical roaster should roast at least 22,259 kg of beans in a year in order to break-even; this involves 93 days of

8-hour operation/day. This is much lower than the actual operating days of 180 days/

year which further shows that using the roaster is highly profitable.

In order for a particular project or business to be viable in the long run, the net

present value must be positive and the benefit cost ratio must be greater than one.

Results in Table 3 show that owning a mechanical roaster and using it for custom work would be a very profitable and viable business. Using the opportunity cost of 12

% as the discount rate, the computed net present value (NPV) is PhP 750,205 and IRR

is 76% which is greater than the opportunity cost of money.

Table 3. Summary of results of the financial analysis of the machine

Item Value

Initial investment cost (PhP) 485,000 Fixed cost /year (PhP) 126,585

Variable cost /year (PhP) 96,000

Total operating cost/yearr (PhP) 222,585

Gross income/yearr (PhP) 432,000

Net income/year (PhP) 209,415

Return on investment (%) 42

Payback period (years) 2.37

Break-even point (kg/year) 22, 258.5

Net Present Value (PhP) 750, 205.75

Benefit Cost Ratio 1.44

IRR (%) 76

Source: Computations based on survey data, 2012


Impact of the Technology to Farmers and Coffee Industry

The development of the microcontroller-based roasting machine could provide

income generating opportunities to small-scale farmers as well as coffee processors. The coffee farmers can process/roast their own coffee either individually or through

the cooperative and sell their processed coffee at a better price rather than selling raw

coffee. Farmers can then demand better price for their product. The additional benefit that can be derived by the farmers from roasting will increase their family income.

Using partial budgeting, the added costs and returns due to the shift from selling dried

berries to selling roasted beans were analyzed. Even with the additional costs due to roasting, results in Table 4 show that an average coffee farmer with one-hectare farm

will get an additional income of PhP 28,240.00 from coffee roasting or an equivalent of PhP 70.6 per kilo of dried berries roasted. This will encourage farmers to plant

more coffee trees, take care of the plantation and engage in coffee processing

business. Consequently, it can be inferred that the different stakeholders in the coffee industry will continue to invest in this roasting technology because of the lure of

higher profit. This investment will result in increased livelihood opportunities for other people in the community. Better livelihood opportunities also translate into

income generation or augmentation for technology adopters and the rest of the

community. Moreover, the development of the roasting technology is a sure way of helping the country‘s ailing coffee industry. Since coffee is dollar earner, coffee

industry can produce more foreign exchange for the country.

Table 4. Additional income of farmers from roasting coffee

Source: Computations based on survey data, 2012

Item Value

A. Income based on the existing practice of the coffee farmers

(selling produce in raw form)

Volume of dried berries (kg/ha) 400 Selling price (PhP per can of 10kg) 400 Income from selling of dried berries (PhP/ ha) 16,000 B. Income from using the roasting machine

Volume of dried berries (kg/ ha) 400 Volume of green (pulped) beans (70% of dried beans) (kg) 280 Volume of roasted beans (60% of green beans) (kg) 168 Cost of pulping (Php 7/kg) 2,800 Cost of roasting (PhP 10/kg) 2,800 Total added cost (PhP) 5,600 Income from selling of roasted beans (PhP 300/kg) 50,400 Net income (PhP) 44,800 C. Additional income from roasting

Per hectare (PhP/ ha) 28,800 Per kg of dried berries roasted (PhP) 72


Sources: Mojica (2003) and IPCC (2005)

Environmental Impact of the Technology

Almost all available roasting machines are commercial in nature and are using

kerosene or LPG as fuel. The developed roasting machine is using electricity as fuel and is therefore very efficient. It emits only 0.18467 kg CO2 equivalent per batch

which compares well with 3.289 kg of CO2 equivalent for kerosene and 2.985 kg of

CO2 equivalent for LPG (Table 5). Energy emission saved is the difference between the total emission from displaced kerosene and from electricity used in coffee

roasting. A savings of 3.10433 kg of CO2eq and 2.8 kg of CO2eq will be incurred per

batch of 10 kg of coffee if electricity were to be used instead of kerosene and LPG,

respectively.

Table 5. Comparison of fuel consumption from different sources and

corresponding CO2 emissions (per 10 kg batch)

Based on the Philippine Statistics Authority, the annual production of dried berries in the Philippines in 2012 is 88,943 metric tons (MT) while the percent

contribution of Cavite is 7.96% or a total volume of 7,085.95 MT per year. Roasting the coffee produced in Cavite alone would entail 921,173.5 liters of kerosene or

708,595 kg of LPG while total national coffee production would require 11,562,590

liters of kerosene or 8,894,300 kg of LPG using the current practice (Table 6).

If coffee farmers in Cavite alone would roast their harvest using the developed

coffee roaster (using electricity) before selling them, the country could save the above liters of kerosene or kilos of LPG and could avoid net greenhouse gas emission of

2,288 metric tons of CO2 equivalent if kerosene were used and 2,072 MT of CO2

equivalent if LPG were used as fuel. On the national level, the total kerosene or LPG requirement that could be saved would be 11.56 million liters of kerosene or 8.89

million kg of LPG using the current practice and could avoid emission of 28,724 MT

of CO2 equivalent if kerosene were used and 26,009 MT of CO2 equivalent if LPG

were used as fuel.

Therefore the developed technology does not only result in less importation of

imported fossil fuel but it is environment friendly as well.

Item

Type of Fuel for Roaster

Kerosene LPG Electricity

Fuel consumption 1.3 li 1 kg 0.3292kW-hr

Heating value 35.2MJ/li 47.31MJ/kg 2.5 Kw-hr

Total energy 45.76 MJ 47.31 MJ 4.50 MJ

Unit emission 2.531 kg CO2eq/li 2.985 kg CO2eq/kg 0.5610 kg CO2eq/kW-hr

Total emission savings

In CO2 emission/batch

3.289 kg CO2eq

3.10433 kg CO2eq

2.985 kg CO2eq

2.80033 kg CO2eq

0.18467 kg CO2eq

not applicable


Table 6. Total emission avoided if electricity were used to roast Cavite coffee

production and total national production per year using 2012 data

Source: Philippine Statistics Authority (2014)

Market Prospects of the Coffee Roasting Machine

Coffee roasting in the Philippines is predominantly controlled by Nestle

Company and few other multinational companies. Commercial coffee roasters being used by these companies are of medium to large scale capacities. Farmers are forced

to sell their coffee in raw form due to the absence of farm level processing facilities including roasting. Raw coffee beans are being sold to middlemen or directly to the

companies mentioned above at very low price ― making coffee farming a less

profitable venture. For this reason, farmers are forced to shift to other crops or look for other means of employment which in turn results in the drop in the production of

coffee in the Philippines.

Table 7 shows the annual production of dried berry in the Philippines in 2012

and the percent contribution of Cavite was 7.96% or a total volume of 7,085.95 MT

per year . The developed coffee roaster can roast 30 kg of berries per hour or 240 kg/day at 8-hour operation/day. With an average 180 days of operation per year, a

roaster can process 43.2 MT per year. In the province of Cavite alone, the 7,085.95

MT of coffee would require 164 units of coffee roasters. In the top producing province of Sultan Kudarat, 526 units of coffee roasters would be needed. This shows

the high market potential for the developed roasting technology.

Table 7. List of top coffee producing provinces and their volume of production,

2012

Source: Philippine Statistics Authority (2014)

Item Cavite Philippines

Annual coffee production, MT 7,086 88,943

Volume of fuel needed

Kerosene, li 921,174 11,562,590

LPG, kg 708,595 8,894,300

Electricity, kW-hr 233,269 2,928,004

Avoided emission kg CO2eq

Kerosene 2,288,412 28,724,193

LPG 2,072,078 26,008,768

Electricity Not applicable Not applicable

Province Annual Production (MT)

Sultan Kudarat 22, 709.43

Davao del Sur 4,830.93

Bukidnon 3,948.00

Cavite 7,085.95

Philippines 88,943.00


Majority of the coffee farmers in Cavite were not familiar with the mechanical roaster for coffee and other crops and had no experience in the operation of a

mechanical roaster. Interview with key informants revealed that very few farmers had ever owned a single machine or equipment for coffee. Thus, they felt that having a

coffee roasting machine is a necessity in the area.

Also, a number of coffee farmers, cooperatives, coffee shop owners and entrepreneurs have already signified their intention to buy a unit of micro-controlled

coffee roasting machine. Prospective buyers of the coffee roaster showed interest in

buying at least one unit of coffee roasting machine. The potential adopters/buyers of the machine can be found in the different provinces in the country ― from Apayao in

the north to Sulu in the south. Potential buyers include foreign buyers from United States of America. This further confirms the favourable market prospects of the

coffee roaster and its importance to the coffee industry.

Commercialization of the coffee roasting machine could be done through partnership with any machine fabrication shop accredited by the Department of

Science and Technology (DOST); this would ensurethat the commercial model would satisfy industry standards. With the documented positive impact of coffee roaster

technology to the stakeholders in the coffee industry, the results of the study could

serve as a guide on how the government and the private sectors can expedite the commercialization and utilization of this technology. Strategies to increase awareness

among farmers and businessmen such as trade fairs or exhibits, distribution of flyers

and posters and seminars would be needed.

Conclusion and Recommendation

The farmers described the coffee roaster as easy to operate, compatible with their needs and field conditions, and showed very visible difference from their own

practice. Economic analysis of the roasting machine shows favorable results. Using

the machine for custom work could be a profitable business. Selling roasted beans instead of dried beans could give additional income to coffee farmers. The machine

could also provide income generating opportunities to small-scale coffee farmers as well as processors. The roasting operations could also provide additional employment

for the farmers and other people in the community, especially after harvest time when

farmers and farm laborers have less employment opportunities. Moreover, the technology is environment-friendly and can reduce our dependence on imported fossil

fuel. This further shows the importance of the roasting machine not only to the coffee

industry but to the economy as a whole.

The development of the technology is a sure way of helping the Philippine

coffee industry. Given the potential positive impact of coffee roaster technology to the stakeholders in the coffee industry, the results of the present study could serve as

a guide on how the government and the private sectors can expedite the

commercialization and utilization of this technology. Strategies to increase awareness among farmers and businessmen such as trade fairs or exhibits, distribution of flyers

and posters and seminars are needed.


References

Anenias, L. C. 2001. ―The Philippine Coffee Industry: A Profile.‖ BAR Digest. July –

September 2001. Vol. 3 No. 3.

Clarke, R. J. and R. Macrae. 1987. Coffee Technology. Elsevier Science Publisher

LTD, Vol. 2. (pp. 73, 89-97).

Department of Agriculture - Bureau of Postharvest Research and Extension. 2006. The Postharvest Industry Situationer in the Philippines.http://www.postharvest%

20industry%20.... 9/18/2006

Intergovernmental Panel on Climate Change (IPCC). 2005. Guidelines for National Greenhouse Gas Inventories. Volume 3. Greenhouse Gas Inventory Manual and

Coffee Commodity fact Sheet 2005.

Mojica, R.M. 2003. ―Design, Construction and Evaluation of A Batch-Type Coffee

Roaster for Small-Scale Roasting.‖ Unpublished MS Thesis. University of the

Philippines Los Baños, College, Laguna, Philippines.

Mojica, R.M. 2010. ―Development, Evaluation and Optimization of a Microcontroller

-Based Coffee Roaster.‖ Unpublished Dissertation. University of the Philippines

Los Banos, College, Laguna, Philippines.

Philippine Council for Agriculture and Resources Research and Development

(PCARRD). 1977. ―The Philippines Recommends for Coffee.‖ 1977. Los Baños, Laguna, Philippine Council for Agriculture and Resources Research and

Development.

Philippine Statistics Authority. 2014. http://agstat.psa.gov.ph/coffee


1 Socio-Economics Research Division, Philippine Council for Agriculture, Aquatic and Natural Resources Research and

Development, Paseo de Valmayor, Timugan, Los Banos, Laguna

Email: [email protected] 2 College of Economics and Management, University of the Philippines Los Banos

Email: [email protected] and [email protected]

Abstract

The practice of continuous flooding of rice fields in the national irrigation systems

has attendant issues, including the inefficient use of water, unequal distribution of water

along the irrigation canals, and greenhouse gas emissions. An ex-ante analysis was

conducted to evaluate alternative policy options in the Upper Pampanga River Integrated

Irrigation Systems, using both market and non-market approaches. The market approach

estimated the greenhouse gas avoided and water savings due to the shift in policy. The

non-market approach used choice experiment to estimate willingness to pay of farmers to

shift the current to an alternative policy.

The study concludes that there is room for a policy change that addresses greenhouse

gas emissions from rice cultivation. This is indicated by the farmers‘ willingness to pay for

this policy change. The important consideration, however, is that the policies should

address factors including water availability, greenhouse gas reduction potential of the

technology or water system being promoted by the policy, and the irrigation water price

which should be lower than the current irrigation service fee to serve as incentive to policy

adoption.

Keywords: greenhouse gas emissions, irrigation policy, choice experiment,

willingness to pay

Introduction

Water is an important input used in increasing rice productivity, with yields from

irrigated rice being consistently higher than rainfed rice. Thus, of the 4.7 million (M) hectares (ha) of riceland in the Philippines, around half is irrigated as of 2012 (BAS

2013).

In the national irrigation systems, farmers practice continuous flooding (CF) in their rice fields. This practice, however, has a number of associated issues. One of the

most important is methane emissions, as CF is conducive for the methane-producing bacteria. According to the United Nations Framework Convention on Climate Change

(UNFCCC 2005), the Philippines emitted 100,866 Gg in CO2 eq. Of these, 33% came

from agriculture. More than 60% of the total methane emissions came from

agriculture, with rice cultivation contributing around 62%. The rate of anthropogenic

greenhouse gas (GHG) emissions is now higher than the natural emission rates. Thus mitigating GHG emissions to avoid the catastrophic effects of climate change is an

important endeavour.

Another associated problem is the use of too much or unnecessary water. According to PhilRice (2007), only around 2,000 liters of water is needed to produce a

kilo of milled rice. With CF, farmers are using double this amount.

Farmers’ Willingness to Pay for an Alternative Irrigation Policy to Reduce

Greenhouse Gas Emissions from Rice Farming in the Upper Pampanga

River Integrated Irrigation System

Fezoil Luz C. Decena1 and Isabelita M. Pabuayon2

36 Fezoil Luz C. Decena and Isabelita M. Pabuayon

This leads to the unequal distribution of water along the irrigation canals. Farms located at the head/middle of the canal generally have sufficient water throughout the planting

season, while those at the tail end had to use pumps especially during land preparation

when water is important.

A policy to address GHG emissions from rice cultivation is needed. However, a

successful policy implementation is premised on a good policy analysis. The main consideration is determining the impact of the policy on the welfare of farmers. There is

a need to understand farmer adoption behaviour because the success of the policy hinges

on farmer acceptance.

This paper presents the result of the choice experiment that measured farmers‘

willingness to pay (WTP) for a change in irrigation policy and the associated welfare

effects. The paper also determined the features of the policy that is important and

acceptable to farmers.

Theoretical Framework

Choice experiment is based on the microeconomic model of utility maximization.

In the theory of consumer behaviour, individuals aim to maximize utility subject to a budget constraint. Individuals derive utility from the characteristics of the goods rather

than directly from the goods themselves. As a result, a change in prices can cause a

switch from one bundle of a good to another that will provide the most cost-efficient

combination of attributes.

Formally, individuals solve the maximization problem (Alpizar et al., 2001 pp. 6-8) :

Maxc,x U [ c1(A1)...,cn(An);z] (1)

st i. (2)

ii.

iii. z ≥0, ci (Ai) ≥ 0 for at least one i

where U is the utility function, c1(A1) is the alternate combination i (profile i) as a function of the attributes A, z is the composite bundle of ordinary goods, A is the vector

of attributes of a good, i is the number of alternative combinations of goods and

attributes, ci is the profile (of the goods) defined for all relevant alternatives, defined to be fixed and given (e.g., availability of water during critical periods), p is the price

related to the profile of the alternative ci(Ai), and cicj means the need for a single choice

only.

The amount of good z that can be purchased is also fixed. The individuals must

choose a non-negative quantity of composite goods.

In a discrete choice, the selection of a particular profile cj(Aj) implies that for a given income, the amount of ordinary good z that can be purchased is also fixed. If a

single profile cj can be chosen, then

z = y - Pj Cj ( 3)


A

A

A

A

Solving the maximization problem results to a conditional utility function:

Uj = Vj[cj(Aj),pj,y,z] = Vj (Aj, y- pjcj) (4)

And the unconditional indirect utility function:

V(A,p,y) = max [V1(A1, y-p1c1)...Vn (An, y-pnc1)] (5)

The function V[..] captures the discrete choice. It follows that the individual

chooses profile j, iff:

Vj(Aj, y-pjcj) > Vi(Ai, y-pici), i ≠j (6)

In stated preference behaviour surveys, revealed preference may sometimes seem

inconsistent with the deterministic model. This is due to the unobservable components

of the individual such as their characteristics and other attributes of the alternatives

that were not included in the experiment or measurement error. The random utility

model (RUT) is used to link the deterministic model with a statistical model to address

this. Thus a random disturbance with a specified probability distribution, ɛ is

introduced, where an individual will choose profile j, iff:

Vj(Aj, y-pjcj,ɛj ) > Vi(Ai, y-pici,ɛi), i ≠j (7)

In terms of probabilities:

P{choose j} = P{Vj (Aj, y-pjcj,ɛj ) > Vi(Ai, y-pici,ɛi); i ≠j } (8)

The error term commonly enters the utility function as an additive term, thus:

P{choosej} = P{Vj(Aj, y-pjcj)+ɛj ) > Vi(Ai, y-pici)+ɛi); i ≠j } (9)

To specify the functional form of V(...), relevant attributes (Ai) that determine

the utility derived from each alternative are selected and included in the experiment.

Methodology

The Study Area and Respondents

The study was conducted in the province of Nueva Ecija in the Central Luzon

Region in the Philippines. The survey was conducted covering data for the dry season

in 2013. The region is the largest producer of rice, contributing more than 10% to the

national output in 2012. Nueva Ecija is the single largest producer within the region,

contributing 60% of the regional production (BAS 2013). The region hosts one of the

largest irrigation systems in the country, the Upper Pampanga River Integrated

Irrigation System (UPRIIS). The specific study area was in Division 1 of UPRIIS,

particularly along Lateral Canal G which is located in Nueva Ecija.


Respondents were divided into two groups of 150 respondents per group: one group was for rice fields located at the head and middle of Canal G and have

sufficient water, and another representing the tail end of the canal where some have to use pumps due to insufficient water in the dry season. Choice experiments require a

minimum of 50 responses per attribute (Kjaer 2005).

Steps in the Choice Experiment

There are five major steps in the choice experiment (CE): (1) identification of

attributes and levels; (2) experimental design; (3) questionnaire design; (4) data

collection; and (5) data analysis.

The first step is necessary because CE requires respondents to choose one

alternative from a set of alternatives. Each alternative has its own description known

as the attributes. The variation across the alternatives is due to the different level of

each attribute. Attributes refer to the characteristics of the good under study. Levels

are the specification of said characteristics. In this study, the goods valued were the alternative irrigation policy options for water management for reduced GHG. The

attributes were the characteristics of these irrigation policies. Each characteristic of

the policy options had several levels.

The attributes and levels that were included in the choice set were initially taken

from the literature review. These were validated through key informant interviews (KIIs) and discussions with National Irrigation Administration (NIA) officials in

Laguna and UPRIIS, irrigators association officials and selected farmers in the study

area. From these KIIs, four attributes and two levels per attribute were selected

(Table 1).

Table 1. Attributes and levels for a choice experiment involving willingness to

pay for alternative irrigation policy in Division I, UPRIIS, dry season,

2014

Attribute Description Level

Water

availability

The system of irrigation water regime

that will be followed by farmers during

rice planting in the dry season. This

system is the basis of NIA‘s release of

water to the canals.

1) Continuous floodinga/

2 ) Single drainage

3) Multiple drainage

Profitability The degree of change of profits due to

the change in irrigation water regime in

rice planting.

1) same a/

2) 10% higher

3) 10% lower

GHG

emission

reduction

The reduction in emission of greenhouse

gases (methane and nitrous oxide) as a

result of the change in irrigation water

regime in rice planting.

1) None a/

2) 30% GHG emission

reduction

3) 60% GHG emission re-

duction

Price of

water

Irrigation water price or irrigation ser-

vice fee

1) 3.5 cavans (PhP 2,975) a/

2) 2 cavans (PhP 1,700)

3) 1 cavan (PhP 850) a/ represents status quo


Water availability. While CF was the current system, the proposed water management strategies under the alternative policies were (a) single drainage (or

midseason, at 50 to 60 days after planting, drained for 10 days) and (b) multiple drainage (or alternate wetting and drying, allowing the field to dry maintaining up to

5 cm of water below the surface before flooding again). Water was always available

for irrigated farms during both wet and dry seasons. Some farms at the tail end of the

canal, however, experienced water shortages in the dry season.

Profitability. Based on technical studies, single and multiple drainage did not

significantly affect yield nor increase inputs (Fowler 2011, 2012; IRRI 2011; Rejesus et al. 2011). Based on KIIs, farmers and some experts claimed decreases in

yield and/or increases in inputs like labor and herbicides because of weeds. Farmers are concerned with the effect of change in water regime on the costs due to changes in

inputs, yield, and level of labor and fertilizer use, among others.

GHG emission reduction. Farmers are also concerned with the environment, especially if they are made to understand their role in GHG emissions, the effect of

these emissions on the climate, and the adverse impact of climate change on rice productivity in the form of prolonged drought or flooding. Based on literature review,

the practice of field draining generally reduces GHG by 30% to 80%.

Price of irrigation water. To provide incentive for farmers to practice single or multiple drainage, water prices for the alternative water management were pegged

lower than the irrigation service fee (ISF) of 3.5 cavans/ha. The proposed per hectare

alternative prices were 2 cavans and 1 cavan. The price of palay used was PhP 17/kg,

consistent with National Food Authority (NFA) support price used by NIA.

The second step, experimental design, involves drawing up of simulated choice sets by combining attributes and levels of the alternatives. Given the 4 attributes and 2

levels per attribute excluding the status quo, a total of 16 combinations (24 = 16) or

scenarios were generated. Using orthogonal design, the optimal number of scenarios and choice sets were reduced to 8. Each choice set consisted of combined two policy

attributes and the status quo as the opt out scenario. Pictures such as visual aids containing illustrations of each attribute were used. An example of a choice set from

which farmers selected their choices is presented in Table 2.

The questionnaire was of three parts. The first part included the basic socio-economic characteristics and the farming practices of the respondents. The

second part was the CE. The third part consisted of the follow-up questions to

determine validity and reliability of responses.

In data collection, the farmers‘ CE started with an introduction about the study.

In particular, farmers were informed about a) the rationale and objectives of the study, b) the reason for choosing the respondent, c) the description of the overall scenario of

water use and GHG emissions in rice farming, and d) the role of GHG in climate

change. Each farmer respondent was asked to choose one option from each of the

simulated choice set.

Under data analysis, information on socio-economic profile of respondents,

farming practices, attitudes towards climate change and irrigation water were tabulated and analyzed using descriptive statistics such as averages, frequency counts

and percentages.


Table 2. Sample choice set

Katangian Policy A Policy B Status Quo

Water

availability or

regime

Single drainage

Multiple drainage

Continuous flooding

Profitability

10% lower

10% higher

same

Greenhouse gas

emission

reduction

30% GHG emission

reduction

60% GHG emission

reduction

none

Irrigation

water price

PhP 850/ha

(1 cavan)

PhP 1,700

(2.0 cavans)

PhP 2,975/ha

(3.5 cavans)

I choose

(Check 1)


Conditional logit model. The CE data was first analyzed by fitting a conditional logit model (CLM). CLM estimates how individual specific variables affect the

likelihood of observing a given outcome. Thus, CLM allows the estimation of the effect of choice-specific variables on the probability of choosing a particular

alternative. The model is estimated by maximum likelihood estimation (MLE).

The idea in CE is that consumers derive utility or satisfaction from the goods through the attributes that the goods provide. Applying this in the study, farmers

derive utility or satisfaction from the specific characteristics of the alternative

irrigation policy option for water management.

The estimating model for the indirect utility function was of the form:

Vin = ß1 Water + ß2 Profit + ß3 GHG + ß4 ISF (11)

where Vin is the indirect utility for farmer i associated with alternative irrigation policy n; β’s are the coefficients to be estimated; Water is water availability; Profit

refers to profitability; GHG is the GHG emission reduction; and ISF is the water

price.

CLM is built on the assumption of independence of irrelevant alternatives (IIA)

which stipulates that the ratio of the probabilities of choosing any option will be

unaffected by the attributes or availability of other options (Pearce et al. 2002).

Random parameter logit model.When the CLM models failed the IIA test, the

random parameter logit (RPL) model was used, utilizing the CLM as the base model. The RPL makes no IIA assumption and incorporates unobserved heterogeneity of

preferences and tastes of respondents in the model by allowing the parameters to vary over individuals (Hensher et al. 2005). Thus, the RPL model extends the CLM model

by allowing one or more of the parameters in the model to be randomly distributed.

By allowing the coefficients to vary, the model implies that different decision makers

have different preferences.

Estimation of willingness to pay. From the model, the farmers‘ marginal willingness to pay (MWTP) for each attribute of the irrigation policy option was

computed using (Pearce et al. 2002):

(12)

where bx is the coefficient of attribute X, or the utility from an extra unit of

attribute x, and β is the coefficient on price, or the value in money terms of one more

extra unit of attribute x.

Welfare effects of policy. The welfare effect was derived from the MWTP values by calculating the compensating surplus (CS) of farmers, following Birol

(2006):

CS = (- (V0 - V1)) / β(monetary attribute) (13)


Where CS is the compensating surplus; V0 is the indirect utility derived from the status quo; V1 is the indirect utility derived from the alternative policy; and

βmontery attribute is the coefficient of the monetary attribute derived from the model. The functional form of this equation was derived using different scenarios associated with

the policy options advocating for the different water management technologies.

Measuring aggregate welfare. Aggregation of welfare benefits followed Pearce

et al. (2002):

Aggregate WTP = N x MWTP (14)

where N is the number of people in the population and MWTP is the marginal willingness to pay. In this study, MWTP was first computed per farmer-respondent.

This was then converted to per hectare to be consistent with the estimation of market

values.

The aggregate welfare was calculated by upscaling the per hectare values. This

was done by multiplying the per hectare value with the number of hectares in

Division 1 of UPRIIS.


Socio-Economic Profile of Respondents and Farm Characteristics

Farmer-respondents have similar socio-economic characteristics across farm

locations. Around two-thirds are landowners, male and were at the prime of their lives, with more than 60% within the age range of 40 - 60 years old. All of them

attended school, with around 50% having reached high school (Table 3).

Table 3. Socio-economic characteristics of farmer-respondents, Canal G,

Division 1, UPRIIS

Head/Middle (n=150)

Tail (n=150)

All (n=300) Characteristic

No. % No. % No. %

Gender

Male 102 68 125 83 227 76

Female 48 32 25 17 73 24

Average household

size 4 4 4

Age (years)

19 - 40 12 8 10 7 22 7

40 - 60 93 62 106 70 219 67

over 60 45 30 34 23 93 26

Years in farming

0 - 10 34 22 23 15 57 19

11-20 40 27 40 27 80 27

21-30 41 28 48 32 89 29

more than 30 35 23 39 26 74 25


All farms at the head/middle of the canal have sufficient irrigation water

provided by the National Irrigation Administration (NIA) as their main and only source of water (Table 4). Farmers from this location practice CF from land

preparation up to a month before harvesting. The level of water varies, with some

farmers claiming that they reduced the standing water levels during fertilization

period.

Table 4. Farming practices of farmer-respondents, Canal G, Division 1, UPRIIS

Education

Elementary level 44 29 40 27 84 28

High school level 72 48 79 52 151 50

College level 34 23 31 21 65 22

Tenure

Owner 100 67 125 83 225 75

Tenant 42 28 13 9 55 18

Others 8 5 12 8 20 7

Head/Middle (n=150)

Tail (n=150)

All (n=300) Characteristic

No. % No. % No. %

Head/ Middle (n=150)

Tail (n=150)

All (n=300) Farming Practice

No. % No. % No. %

Main source of irrigation water in the DS

NIA 150 100 150 100 300 100

Pump 82 55 82 27.3

Others 1 0.3

Irrigation water sufficiency

Sufficient 150 100 67 45 217 72

Not sufficient 83 55 83 28

Irrigation practice

Continuous flooding 150 100 110 73 260 87

Intermittent irrigation 0 40 27 40 13


Numbers in parentheses are standard errors

*** Significant at 1% level; **Significant at 5% level

ns = not significant

In contrast, 55% of the farmers at the tail of the canal said that irrigation water from NIA was not sufficient. Although all farmers recognize that NIA water is their

main source, these farmers get supplementary water from pumps at various stages of farming operations, most notably in seedbed preparation, land soaking and land

preparation, and even during the normal irrigation period. Thus, about 27% of

farmers at the tail end claim to practice intermittent irrigation to save on fuel and

other costs associated with pumping.

Factors Affecting the Policy Choice: From the Conditional Logit Models

Model for the Complete Sample Set

Table 5 presents the three conditional logit models (CLM): head/middle,

representing sufficient water condition; tail, representing insufficient water condition

and use of pumps to augment or supplement water during critical periods; and all

samples, representing respondents from both water conditions.

Table 5. Conditional logit models for choice experiment for alternative irrigation

policy options, all samples, Canal G, Division 1, UPRIIS, 2014

In CLM, the test of significance used is the Wald statistic (z statistic),

interpreted in the same way as that of t or F statistic. If the absolute value of the

statistic given in the output is greater than the critical value of 1.96, the hypothesis that the parameter equals 0 is rejected and the explanatory variable is considered

statistically significant (Hensher et al. 2005).

Except for the variable profitability at the head/middle sample, all coefficients

for the three models are significant. This means that these factors are considered by

farmers as important in any alternative policy for irrigation for reduced GHG from

rice production in the national irrigation system.

Policy Attribute Head/Middle

(n=150)

Tail

(n=150)

All

(n=300)

Water availability 0.1911** 0.2286*** 0.1878***

(0.0795) (0.0718) (0.0526)

Profitability 0.0051ns 0.0082** 0.0066**

(0.0046) (0.0041) (0.0030)

Greenhouse gas

reduction

0.0215***

(0.0027)

0.0214***

(0.0024)

0.0203***

(0.0018)

Water price -0.0116*** -0.0002*** -0.0006***

(0.00009) (0.00007) (0.00006)

Pseudo R2 0.0601 0.0577 0.0248

Log-Likelihood -1904.49 -1909.36 -3951.98

No. of Observations 3600 3600 7200


Survey data indicate that farmers at the head/middle location of the canal realized higher profits compared with farmers at the tail. Thus, farmers at the head/

middle may already be confident that their profitability will not be affected by a

policy change.

The positive sign of the coefficients means that a change in water availability,

profit and GHG reduction potential of the policy would contribute positively to the utility of farmers. The negative sign of the coefficient for the water price indicates that

an increase in price would decrease utility of farmers. This is consistent with

economic theory of the inverse relationship of price and utility of consumers.

Utilizing equation 12 to determine the implied ranking of the attributes, the most

important policy attribute is water availability, followed by GHG reduction and

profitability (at the tail). This implies that any policy change should take these

attributes into consideration in designing the policy.

The MacFadden R2or the Pseudo R2 was low at 0.0709, compared with the conventional good fit of 0.2 to 0.4 (Pearce et al. 2002; Kjaer 2005). According to

Hensher et al. (2005), an R2 of 0.3 in CLM, estimated using MLE, is approximately equal to 0.6 in ordinary least squares estimation. The low R2 for this model implies

that the model does not represent reality and therefore has a very low predictive value.

Because of this, there is a need to validate the data further to determine whether the

model can be of use in policy analysis.

Protest Bids and Validity Responses

One of the important ways of validating the CE data is by identifying protest bids. In this study, status quo responses are considered as protest bids. The preference

for status quo is a clear indication that farmers would not be willing to pay for any change in policy. Table 6 summarizes the farmers‘ choice of policy. A high

percentage of protest bids were observed in the head/middle sample group (42%),

compared with the sample at the tail (9%). Overall, 25% of the respondents prefer the

status quo.

Table 6. Distribution of farmer responses according to policy choice, Canal G,

District 1, UPRIIS

Head/ Middle Tail All Policy Choice

No. % No. % No. %

Status quo 63 42 13 9 76 25

Alternative (Policy A or

Policy B)

87 68 137 91 224 75

Total 150 100 150 100 300 100


Follow-up questions were made to determine validity of responses. These questions focused mainly on the knowledge of farmers about the relationship of rice

planting, GHG emissions, climate change and its effect on rice planting. A question on farmer belief that irrigation water from NIA should be free of charge was included.

If respondents believed that water should be free, then responses to any change in

policy on irrigation would not be valid. This was because of the premise that if they believed that water should be free, then they will not be willing to pay for its use

regardless of the policy being implemented.

Many respondents at the tail end of the canal think that water from NIA should be free (43%) (Table 7). It must be noted that these respondents were given discounts

ranging from 10% to 50% during the dry season because water was either late or insufficient (personal conversation with Engr. Galapon of NIA). In fact, survey

results indicate that many farmers at the tail do not pay the ISF despite using

irrigation water when this becomes available. For these farmers, paying for a poor service of the NIA was not acceptable. Almost one-third of the farmers at the head/

middle of the canal also thought that irrigation water should be free. This is the case

despite the sufficient water they get from NIA.

Table 7. Farmer-respondents’ perception on greenhouse gases (%)

Half of the farmers at the tail and 27% at the head/middle end of the canal

believe that rice planting emits GHG. The high negative responses of 72% in the

head/middle would explain the high status quo responses of farmers in this location. However, upon probing, this perceived emission of GHG is related to the heat

generated by fertilizers and felt by farmers during fertilization.

Farmers‘ knowledge about GHG is high as reflected by their responses to the

question about GHG causing climate change, and climate change having an adverse

effect on rice farming. The survey period (December 2013 to March 2014) was done a month after typhoon Haiyan or Yolanda (November 2013), and the recent storm in the

area destroyed the crops of many farmers.

The responses to the validity questions were also used in identifying protest bids other than the status quo response. In particular, the study used the question on the

farmer belief that ―irrigation water should be free‖ to gauge their willingness to pay

(WTP).

Head/Middle

(n=150)

Tail

(n=150)

All

(n=300) Farming Practice

Yes No Yes No Yes No

Irrigation should be free 31 69 43 57 37 63

Rice planting emits

greenhouse gas 27 72 50 45 39 58

Greenhouse gases causes

climate change 72 15 87 13 79 14

Climate change has adverse

effect on rice farming 94 6 85 15 89 11


As many as 31% of farmers at the head/middle and 43% of farmers at the tail believed that irrigation service should be free. This attitude implies that they were

indifferent between the irrigation policies. Therefore, their responses will not matter because they will be just answering for the sake of answering the experiment

questions. This was the reason for considering them as protest bids.

The model was reestimated by dropping these protest bids. Results of the reestimation are shown in Table 8. Compared with the CLM estimates for the

complete sample, the magnitudes of the coefficients of these new set of models are

slightly higher. The log likelihood ratio for all three sample groups improved, implying that this is the better model. The McFadden R2 also improved. In the new

models, however, the profitability attribute was found to be not significant for all

three sample groups.

Table 8. Conditional logit model for choice experiment for alternative irrigation

policy options excluding status quo and invalid responses, Canal G,

Division 1, UPRIIS, 2014

The coefficients of the model at the tail are slightly lower than those at the head

and middle of the canal. In addition, the coefficient for water price is not significant at

the tail, indicating that farmers using pumps are not willing to pay for a shift in irrigation policy. These results contradict the idea that because farmers at the tail have

insufficient water, they would automatically be willing to pay or would prefer to have policies that would ensure water for their farms at all times. This attitude could be due

to distrust with how policies or regulation on water distribution are being

implemented based on their experiences. It is also possible that farmers are already quite used to utilizing the pumps, and this gives them the excuse not to pay NIA even

though NIA water is being used during the normal irrigation period (NIP).

However, the IIA test for this reduced model indicates that IIA was violated. This means that dropping any alternative policy will significantly affect the choice for

the other policies.


(n=75)

Tail

(n=69)

All

(n=144)

Water availability 0.8143*** 0.3880*** 0.5942***

(0.1067) (0.1055) (0.0744)

Profitability 0.0080 ns 0.0077ns 0.0077*

(0.0062) (0.0061) (0.0043)

Greenhouse gas reduction 0.0425***

(0.0036)

0.0260***

(0.0035)

0.0334***

(0.0025)

Water price -.0009*** 0.00002 -0.0005***

(0.0001) (0.0001) (0.00008)

Pseudo R2 0.1804 0.1306 0.1481

Log-Likelihood -830.31 -810.37 -1657.14



***Significant at 1% level; *Significant at 10% level


Factors Affecting the Policy Choice: From the Random Parameter Logit Model

Since the CLMs violated the basic assumption of IIA, the random parameter

logit (RPL) was estimated. The RPL or ML allows for the heterogeneity of individual preferences and enhances the accuracy and reliability of the estimates. It uses the

CLM as its baseline model (Hensher et al. 2005).

The RPL model was estimated using the CLM model for reduced sample set. The structure of the data set was the same as those of the CLM. Three models were

also estimated: head/middle, tail and all samples. The standard deviation of the

random variables determined whether the distribution of the responses varied across samples or that these responses exhibit heterogeneity across samples. Whenever the

standard deviation turned out to be not significant, the variable was treated as a fixed

variable in the succeeding simulations. The final models were estimated with the

attributes price, profitability and GHG reduction as fixed parameters, and water

availability as random variable.

The coefficients in the RPL model have lower magnitudes or values but have

the same signs as those estimated from the CLMs in Table 8 (Table 9). The significance of the coefficients is also the same, with the coefficient for profitability

variable for the head and middle and all sample groups not significant, as well as the

price coefficient for the tail sample group.

Table 9. Mixed logit (random parameter) model for choice experiment for

alternative irrigation policy options, excluding the no status quo and

invalid responses, Canal G, Division 1, UPRIIS, 2014


(n=75)

Tail

(n=69)

All

(n=144)

Water availability 0.5328***

(0.0938)

0.2751**

(0.0900)

0.3957***

(0.0642)

Profitability 0.0051

(0.0049)

0.0047

(0.0048)

0.0048ns

(0.0040)

Greenhouse gas

reduction

0.0280***

(0.0029)

0.0175***

(0.0028)

0.0224***

(0.0020)

Water price

-0.0005***

(0.0001)

-0.00006ns

(0.00009)

-0.0002***

(0.00007)

Log-Likelihood -534.50 -519.27 -1064.71



***Significant at 1% level; **Significant at 5% level

ns not significant

The results show that water availability has the highest contribution to the utility of farmers, followed by GHG emission reduction, profitability and price. This means

that for policies to be acceptable to farmers, the focus should be given to these

variables.


Farmers' Willingness to Pay for a Change in Irrigation Water Policy

Two levels of welfare effects were estimated from the CE. The first was the

WTP of the farmers for a shift from the current policy (status quo) to an alternative irrigation policy. The second was the WTP of the farmers for a shift from the current

policy (status quo) to the specific policy where the technology for GHG emission

reduction and other levels of the policy attributes are already specified.

Willingness to pay for a shift to alternative irrigation policy. The RPL model for

all samples from Table 9 was used in the estimation of the WTP. Applying equation

12, the individual farmers‘ WTP for alternative policy options for reduced GHG

emissions was calculated. Results are presented in Table 10.

Table 10. Marginal willingness to pay estimates (PhP) of rice farmers for

alternative irrigation policy options, Canal G, District 1, UPRIIS

In calculating the WTP, it is important that both attributes are statistically significant, otherwise no meaningful WTP can be established (Hensher et al. 2005).

For the tail model, the price coefficient was not significant; hence, no WTP can be

inferred.

Results show that farmers at the head/middle are willing to pay PhP 975 for a

change in policy that has features of water availability (either single drainage or multiple drainage systems), and PhP 51.31 for a policy with features that would

reduce GHG. In contrast, farmers at the tail are not willing to pay for policies

containing any of the four attributes. The non-significance of the coefficient for the price indicates that farmers‘ utility or happiness is not affected by the price of water.

Therefore, they may not care at all for any change in policy. Overall, the individual farmers‘ WTP for a change in policy that contained the attributes of water availability

and GHG reduction was PhP 1,767.39 or approximately equal to 2 cavans.

Willingness to pay for a shift to a specific irrigation policy. The estimation of the welfare effects because of the change into a specific policy was done using the

compensating surplus (CS) approach.

The following scenarios were used in computing the CS: the status quo scenario (V0) features CF, current profitability level; no GHG reduction and ISF of PhP 2,975;

the medium impact scenario (V1) features water regime of midseason drainage of the field, 60% reduction in GHG emission and no change in ISF; and finally the high

impact scenario (V2) features multiple drainage water regime, 60% reduction in GHG

emission and no change in ISF. The estimated RPL model for the all samples was

used.

a/No estimates because model coefficients were not significant


(n=75)

Tail

(n=69)

All

(n=144)

Water availability 975.48 -- a/ 1,757.91

Profitability --a/ -- a/ --a/

Greenhouse gas reduction 51.31 -- a/ 99.48

Total 1,026.79 1,767.39


Following equation 13, the CS was obtained for each farmer. The calculated total MWTP for the high impact scenario was PhP 9,484 per farmer and PhP 7,726 for

the medium impact scenario. Given that the average farm size was 1.4 ha, the per hectare marginal WTP for the high impact scenario of the policy featuring multiple

drainage was PhP 6,744, while it was PhP 5,518 for the medium impact scenario or

for policy featuring single drainage. These values may be compared with the ISFs that were being paid in the communal irrigation systems, where the ISFs can be as high as

10 to 12 cavans/ha in the dry season, most particularly in the Ilocos Region (NIA

2012). At the rate of PhP 17/cavan, these values translates to PhP 8,500 to PhP 10,200/ha. For this study, the per hectare total MWTP is equal to around 8 cavanss or

2.27 times the current ISF for the high impact scenario, and 6.5 cavans or 1.8 times

for the medium impact scenario.

These findings on high WTP implies that farmers are indeed receptive to any

policy changes in the irrigation system, especially if such policies could help mitigate GHGs. The result that the WTP of farmers is higher than the actual ISF implies that

farmers may be willing to pay a higher fee for irrigation water services than the current rates they are paying. Farmers at the tail are already spending more for water

fees because of expenses related to pumping.

The aggregate welfare was estimated following equation 14. Results indicate that on the whole, farmers´ WTP for a policy for a high and medium impact scenarios

could be as high as PhP 125,431,218 and PhP 102,187,717, respectively, for the

whole of Division 1 in UPRIIS. These amounts are double the potential total ISF collections in Division 1 for one season if all farmers pay. With the Division 1 area of

18,515.66 ha and ISF of PhP 2,975/ha, the potential total ISF is PhP 55,082,125.

Conclusions and Policy Recommendations

The results of the study reveal the implied ranking of the factors influencing

farmers‘ willingness to pay. Water availability tops the list, followed by greenhouse gas reduction potential and water price or irrigation service fee. Profitability was

found to be not important. Farmers at the head and middle of the canal were willing to pay PhP 1,026 during the dry season for a policy that promotes an irrigation system

that reduces greenhouse gas emissions. In contrast, farmers at the tail were not willing

to pay any amount at all, given that the coefficient for water price was not significant. All in all, the farmers in the whole of Canal G were willing to pay PhP 1,767, an

amount equal to 2 cavans of palay and lower than the current ISF of 3.5 cavans/ha.

Using compensating surplus approach, shifting from status quo to a high impact

scenario of multiple drainage and 60% greenhouse gas reduction leads to a welfare

effect of PhP 6,774/ha. Shifting to medium impact scenario of single drainage and 30% reduction has a slightly lower welfare effect of PhP 5,518. On aggregate, given

18,515 ha in UPRIIS, the high impact scenario would generate a welfare impact of

PhP 125 million and the medium impact scenario would generate PhP 102 million.

In conclusion, the study finds that farmers are willing to pay for a shift in

irrigation policy from continuous flooding to controlled irrigation. The welfare effect

of shifting to multiple drainage is higher than the welfare effect of shifting to single

drainage, as evidenced by the higher consumer surplus.

A change in irrigation policy could be successful if this would take into consideration the factors that are important to farmers such as water regime or

availability of either single or multiple drainage, the GHG reduction potential of such

policy, and the rate of the irrigation service fee.


References

Alpizar, F., F. Carlsson, and P. Martinsson. 2001. ―Using Choice Experiments for

Non-Market Valuation.‖ Working Paper in Economics No. 52. Department of

Economics, Goteborg University, Sweden.

Birol, E., K. Karousakis and P. Koundouri, P. 2006. ―Using a Choice Experiment to

Account for Preference Heterogeneity in Wetland Attributes: The Case of

Cheimaditida Wetland In Greece.‖ Ecological Economics 60(1): 145-156.

Bureau of Agricultural Statistics (DA-BAS) CountryStat database. (available online

at www.bas.gov.ph). Accessed on September 12, 2013.

Decena, F. L. C. 2015. ―Economic Analysis of Alternative Irrigation Policy Options

for Methane and Nitrous Oxide Emission Reduction from Rice Cultivation in Nueva Ecija, Philippines‖ PhD. Dissertation. University of the Philippines Los

Baños, Laguna Philippines.

Fowler, J. 2011.―Mid-season Drainage of Rice.‖ IREC Farmer‘s Newsletter – Large Area No. 185: Spring 2011. Available at http://irec.org.au/farmer_f_pdf_185/

Mid-season-drainage-of-rice.pdf. Accessed November 17, 2011.

Hensher, D., M. Rose and W. Greene. 2005. Applied Choice Analysis: A Primer.

Cambridge: Cambridge University Press.

Hanley, N., R. Wright and V. Adamowicz. 1998. ―Using Choice Experiments to Value the Environment: Design Issues, Current Experiences and Future

Prospects. ”Environmental and Resources Economics 11(3-4): 413-428.

International Rice Research Institute. 2012. ―Impact Evaluation of Bohol Irrigation Project (Phase 2) in the Republic of the Philippines.‖ Unpublished Terminal

Report submitted to the Japan International Cooperation Agency.

Kjaer, T. 2005.―A Review of the Discrete Choice Experiment – With Emphasis on its

Application in Health Care.‖ Health Economics Paper 2005:1. Health

Economics, University of Southern Denmark.

National Irrigation Administrationwebsite.www.nia.gov.ph. Accessed November 15,

2012.

Pearce, D. and E. Ozdemiroglu. 2002. ―Economic Valuation with Stated Preference

Techniques, Summary Guide.‖ Department for Transport, Local Government

and the Regions: London.

Philippine Rice Research Institute. 2007. ―Controlled Irrigation: Saving Water While

Having Good Yield.‖ Rice Technology Bulletin 2007 No. 59. Philippine Rice

Research Institute.

Rejesus, M., F. Palis, D. Rodriguez and R. Lampayan.2011. ―Impact of the Alternate

Wetting and Drying (AWD) Water-Saving Irrigation Technique: Evidence

from Rice Producers in the Philippines.‖ Food Policy 36(2011) 280-288.

United Nations Framework Convention on Climate Change. 2005. ―Sixth

Compilation and Synthesis of Initial National Communications from Parties Not Included in Annex I to the Convention, Addendum: Inventories of

Anthropogenic Emissions by Sources and Removals by Sinks of Greenhouse

Gases.‖ UNFCCC.



Email: [email protected] 2 Ateneo de Naga University, Ateneo Avenue Bagumbayan, Naga, Camarines Sur

Email: [email protected] 3 Forest Engineering and Watershed Management Lab. University of the Ryukus, 1 Senbaru, Nishihara, 903-0213

Okinawa, Japan

Email: [email protected] 4 Shijyonawate Gakuen University, Osaka

Abstract

The Laguna Lake Region is one of the places in the Philippines that are most

vulnerable to natural disasters because of their location and ecological condition. Among

the natural calamities that regularly hit the country, one of the most frequent and disastrous

is flooding. The most affected are the poor who are faced with the double risk of being

food insecure and living in conditions that are very vulnerable to natural hazards. The most

negative and long-term effect of flooding is reduction in food security because flooding

erodes the asset base of households that in turn results in the adoption of negative

adaptation strategies. The impact of flooding on household food security however depends

on disaster resilience - the capacity of households to absorb the adverse consequences. But

what constitutes resiliency? What factors affect resiliency?

A case study was made in the Sta. Rosa-Silang Subwatershed with the intention

of developing a resiliency index that would indicate the capacity of households to absorb

the negative consequences of flood disasters. Results show that resilience, which is

multidimensional, is determined by demographic and socioeconomic conditions, social

capital, amount of damages or losses, social safety nets, and quality of local governance.

Specifically, the results show that household disaster resiliency is negatively related to the

level of exposure to natural hazards and positively related to the economic capability of

households and the community standard of living. In addition, the study demonstrates that

enhancing household resiliency could be an important component of any strategy to

address food insecurity due to natural hazards. Thus, in disaster-prone areas, measures to

enhance household disaster resilience should be an integral part of food security strategies

and policies. The role of the resilience index becomes crucial to the evaluation of the

conditions of a target population.

Keywords: household disaster resilience, natural hazards, food security, watershed

Introduction

The Laguna Lake Region is among the most vulnerable to natural disasters in the country due to its location and ecological condition. It has a long history of flood

disasters, the most recent and damaging of which were Typhoon Milenyo, Typhoon

Ondoy, Typhoon Pepeng, and Typhoon Santi. These typhoons hit the country

between 2006 and 2009, leaving a total of 1,233 dead, injured and missing, 318,055

damaged residences, and PhP 2.12 billion in damages to agriculture (GFN 2013).

Flooding is not confined to typhoon events and may also be caused by heavy rains, as shown by the flooding that hit the towns of Sta. Rosa, Biñan and San Pedro, Laguna

(Carcamo 2013).

Building Disaster Resilience to Address Household Food Security:

The Case of Sta. Rosa-Silang Subwatershed

Roberto F. Rañola Jr.1, Michael Cuesta2, Bam Razafindrabe3 and Ryohei Kada4

54 Rañola, Cuesta, Razafindrabe and Kada

Recovery from these kinds of event has been difficult for farmers, fishers, small business owners and informal sector workers. Natural calamities bring about

disruptions in consumer income flows resulting in the loss of productive assets, diversion of capital to basic consumption and housing repairs, and depressed market

demand (World Bank 2011). Flood events in the region have been attributed to the

limited capacity of most upstream river channels draining into the Laguna Lake to confine floodwaters, the slow-flow capacity of the outlet channel from the Laguna

Lake, and the declining capacity of the Laguna Lake itself to serve as a detention

reservoir for floodwaters (GFN 2013). Rapid population growth and urbanization characterized by a strong expansion of informal settlers along the lakeshore who are

drawn by economic prospects also add to the risk of a disaster.

A lot of effort has been exerted to respond to these combined sources of stresses

from natural disasters which, coupled with conflict and chronic poverty, have directly

threatened the lives and food security of millions of people. A common concern with these responses is that while they have saved lives, they have not increased the

capacity of affected populations to withstand future shocks and stresses. Another major concern is that there is very little evidence on which among the approaches to

building resilience represents the best ‗value for money‘ (Frankenberger et al. 2012).

Helping people cope with current changes in their environment, such as adapting their livelihoods and improving the ecosystem health as well as the

governance systems, can help build the ability to avoid future problems and enhance

the resilience of vulnerable households. To this end, it is important to understand the properties, principles, and processes that strengthen resilience at the individual,

household, and community levels.

Disaster management efforts have concentrated largely on preparedness and post

-disaster responses. Prevention and mitigation projects are fewer. Placed in the proper

context, resources and properties that can be utilized in the face of a stress or shock are key determinants of exposure, sensitivity and adaptive capacity. According to the

Department for International Development (DFID 2011), resilience-enhancing activities can be usefully classified using the ‗assets pentagon‘ of the sustainable

livelihoods framework. The categories include the following aspects: (1) social/

human; (2) physical/technological; (3) financial/economic; (4) political and (5) environmental/natural and intervention can be done at the global/regional, national,

municipal/local and community/household levels. The projects and programs to

enhance resiliency can range from rural livelihood support to regional disaster insurance mechanisms and from pre-disaster household asset protection to housing

upgrades. These also include the following programs: (1) productive safety net program to protect the asset base of households in times of natural disasters and other

forms of shock; (2) early warning systems, village infrastructure for flood prevention,

cyclone shelters, and climate-resilient crops for strengthening resilience to climate change; (3) risk capacity program to provide participating countries with effective

financial tools and funds to effectively cope to extreme weather events; and (4)

support for the disaster risk reduction in school recovery programs like awareness and

response to natural disasters.


Conceptual Framework for Assessing Resilience

This paper adopts the United Kingdom Department for International

Development‘s (DFID 2011, p. 6) definition of disaster resilience as ―the ability of countries, communities, and households to manage change, by maintaining or

transforming living standards in the face of shocks or stresses – such as earthquakes,

drought or violent conflict – without compromising their long-term prospects.‖

Resilience is often described as the opposite of vulnerability. Both terms can be

used to describe the ability of individuals, households or communities to deal with

stresses or shocks (Bahadur et al. 2010). In natural ecosystems, the concept of resilience covers two separate processes: resistance—the magnitude of disturbance

that causes a change in structure and recovery—the speed of return to the original

structure (Darling and Cote 2010). In line with this, to minimize food insecurity

occurrences among households during natural disasters, it is of value to have a

measure of the household resistance and recovery potential. However, resilience analysis should not be seen as an alternative to vulnerability analysis, but rather as a

complement. Vulnerability analysis tends to measure only the susceptibility of people to damage when exposed to particular hazards or shocks. Often, it focuses on one

specific target variable, usually represented by the household consumption

expenditure (FAO 2010).

Building resilience reduces vulnerabilities of households at risk of natural

disasters, consequent shocks and stresses. DFID (2011, p. 15) listed the principles for

enhancing disaster resilience. Accordingly, resilience-building activities should ―(1) be anchored in national and local actors‘ realities and contexts; (2) be shaped by local

understanding and priorities; (3) be owned at country level; (4) be iterative and flexible, with regular adaptations, revisions, and check-backs; (5) understand and plan

for the fact that women, children, older and disabled people, and politically

marginalized groups are disproportionally impacted; (6) take multi-sectoral, multi-disciplinary approaches that bring together development and humanitarian

efforts and that establish common ground between climate change adaptation, social protection, disaster risk reduction, and work in fragile states; (7) be long-term and

collaborative, building on local relations and new partnerships; (8) be consistent with

international and national commitments; and (9) ensure that overall, the intervention/

response does not undermine resilience.‖

A conceptual framework for resilience assessment that considers these different

elements is shown in Figure 1. The onset of shocks may be rapid or slow as in

earthquakes or drought. It may also be longer-term in nature as in environmental

degradation. Usually, it is easier to mobilize resources to address rapid shocks rather than slow shocks. It is also important to understand that some shocks may affect only

some individuals or households while others may affect whole populations

(Frankenberger et al. 2012).


Figure 1. Resilience assessment framework (modified from Razafindrabe 2009; DFID 2011;

Frankenberger 2012)

Exposure is a function of the magnitude, frequency and duration of shocks while

―adaptive capacity is the nature and extent of access to and use of resources in order to deal with disturbance‖ (Frankenberger et al. 2012, p. 5). Adaptive capacity has

three interrelated elements, namely, livelihood assets; transforming structures and processes; and livelihood strategies. The tangible and intangible livelihood assets

would provide for the basic needs of individuals and households. These include the

financial, physical, political, human, social, and natural assets. The structures and processes, on the other hand, are ―embodied in the formal and informal institutions

that enable or inhibit the resilience of individuals, households and communities‖. The ―distinct or combined strategies that individuals and households pursue to make a

living and cope with shocks‖ are referred to as the livelihood strategies

(Frankenberger et al. 2012, p. 5).

The degree to which an individual, household or community will be affected by

a given shock or stress is referred to as sensitivity. The greater the sensitivity, the

lower is the degree of resilience and vice versa. Individuals or households are in the resiliency pathways if they are able to use their adaptive capacity to manage the

shocks or stresses they are exposed to. Otherwise, if they were sensitive and unable to

manage these shocks and stresses, they would likely go down the vulnerability

pathway. The pathway that they go through would determine whether the individual,

household, or community would be able to meet their needs or objectives, referred to

as the livelihood outcomes (Frankenberger et al. 2012).

The resilience assessment framework is useful in a number of respects. First, it provides policy makers and practitioners a comprehensive understanding of the

factors and processes influencing vulnerability and resilience at the household and

community levels. Second, it is useful for identifying ―gaps in key livelihood assets, the functioning of structures and processes of key institutions, and the livelihood

strategies of vulnerable households.


5 Classification of cities/municipalities based on average annual income derived from recurrent revenues/receipts is

mandated by the Department of Finance. The purpose is to determine the capability of local government units to

provide funding for developmental projects in their locality. 6

LAKEHEAD is the acronym of the research project ―Managing environmental risks to food and health security in

Southeast Asian watershed‖ that was funded by the Research Institute for Humanity and Nature, Kyoto, Japan.

Lastly, the ―extent and nature of community and household responses to shocks and stresses will result either in increased vulnerability or increased adaptive capacity and

resilience over time‖ (Frankenberger et al. 2012, p. 3).

Research Methodology

The analysis of this study has two parts: the determination of the household

resiliency index using a number of parameters and the evaluation of the link between

household resiliency and the levels of food security.

Study Site

The study was conducted in communities located at the Sta. Rosa-Silang Subwatershed, which is one of the 24 basins surrounding the Laguna de Bay. The

subwatershed has an area of about 120 square kilometers which comprises 4% of the total Laguna Lake Basin. It is one of the four elongated basins emanating from the

Tagaytay ridge and draining toward Laguna Lake. By political jurisdisction, Sta.

Rosa-Silang Subwatershed consists of Sta. Rosa City (37%), Cabuyao City (25%), Silang (22%) and Biñan City (16%). It covers a total of 54 barangays (the smallest

administrative unit in the Philippines). Sta. Rosa and Biñan are first class cities based on their income. Cabayao and Silang are second and fifth class municipalities,

respectively5. The natural vegetation, open areas, and farm lands account for 66% of

the watershed, while the rest (34%) consists of built-up areas consisting of residential, commercial, institutional, and large industrial areas. In 2007, the population of the

subwatershed was estimated at 569,199 or 113,839 households. In 2012, poverty

incidence in Laguna was 5.8%, one the lowest in the country (NSCB 2012). In 2010,

average annual employment rate in Laguna was 91.8% (NSO 2010).

The flood hazard map produced by the DENR-MGB (2010) shows that of 56 barangays included in the study, 40 were susceptible to flooding. These barangays are

located in the lowland to mid-lowland portion of the Sta. Rosa-Silang subwatershed

and comprise the majority of the cities of Sta. Rosa and Biñan and the town of Cabuyao. Four barangays in Sta. Rosa City were highly to moderately susceptible to

flooding. This means that in the event of prolonged and extensive heavy rains or extreme weather conditions, these barangays could immediately experience flooding

of more than 1 meter in height. On the other hand, 6 barangays, all located in Sta.

Rosa City, were moderately prone to flooding that could reach 0.5 to 1 meter in height during extreme weather conditions. Meanwhile, the rest of the barangays have

predominantly low susceptibility to flooding.

Dataset

The socioeconomic dataset used to estimate household resilience to flood

disasters was obtained from the results of the households survey in Sta. Rosa-Silang

Subwatershed conducted by the LAKEHEAD6 Project in 2012 (Figure 2).


Figure 2. Map of the Silang-Sta. Rosa Subwatershed (showing the barangays ― marked by black dots ―

where the samples were taken)

A follow-up survey on the same respondent-households used a structured

questionnaire to obtain necessary data that were not covered by the LAKEHEAD Project survey. Among these are household assets, access to public utilities and

services, social safety nets, flood disaster awareness, preparedness and experiences, and the impacts of flood on housing, household assets, health, livelihood, public

utilities, and infrastructure. The follow-up survey involved a total of 178 randomly

selected household-respondents (level of confidence = 95%; variability = .5; precision level = ±10%). While a larger number of samples or cases would have been

ideal to test more variables in the model, the current number of cases satisfies the

minimum requirement (i.e., cases-to-variable ratio) for the analyses.

The respondents are distributed in the seven barangays that were taken to

represent the 40 barangays within the subwatershed that were found to be susceptible

to flooding (Table 1). Around half (50.6%) of the respondents are settled along the

shore of Laguna Lake which is considered as the lowland portion of the subwatershed

(5-10 meters above sea level or masl) while around a quarter (25.3%) of the respondents were located in the upper lowland (11-50 masl). The rest occupied the

midland portion of the subwatershed (51-300 masl). The grouping of the location based on elevation (i.e., masl) followed the one used by the LAKEHEAD Project and

was relevant in identifying areas susceptible to flooding. These areas in the Laguna

Lake region were among the areas in the country hardest hit by Typhoon Ondoy

(Ketsana) in 2009.


Table 1. Sample size and distribution, Sta. Rosa-Silang Subwatershed, 2012

Estimating the Household Resilience Index

The model used in this study considers household resilience as a multidimen-sional latent variable determined by the socio-demographic and economic attributes of

the household. The list of relevant variables used in the estimation of the household

resilience index is presented in Table 2. The selection of indicators and variables was based on a survey of literature on resilience indicators, particularly the works of

Adger (2000), Folke (2006), Twigg (2007), Norris et al (2008), Razafindrabe et al. (2009), Alinovi et al (2009), and Aldrich (2012). Typhoon Ondoy (Ketsana), which

hit the country in 2009 causing a flood discharge that exceeded a 100-year flood

event, was made a common reference in describing household flood experiences.

Table 2. List of domains and indicators used in the analysis of household

resilience, Sta. Rosa-Silang Subwatershed, 2012

Barangay/Town/Province

Total No. of

Households

(As of May 2010)

Total No. of

Household

Respondents

Marinig, Cabuyao, Laguna 7,434 16

Gulod, Cabuyao, Laguna 1,883 23

Aplaya, Sta. Rosa, Laguna 2,834 19

Caingin, Sta. Rosa, Laguna 3,722 14

Sinalhan, Sta. Rosa, Laguna 3,816 18

Banaybanay, Cabuyao, Laguna 4,387 45

Sto. Tomas, Biñan, Laguna 7,798 43

Total 31,874 178

Variable Unit of Measurement

Household asset

Subsistence income

Employment ratio

Workers in the household

Disaster preparedness drills attended

Disaster preparedness seminars attended

Farm lot owned

Housing condition

Disaster relief goods availed of after the

disaster

Household asset index

Monthly household income per capita minus

official monthly subsistence income per

capita threshold (local currency)

Ratio (total number of employed household

members divided by household size)

Count

Count

Count

Number of hectares

Ratio (total walled-in floor area divided by

household size)

Value in local currency


As a measure of the overall living standard, the study used four types of indices pertaining to the built, social, economic, and institutional environments of the

community. Each of these indices represents the weighted mean score that was

obtained by the community with respect to a set of 4-5 parameters taken to indicate the state of a given environment especially with regard to managing a disaster. Based

on their current priorities, local government officials assigned the weight for each

indicator. The various parameters that were used in constructing the indices are outlined in Table 3. The data on household exposure, livelihood assets and quality of

living standards were obtained from parallel studies conducted under the

LAKEHEAD Project and for this study.

Table 3. List of indices and corresponding indicators used in constructing the

overall community disaster resilience index, Santa Rosa-Silang

Subwatershed, 2012

Damage to housing structure

Damage to household assets

Damage to productive assets

Loss in farm income

Loss in non-farm income

Cost of health treatment

Flood height in immediate proximity

Flood duration in immediate proximity

Duration of stay in evacuation center

Repair cost in local currency

Replacement cost in local currency

Replacement cost in local currency




Maximum flood height in feet

Number of days

Number of days

Built-up environment

Social environment

Economic environment

Institutional environment

Weighted mean score, 1-5




Index Indicator

Physical Electricity (availability, access, supply and alternative capacity)

Water (availability, access, supply and alternative capacity)

Sanitation and solid waste disposal (access to sanitation, collection,

recycling and treatment of solid waste after a disaster)

Road network (availability, access, drainage network)

Housing and land use (ownership, compliance with building code,

housing materials, plinth level, safety, built areas, ―green space‖)

Social Population (informal settlers)

Health (availability, access and quality of health facilities/services,

prevalence of waterborne and vector borne diseases after a disaster)

Education and awareness (literacy rate, availability of disaster

awareness program and preparedness drills, access to internet)

Social capital (participation of residents and ethnic groups in

community activities, participation in decision-making processes)

Community disaster preparedness (voluntary contribution and

participation in relief efforts among residents, support from NGOs/POs,

voluntary evacuation)


The study used the principal components analysis (PCA) to measure the whole

system. PCA is a variable reduction method used to identify groups of observed

variables that tend to converge empirically. In mathematical terms, from an initial set of n correlated variables, PCA creates uncorrelated indices or components, where

each component is a linear weighted combination of the initial variables (Vyas and

Kumaranayake 2006). PCA produced two key results that are important in understanding and measuring resilience. First, PCA enabled the study to interpret

factor structures by yielding factors that are linearly related to each of the principal components. Factor loading, which can be interpreted as a standardized regression

coefficient, readily indicates the strength of this relationship. Hence, it was possible to

interpret factor structure by considering prior knowledge that could be responsible for

the observed pattern of positive and negative loadings.

Economic Income (number of income sources, income from

employment in informal sector, households experiencing

reduced income after a disaster)

Employment (informal sector, employment among women;

women employed in formal sector, prevalence of child

labor)

Household assets (ownership of appliances, motorized/non-

motorized vehicle)

Finance and savings (availability, access and quality of

credit facilities, capacity to save, insurance coverage)

Budget and subsidy (availability and sufficiency of

barangay fund for Disaster Risk Management or DRM,

availability of barangay subsidies/incentives for house

reconstruction, livelihood and health care services after a

disaster)

Institutional Disaster management (existence and effectiveness of

―emergency response teams and volunteer groups, existence

and effectiveness of early warning system and disaster

drills, capacity of evacuation centers)

Mainstreaming of Disaster Risk Reduction (DRR ) and

Climate Change Adaptation (CCA) (technical and logistical

capacity to formulate barangay development plans,

integration of CCA and DRR in barangay development

plans, use of hazard maps in planning, community

participation in barangay development planning processes)

Knowledge dissemination and management (availability of

disaster training programs for emergency workers,

effectiveness to learn from previous disasters, Information

Extension and Communication materials ( IEC) for disaster

awareness and preparedness, community satisfaction with

disaster awareness programs)

Institutional collaboration with other organizations and

stakeholders (dependence on external support,

collaboration between and among different barangays

councils, cooperation among local officials/leaders,

collaboration with NGOs )


Second, PCA commonly provides for the factor scores and the proportion of the variance in the principal component that is explained by each factor. Hence, it was

possible to estimate the household resilience index as the weighted sum of the factor scores each multiplied with its own proportion of the variance explained. Formally,

the household resilience index, HRI, is given by:

(1)

where x is the factor score, and w is the weight.

The limitation of this approach becomes evident when the resilience indices are

to be compared among different sets of samples, i.e., different locales or social groups. Nonetheless, the role of the resilience index is crucial to the evaluation of the

conditions of a target population.

Linking Household Resilience to Food Security

A logistic regression model was constructed to evaluate the significance of the

estimated household resilience index as a measure of food security. The aim is to construct a parsimonious model, from which variables previously used to estimate

household resilience are excluded. To do this, the following logistic regression model

used:

(2)

where Y is the dichotomous outcome of interest (1 = food secure, 0 = food

insecure); π is the probability of the event; α is the Y intercept; β’s are regression coefficients; HRI is the household resilience index; DGen is a dummy variable for the

gender of the household head (1 = female, 0 = male); DSrc is a dummy variable for primary source of income (1 = non-agriculture, 0 = agriculture); and DLoc is a

dummy variable for location of residence (1 = lowland, 0 = upper lowland/midland).

The α and β’s are typically estimated by the maximum likelihood (ML) method, which is designed to maximize the likelihood of reproducing the data given the

parameter estimates. The marginal effects after logit was then computed to determine the percentage increase in the likelihood that a household will be food secure for

every unit increase in the predictor variable (e.g., HRI) with all other factors held

constant.


Profile of the Respondents

The results show that a large majority (76.4%) of the household heads are males. The average household size was five with 38.8% of the households composed of 3-4

members. A large majority (73.0%) live in their own home lot while 15.2% live in lots rent-free either with or without the owner‘s consent. The average floor area of a

dwelling unit is about 84 square meters.


Table 4. Profile of the respondents, 2012

The mean worker-to-dependent ratio is 0.45, which means that a single worker

was supporting two non-working individuals in the household. Only 10.1% and

11.8% of the households were engaged in farming and fishing, respectively, while the

large majority (78.8%) drew their income mainly from non-agriculture related work.

About a quarter (24.7%) of the households has a monthly per capita income that

is below the food subsistence threshold of PhP1,012 estimated by the National

Statistical Coordination Board (2009) for Region IV-A; these households are considered subsistence poor. But considering other factors apart from income (e.g.,

anxiety over the quantity and quality of food in the household, incidence of food substitution and/or reduction), the survey data show that a substantial majority (82%)

of the households may be considered as food insecure.

Variable Result

(n = 179) Variable

Result

(n = 179)

Place of residence (%) Mean floor area (sq. m.) 84

Sta. Rosa City 28.7 Mean employment ratio 0.45

Biñan City 24.1 Source of income (%)

Cabuyao City 47.2 Non-agriculture 78.1

Location (%) Farming 10.1

Lowland 50.6 Fishing 11.8

Upper lowland 25.3 Monthly per capita income

(%)

Midland 24.2 PhP 1,000 or less 24.7

Gender of household head (%) PhP 1,000 – P2,000 24.7

Female 23.6 PhP 2,001 – P3,000 15.2

Male 76.4 PhP 3,001 or more 35.4

Household size (%) Mean (PhP) 3,375

3 members or less 24.2 Food security (%)

4-5 members 38.8 Food secure 18.0

6-7 members 29.2 Food insecure 82.0

8 members or more 7.9 Experienced flooding TY

Ondoy (%)

Mean (count) 5 Yes 53.4

Ownership of home lot (%) No 46.6

Owned 73.0 Evacuated due to TY Ondoy

(%)

Rented 11.8 Yes 22.5

Rent free with owner‘s

consent

13.5 No 77.5

Rent free without consent 1.7

Note: Percentages may not total 100 due to rounding off


In terms of flood experience, slightly over half (53.4%) of the households have actually experienced flooding first hand, specifically in 2009, when Typhoon Ondoy

(Ketsana) devastated many parts of the area. Around 22.5% of the households were

reportedly forced to evacuate their homes.

Principal Component Analysis

The results of the PCA revealed that only three components satisfied the eigenvalue-1 latent root criterion. Thus, only the first three components were retained

for rotation. The corresponding variables for each of the components and their

respective factor loadings are presented in Table 5. Combined, components 1, 2 and 3 accounted for 85% of the total, satisfying the cumulative proportion of total variance

criterion of 60%.

In interpreting the rotated factor pattern, a variable was said to load on a given

component if its factor loading was 0.40 or greater for that component, and was less

than 0.40 for the other. Using these criteria, four out of the original 29 variables subjected to PCA were found to load on component 1, which was consequently

labeled flood exposure component. These variables were repair cost of flood damages to housing structure, flood height, flood duration and duration of stay in evacuation

center. Noting that these variables are in the reverse scale, the positive signs suggest

that reduction of exposure to flood hazards contributes to the increase in household resilience as reduction of exposure to flood hazards lessens the risk of asset loss and

displacement.

Table 5. Rotated factor pattern and final communality estimates from PCA,

Sta. Rosa-Silang Subwatershed, 2012

Note: Variables for repair cost of damage to housing, flood height, flood duration, duration of stay in evacuation

center, and built-up environment index are in the reverse scale.

Two variables, the economic environment index and the built-up environment

index, were found to load on component 2, which was labeled the community living standards component. Noting that the built-up environment index is also in the

reverse scale, its positive sign appears to corroborate the hypothesis that increasingly

urbanizing areas face greater risks from flooding because of increasing population

density and increasing housing stock, among others.

Variable Communality

Estimates

Component

1 2 3

Cost of flood damage to housing .868 .926 .045 .092

Flood height .881 .925 .150 .054

Flood duration .895 .924 .202 .031

Duration of stay in evacuation center .644 .775 .200 .053

Built-up environment index .924 .206 .938 -.035

Economic environment index .923 .170 .945 -.032

Employment ratio .867 .095 -.027 .926

No. of workers in the household .868 .046 -.032 .930


On the other hand, the positive sign of the economic environment index suggests that increasing employment opportunities in the community contributes to increasing

household resilience, as it would allow them to diversify sources of income. In the event of a disaster, more sources of income means less risk of livelihood

displacement.

Finally, two variables - the household employment ratio and the number of workers in the household - were also found to load on the component 3, which was

labeled the household economic capability component. While the community may

provide employment opportunities, it is the economic capability of households that ultimately translates these opportunities into livelihood outcomes. Enhancing the

economic capability of households to avail of these opportunities may enable them to become more flexible and diverse in terms of income sources, which is an important

coping strategy in times of crisis.

Linking Household Flood Disaster Resilience to Food Security

The results of the logistic regression show that the model is statistically

significant (Prob> chi2 = .024), indicating that at least one of the regression coefficients in the model is not equal to zero or has an effect on the response variable

(Table 6).

Table 6. Estimated coefficients and marginal effects from the logistic regression,

Sta. Rosa-Silang Subwatershed, 2012

Notes: Standard errors are in parentheses; ** significant at 5%; *** significant at 1%; ns means not significant;

(a) dy/dx is for discrete change of dummy variable from 0 to 1.

The results also show that resilience was the most important factor (p< .01) and

therefore, a key indicator in household food security analysis. Further, the results of

computing for the marginal effects after logit indicated that for each point increase in the level of household resilience, the likelihood of being food secure increases by

17.2%. The dummy variable for location of residence was also found to significantly

predict household food security (p< .05).

Variable Coefficient Marginal Effect

Household resilience index 1.350*** .172***

Dummy for location of residencea

(1 = lowland, 0 = upper lowland/midland)

1.018**

(.494)

.130**

(.062)

Dummy for gender of household heada

(1 = female, 0 = male)

.700ns

(.460)

.101ns

(.073)

Dummy for primary source of incomea

(1 = non-agriculture, 0 = agriculture)

.566ns

(.566)

.064ns

(.057)

Constant -2.865***

(.666)

-

No. of observations 169

LR chi2(4) 11.24

Prob> chi2 0.0240

Pseudo R2 0.0725


The computed value for the marginal effect indicates that a discrete change in the place of residence, that is, from upper downstream/midstream to downstream

increases the likelihood of household food security by 13%. But this should be interpreted with caution for two reasons. First, while upper lowland and midland

communities are less exposed to flooding and relatively more progressive in terms of

built-up and economic environments, households in these areas have significantly lower economic capability to actually benefit from these opportunities as compared to

those in lowland areas. The other is that, compared to their counterparts in the upper

lowland/midland areas, it was observed that most sample households in lowland areas reside in relative close proximity with the central market or business districts and may

have more access to livelihood opportunities. Hence, location alone does not strictly contribute to resilience, but the opportunities that these locations offer along with the

capability of the household to actually avail of and translate these opportunities into

livelihood outcomes could contribute to resilience.

Conclusion

The results show that household resilience is multi-dimensional. In the context of the households in the Sta. Rosa-Silang Subwatershed, resilience is a function of

households‘ level of exposure to natural hazards, their economic capability and the

community standard of living. Exposure refers to the nature of the hazard affecting the household (e.g., flood height, flood duration) and its immediate impacts (e.g.,

property damage, displacement from home). Economic capability, on the other hand,

refers to the quantity of resources that are actually and potentially available to the household (e.g., number of household labor) as well as the mobility of these resources

(e.g., household employment ratio) especially in coping with a disaster. It is reflective of the adaptive capacity of households to cope with the adverse effects of natural

hazards. Community living standard in turn describes the overall capacity of the

natural, built, social, economic and institutional environments of the community to

support the households in the community.

Results show that the economic capability of households and the conditions of the economic and built environments of the community have a significant influence

on household resiliency. The economic environment, for example, may enhance

household resilience by providing opportunities to diversify sources of income but at the same time, the built environment may pose as a constraint by increasing exposure

because of unregulated land use that allows human settlements in areas highly

susceptible to flooding. The results of the study provide empirical evidence that community conditions directly affect the level of household resilience, either

positively or negatively.

Lastly, the study demonstrates that resilience directly affects household food

security. In fact, the results show that the estimated household resilience index plays a

significant and prominent role in predicting household food security. This suggests that building household resilience could be an important component of any strategy to

attain household food security.

But there are limitations inherent to the approach taken in this study. First, the indices are apparently dependent on the quality of the data obtained in the surveys

and maps.


They are not exhaustive; hence they are only indicative and not readily comparable to different sets of samples. The robustness of these indices may still be improved as

more data (e.g., higher resolution hazard maps, population data) become more available. Second is a limitation related to the time scale of the analysis. By focusing

on the current capacities of both the household and the community, social change is

held constant in this study. This could be addressed in subsequent studies by

incorporating future scenarios of change in the model (e.g., sensitivity analysis).

All things considered, the study demonstrates the usefulness of resiliency

assessments to researchers, local policy makers, and other practitioners in the field of disaster risk reduction seeking to evaluate the conditions of a target population for

possible intervention.

References

Adger, W.N. 2000. ―Social and ecological resilience: are they related?‖ Progress in

Human Geography 24(3): 347–364.

Alinovi, L, E. Mane and D. Romano . 2009. ―Measuring household resilience to food

insecurity: application to Palestinian households.‖ Working Paper. EC-FAO Food Security Programme. Linking Information and Decision Making to

Improve Food Security. www.foodsec.org

Aldrich, D.P. 2012. Building Resilience: Social Capital in Post-Disaster Recovery.

Chicago and London: University of Chicago Press. London.

Bahadur A, M. Ibrahim and T. Tanner. 2010. ―The Resilience Renaissance?

Strengthening Climate Resilience‖ Discussion Paper 1. 2010. http://ipcc-

wg2.gov/SREX/report/njlite?chapter=&page=3.Cited 1 February 2013.

Carcamo, D. 2013. 3 Laguna Towns Placed Under State of Calamity. Aug. 19, 2013

Philstar Nation. http://www.philstar.com

Center for Research on the Epidemiology of Disasters (CRED) . 2013. International

Disaster Database.http://www.emdat.be/result-country-profile. Cited 19 January

2013.

Darling, E. and I. Cote . 2010. Rethinking Ecosystem Resilience in the Face of Climate Change. PLoS, 1-5.http://www.mendeley.com/catalog/rethinking-

ecosystem-resilience-face-climate-change-10/. Cited 31 January 2013.

Department for International Development (DFID). 2011. Defining Disaster Resilience: A DFID Approach Paper. DFID 2011. UK Department for

International Development: London.

Food and Agriculture Organization of the United Nations (FAO). 2010. Measuring Resilience: A Concept Note on the Resilience Tool. FAO. http://www.fao.org/

docrep/013/al920e/al920e00.pdf. Cited 30 January 2013.

Frankenberger, T., T. Spangler, S. Nelson and M. Langworthy. 2012. ―Enhancing

Resilience to Food Security Shocks in Africa.‖ Discussion Paper, (November),

http://www.fsnnetwork.org/sites/default/files

discussion_paper_usaid_dfid_wb_nov._8_2012.pdf . Cited 31 January 2013.


Global Footprint Network (GFN). 2013. Ecological Footprint Report: Restoring

Balance in Laguna Lake Region. Oakland, CA: Global Footprint Network.

Norris, F., S. Stevens, B. Pfefferbaum, K. Wyche and R. Pfefferbaum 2008. ―Community Resilience as a Metaphor, Theory, Set of Capacities, and Strategy

for Disaster Readiness.‖ American Journal of Community Psychology, 41(1),

127-150.

National Statistical Coordination Board. 2012. First Semester Per Capita Poverty

Threshold and Poverty Incidence among Families, by Region and Province:

2006, 2009 and 2012. http://www.nscb.gov.ph/poverty/defaultnew.asp Cited 1

May 2013.

National Statistics Office (NSO). 2010. Provincial Estimates of the Employment

Rates and Labor Force Participation Rates. https://psa.gov.ph/old/data/

sectordata/datalfs.html. Cited 1 May 2013.

Razafindrabe, B., G. Parvin, A. Surjan, Y. Takeuchi and R. Shaw. 2009. ―Climate disaster resilience: focus on coastal urban cities in Asia.‖ Asian Journal of

Environment and Disaster Management (AJEDM), 1 (1):101-116.

Twigg, J. 2007. Characteristics of a Disaster-resilient Community: A Guidance Note

Version 1. DFID Disaster Risk Reduction Interagency Coordination Group.

UK Department for International Development (DFID). 2011.―Defining Disaster Resilience: A DFID Approach Paper.‖.http://www.dfid.gov.uk/Documents/

publications1/Defining-Disaster-Resilience-DFID-Approach-Paper.pdf. Cited 16

January 2013.

Vyas, S. and L. Kumaranayake. 2006. ―Constructing socio-economic status indices:

how to use principal components analysis.‖ Health Policy and Planning, 21

(6):459-68.

World Bank (WB). 2011. The social impacts of tropical storm Ondoy and typhoon

Pepeng - The recovery of communities in Metro Manila and Luzon. Washington, DC:WorldBank.http://documents.worldbank.org/curateden/2011/12/16279435/

social-impacts-tropical-storm-ondoy-typhoon-pepeng-recovery-communities-metro-manila-luzon-social-impacts-tropical-storm-ondoy-typhoon-pepeng-

recovery-communities-metro-manila-luzon . Cited 12 March 2013.



Email: [email protected] 2 College of Economics and Management, University of the Philippines Los Baños

Email: [email protected] 3 College of Economics and Management, University of the Philippines Los Baños

Email: [email protected] 4 BS Agricultural Economics Graduate, University of the Philippines Los Baños

Email: [email protected]

Abstract

The objective of this study is to provide a better understanding of the snail dredging

industry and its current situation. Specifically, it describes the snail dredging activities

done in selected areas in Laguna Lake, determines the profitability of snail dredging, and

identifies the market destinations of the snails dredged. Snail dredging has been banned

because of its adverse effects on the lake environment. However, the study finds that snail

dredgers continue to operate all year round. On a daily basis, dredging operations range

from 4 to 10 hours. Snails dredged from Laguna Lake are brought to the provinces of

Pampanga (41%), Bulacan (13%), Quezon (9%) and Laguna (37%). Gross margin

analysis was employed to determine the profitability of the snail dredging business. With a

GM/GS ratio of 0.15 for both Rizal and Laguna dredgers and 0.14 for Pasig dredgers, snail

dredging is not a lucrative business that can provide the snail dredgers a high income.

Despite the prohibitions and low profitability, snail dredgers are likely to remain in the

business due to the following reasons: (1) the operators and hired workers lack alternative

high paying jobs and sure source of income, (2) there is high demand for snails from duck

and prawn farms, and (3) high investment cost. Further research is needed to determine

how snails can be profitably grown without aggravating the lake‘s current condition.

Keywords: snail dredging, Laguna Lake, gross margin analysis

Introduction

Open fishery production in Laguna Lake consists mainly of fish, shrimps and

snails. The lake‘s relative shallow depth, muddy smooth bottom and its almost perennial turbidity make it ideal for the growth and proliferation of snails. Palma,

Mercene and Goss (2005) identify ten species of snails in Laguna Lake. Some of the

common species found in the lake are Vivipara angularis, Thiara sp. and Simpsonella subcrassa. In 1963, snail was the biggest aquatic output of Laguna Lake,

recorded at 247.8 metric tons (MT) or 71% of total production of 349.7 MT

(Delmendo 1966).

Snails gathered from the lake are mainly used as feeds for ducks. According to

duck raisers, snails are cheap and serve as the main source of protein for ducks that is necessary for forming richer egg yolks and thicker egg shells (Atienza et al. 2013).

The demand of the duck farms for freshwater snails gave rise to the snail dredging

business in areas around Laguna Lake.

Snails can be gathered from the lake bottom with the dredging gear, locally

called kaladkad or pangahig. The gear is made up of ropes, nets, wood and steel. The net or collecting bag is attached to a rectangular wood frame with the lower bar made

of steel. The weight of the steel frame and the force from dragging the frame as the

boat tows the gears enable the frames to dig a few inches below the lake bottom to get

the snails.

Operations and Profitability of Snail Dredging in Laguna, Rizal and Pasig City

Ma. Eden S. Piadozo1, Roberto F. Rañola Jr.2, Ma. Joy N. Malabayabas3 and

Dominic M. Hamada4

70 Piadozo, Rañola, Malabayabas and Hamada

The snail dredging operation in the lake flourished together with the success of the duck raising industry in the 1940s, specifically from 1945 to 1947. The expansion

of the snail dredging operation resulted in competition and conflicts among commercial fishermen because the dredging gears could damage the fishing gears of

other fishermen. To resolve this conflict, the Fisheries Administrative Order No. 30

under Republic Act No. 177 and Act No. 4003 or the Fisheries Act were enacted in April 1952. The use of dredging gears was not allowed during a particular period of

time − from 5 pm to 5 am the following day. Dredgers were also banned from

gathering snails within 200 meters of any fish corrals (baklad) authorized by the

municipality.

On January 27, 1955, another administrative order (Fisheries Administrative Order No. 41) was enacted to regulate the operation of several fishing gears including

dredging gears. The new administrative order was implemented for the protection,

promotion and conservation of municipal fisheries in the Laguna Lake area. It prohibited snail dredging within any area of Laguna Lake which is occupied by

lungga sa biya (a stationary underwater fish shelter or trap made of either curved-out adobe rock, pre-cast cement, empty trench mortar shell, or one to two-joint

bamboo). However, licensed snail dredgers can operate 50 meters away from

established fish traps. Possible penalties for the violation of the said provisions were the following: (1) fine of not more than two hundred pesos, (2) imprisonment not

exceeding six months, or (3) both fine and imprisonment.

Despite the prohibitions and penalties, snail dredging activities remained rampant in the 1960s (Ecological Footprint Report 2013). The activity continued

unabated even in the 1980s and the 1990s when truckloads of snails from the lake were being harvested to meet the requirements not only of duck farms but also prawn

farms in the provinces of Quezon and Bulacan (Santos-Borja and Nepomuceno 2006).

This resulted in a significant decline in the lake‘s snail population. The reduction in the snail population was the major reason for the decline in the duck population

around the lake. According to the data from CountrySTAT (2015), the population of ducks in commercial and backyard farms around Rizal and Laguna decreased from

444,544 in 1994 to 185,448 in 2014.

The extensive snail dredging activities in the lake also aggravated the existing conflicts between the dredgers and other fisher folk. The snail dredging operations not

only damaged the fishing gears of other fishermen but also destroyed the lakebed.

These gears not only caught the snails but also all forms of aquatic life that lay along their path. Republic Act 8550 (Fishery Code of 1998) was enacted to address these

conflicts. Under this law, dredging gears are reclassified as active gears and the use of these gears is prohibited in municipal waters and in all bays as well as other fishery

management areas. As defined in RA 8550, an active gear is a fishing device that is

characterized by gear movements and/or getting the target species by towing, pushing gears, surrounding, pumping, dredging and scaring the target species to

impoundments. Anyone caught violating the said prohibition shall be meted the

penalty of 2 to 6 years imprisonment and be made to pay a fine amounting to

PhP2,000 to PhP20,000 upon the decision of the court.


The ban on dredging gears in the lake and the strict penalties, however, was not able to completely stop the dredging operations in the lake. There still remain some

operators in the lake, which, according to officials, are contributing to the degradation of the lake. Various studies had mentioned the adverse effects of

dredging that include the decline of fish catch, the discharge of poisonous odor and

the release in overlaying water of heavy metals present in the lake.The toxic and non-toxic sediments at the bottom of the lake, when disturbed, may cause danger to the

health of people (Colting-Pulumbarit and Lapitan 2009).

Methodology

The study was conducted in 2013. To meet the objectives of the study a

complete list of snail dredgers in each shoreline municipality of the provinces of

Rizal and Laguna was requested from either the concerned Fisheries and Aquatic

Resources Management Council (FARMC) or the municipal agriculturist. There is

no list available from such sources because snail dredging has been considered an illegal activity in the lake. Only 14 respondents (snail dredger operators) were

interviewed. These were the snail dredgers identified by key informants in the study areas. The study areas were Victoria and Sta. Cruz in Laguna, Binangonan in Rizal

(Talim Island), and Pasig City (Figure 1).

Figure 1.Map of the study sites around Laguna Lake, 2012


Primary data were gathered using a pre-tested questionnaire. Data collected included the following: socio-economic profile; snail dredging practices; volume of

snails dredged and sold; marketing practices; capital investments; and the costs and

revenue from snail dredging.

To determine the profitability of snail dredging business, the gross margin

analysis was used. The gross margin represents the percentage of total sales revenue that a snail dredger retains after incurring the direct costs associated with the

operation:

Gross Margin = Net revenue / Gross Revenue x 100

Gross margin is an important indicator of the financial health of the snail

dredging business. The gross margin of snail dredgers should be high enough to cover

costs and provide profits. Without an adequate gross margin, snail dredgers will be

unable to pay their expenses and continue their operations.

Results

Socio-economic Characteristics of Respondents

Fifty-eight percent of the snail dredgers interviewed came from the province of Laguna while 29% and 13% came from the province of Rizal and the City of Pasig,

respectively. The duration of their engagement in this activity ranges from one to 43

years (Table 1). All of respondents are owner-operators. The oldest owner-operator interviewed was 62 years old while the youngest was 28 years old; they are both from

Sta. Cruz, Laguna. The younger operators inherited the business from their parents.

Table 1. Socio-economic characteristics of respondents, Rizal, Laguna and

Pasig City, 2012

Characteristic

Rizal Laguna Pasig All

Locations (N=14)

Binangonan

(n=4)

Sta. Cruz

(n=4)

Victoria

(n=4)

Pinagbuhatan

(n=2)

Age (years)

Range 30-49 28-62 30-60 34-59 28-62

Average 36 48 44 47 43

Educational attainment (years)

Range 10-12 6-12 5-13 6-14 5-15

Average 11 9 10 11 10

Years in snail dredging

Range 1-32 2-40 2-30 10-43 1-43

Average 16 18 15 27 19

Dependence

Fully dependent 1 2 1 2 6

(%) 25 50 25 100 42.86

Not fully

dependent 3 2 3 - 8

(%) 75 50 75 - 57.14


Survey results show that majority (57%) of the operators are not fully dependent on snail dredging. Seventy-five percent of the owner-operators in Victoria, Laguna

and Binangonan, Rizal are only partially dependent on snail dredging for their livelihood. In contrast, all of the respondents from Pinagbuhatan are full-time snail

dredgers.

The prohibition of snail dredging in Laguna Lake in 1998 by virtue of RA 8550 has had significant impact on the snail dredging industry and consequently on the

duck industry. Most owner-operators reduced their snail dredging operations because

quite a number of them were arrested and their boats were confiscated and even destroyed. The reduction in the supply of snails greatly affected the duck farms

around the lake. In addition, there were typhoons (Ondoy and Pepeng in 2009) which brought high flood waters that devastated a lot of duck farms especially those in

Laguna. These series of events caused a decline in the duck industry and

consequently forced a lot of snail dredgers to stop their operations.

The slowdown in the snail dredging operations greatly affected the income of

the snail dredgers. They therefore turned to other sources of income such as duck raising, selling salted eggs, relative‘s remittance, furniture making/selling, boat rental,

apartment rental and tricycle driving. But quite a number of them continued to

operate their snail dredging operation.

Capital Investments

In snail dredging, operators invest in motorized boats, dredging gears

(kaladkad), vehicles, nets, ropes and pails (Table 2). The largest capital investment is on motorized boats which account for 30% to 90% of total investment. The highest

average investment on boats is found among operators in Rizal Province (PhP 383,333) followed by those in Pasig (PhP 350,000) and those in Laguna (PhP

128,438).

Motorized boats are used for snail dredging operations because they are fast, efficient and have the necessary power to drag the dredging gears from the bottom of

the lake. These boats are usually made of fiber glass, wood and bamboo poles and are bigger than the regular boats used for fishing. They are designed to carry heavy loads

that could reach thousands of kilograms of snails. The highest maximum holding

capacity of a motorized boat is 8,000 kilograms while the minimum is 500 kilograms.

Another important capital investment is the dredging gear (kaladkad or

pangahig). As mentioned earlier, the gear is a composite unit attached to the boats

and is made up of nets with fine mesh, ropes, wood and steel (Figure 2). Owner-operators usually own more than two gears that are used simultaneously. They

usually attach four individual units of dredging gears so that they can gather a high volume of snails in one operation. Snail dredgers in Laguna and Pasig have more

dredging gears than those in Rizal because majority of the owner-operators in Rizal

are only part-time snail dredgers.


Table 2. Average number and cost per unit of capital investments used in snail

dredging operation in Rizal, Laguna, and Pasig, 2012

Capital Investment

Rizal Laguna Pasig

Number

Owned

Average

Cost (PhP)

Number

Owned

Average

Cost (PhP)

Number

Owned

Average Cost

(PhP)

Boat

Motorized 1 383,333 1 128,438 2 350,000

Non-motorized 1 15,000 1 24,000 - -

Dredging gear

(Kaladkad) 5 633 6 705 15 625

Net (meters) 12 75 27 302 - -

Rope (meters) 11 222 54 294 59 82

Pail 2 53 4 64 3 50

Vehicles

Horse and wooden

cart (kalesa) - - 1 29,250 - -

Tricycle - - 1 50,000 - -

Van 150,000

Trucks - - - - 1 150,000

Total 399,316 383,053 500,757

Source of data: Survey data, 2012

Figure 2. Dredging gears used in snail dredging operations

Source: Santos-Borja(2013) and Hamada (2014)


Vehicles used for delivering the snails to their respective buyers vary according to location. One owner-operator in Victoria, Laguna uses a horse with a wooden cart

to deliver the snails to the nearby duck farms while the other operators use tricycles or mini-vans. Pasig dredgers, on the other hand, use a bigger vehicle (truck) in delivering

the snails to their market outlets, most of which are located outside Metro Manila.

Owner-operators also invest in pails and extra ropes and nets. Pails are used as temporary storage container of snails. It is used in transferring the snails to the sacks

before delivery. Most of the time, snails are sold on a per pail basis.

Overall, owner-operators in Pasig have the highest capital investment amounting to PhP500,757. They have a larger scale of operation which requires more dredging

gears, bigger boats and transport vehicle.

Weekly Operating Costs

On a weekly basis, owner-operators usually spend some amount on fuel, labor,

sacks, food, and cigarette of workers (Table 3). The labor cost accounts for 68%, 64%, and 53% of the operating costs in Rizal, Laguna and Pasig, respectively. The average

wage rate across all locations is PhP 300.00 per day. The average number of workers hired in Rizal is four while Laguna and Pasig hires an average of ¬ 3 and 10 workers,

respectively. The hired workers perform various tasks − from the actual snail dredging

operations to the hauling of snails. Owner operators usually hire boat drivers and workers to perform dredging operation and the gathering of snails. During snail dredg-

ing operations, owner-operators provide food and cigarettes to their workers. Food

provided varies from coffee and biscuits to rice and actual meals. Pasig snail dredgers also incur the highest cost for food and cigarette – they employ more workers com-

pared with Rizal and Laguna operators.

Table 3. Average weekly operating cost incurred in snail dredging business in

Rizal, Laguna and Pasig, 2012

Item Average Weekly Operating Cost (PhP) All

Locations Rizal Laguna Pasig

Fuel 1,988

(15%)

2,783

(24%)

5,650

(25%)

2,967

(22%)

Sacks 276

(2%)

635

(6%)

1,200

(5%)

538

(4%)

Labor cost 9,013

(68%)

7,315

(64%)

12,075

(53%)

8,480

(62%)

Other costs

Food 1,715

(13%)

653

(6%)

3,500

(15%)

1,540

(11%)

Cigarette 224

(2%)

- 336

(1%)

252

(2%)

Total 13,216 11,386 22,761 13,777

Source of data: Survey data, 2012

Note: Figures in parentheses are percent of total


The dredgers from Pasig incur the highest fuel costs (PhP 5,650) while those from Laguna have the lowest (PhP 2,783). Compared with the other two locations,

Laguna and Rizal, Pasig is farther away from the lake, thus Pasig operators use more

fuel in going to and moving from the dredging area.

Owner-operators also spend some amount on sacks. Snails are packed in sacks

before they are delivered to buyers. The price of sacks ranges from PhP 2.00 to PhP

5.00 per piece. Some owner-operators buy recycled sacks because they are cheaper.

Dredging Location

Survey results show that dredging is done in several locations. In general, operators in Rizal and Laguna dredge in the municipalities within their province while

the operators in Pasig gather snails in the areas of Rizal and Napindan (Figure 3). The

municipalities of Angono, Tanay and Cardona in the province of Rizal are the most

common dredging areas.

Figure 3. Dredging locations in Laguna, Rizal and Pasig, 2012

Snail dredgers in Laguna Lake consider two factors when choosing their location:

(1) municipal regulations, and (2) abundance of snails. While snail dredging is completely prohibited in Laguna Lake, there are municipalities that allow snail

dredgers to operate in their area as long as they do not damage other fish structures or

fishing gears.


According to a FARMC official, some municipalities in Laguna (e.g., Santa Cruz, Victoria, Lumban, Kalayaan, Siniloan and Mabitac) and Rizal (e.g., Binangonan,

Cardona and Tanay) allow snail dredging operations in their areas because they are aware of its importance to various duck farms that use snails for feeds. Other

municipalities (e.g., Bay and Biñan), on the other hand, strictly enforce the regulation

that prohibits snail dredging in the lake. Some dredgers found operating in the lake were apprehended by the Bantay Lawa (Lake guards) and were made to pay the fines

for the release of their boats and their crew members.

According to the respondents, snails are found abundantly in certain areas of Laguna Lake that are not heavily polluted. Some respondents in Laguna mentioned

that they gather snails as far away as Rizal because the catch in Laguna has declined. Snail dredgers in Sta. Cruz, Laguna attributed the lack of snails to the strong

earthquake which hit the municipality in 1990. They believe that the earthquake

destroyed snail fry within the lakebeds so that the productivity of snails declined in

the succeeding years.

The snail dredging industry has also been adversely affected by the proliferation of fish pens that have occupied large areas of the lake, resulting in the reduction of

areas that can be dredged. Some of the dredgers resort to paying guards/caretakers of

fish pen in order to be allowed to dredge inside the pens after the fish is harvested. Snail dredgers claim that there are plenty of good quality snails inside fish pens. They

believe that the feeds given to the fish in pens also contribute to the growth of snails

inside the structure.

Dredging Days and Dredging Hours

All the dredgers interviewed mentioned that snail dredging is an all-year-round operation. They engage in dredging everyday. However, they consider the rainy

months as lean months due to the difficulty of locating the snails. When there are

strong winds, snails are swept away by the current.

While all the dredgers operate daily, the dredging hours vary according to the

volume of snails required by their buyers. The buyers, usually traders and duck farm owners, would preorder the amount of snails they would require before these are

dredged. Snail dredgers start their operation early in the morning (normally from

4 a.m. to 6 a.m.). The number of hours spent dredging varies. Dredgers in Laguna spend 6-10 hours while those in Rizal and Pasig spend 4-6 hours and 4-8 hours,

respectively. Those in Rizal spend the least number of hours dredging because of their

proximity to dredging areas.

Volume of Snails Dredged

As mentioned earlier, while the rampant snail dredging activities during the early 1960s caused a decline in the lake‘s snail population (Ecological Footprint

Report 2013), the volume of snails gathered is still sufficient to meet the requirements

of the different buyers. The average volume of snails dredged across all locations is

158,546 kilograms (158. 54 tons) per week.

About 48% (75,904 kg) of the total weekly volume of snails dredged come from

Laguna while 32% and 20% are supplied by dredgers from Rizal and Pasig, respectively (Figure 4). The large percentage of snails coming from Laguna can be

attributed to the larger number of dredgers operating in the province.


As discussed earlier, more than half (58%) of the dredgers operating in the lake come

from Laguna.

Although Laguna dredgers contribute the largest share of the total volume of snails, they have the lowest average weekly volume dredged with 9,488 kilograms per

week. Dredgers in Pasig have the highest average volume of snails dredged, with

15,750 kilograms per week (Figure 5). This is not surprising since Pasig dredgers are fully dependent on dredging and their operation is larger compared with the Rizal and

Laguna operations. Pasig operators also have bigger boats and use more dredging

gears during their operations.

Figure 4. Percent share of snails dredged from Rizal, Laguna and Pasig, 2012

Location

Figure 5. Average volume of snails dredged from Rizal, Laguna and Pasig, 2012

Rizal Laguna Pasig


Snail Marketing

Thirty percent of the dredgers in Laguna and 25% of the dredgers in Rizal own a

duck farm in their municipalities. About 3% (4,055 kg or 4.06 tons) of the total volume of snails dredged were used in their respective farms while 154,070 kilograms

(154.07 tons) of the total volume of snails dredged were sold. The snails are usually

sold on a per sack or per pail basis. One sack is usually equivalent to 2-3 pails and contains 15-30 kilograms of snails. The average price of snails per pail across all

location is PhP14.00 or PhP1.44 per kilogram.

Snail dredgers have regular buyers from different provinces. The snails dredged from Laguna Lake are either delivered or picked up by the said buyers. In the case of

Rizal dredgers, traders procure snails at the Binangonan port on a daily basis. The

sales and transport of snails to the trader‘s truck usually occur at around 4 p.m. Pasig

dredgers, on the other hand, deliver directly the snails to traders that ordered them.

Snails dredged from Laguna Lake are brought to the provinces of Pampanga, Bulacan, Quezon and Laguna. Snails sold in Laguna are solely used as feeds for

ducks while those supplied in Quezon, Bulacan and Pampanga are not only fed to ducks but also to prawns grown in ponds. Figure 6 shows the geographic flow of

snails dredged from Laguna Lake. The snails dredged by Pasig and Rizal operators

are brought to the provinces of Pampanga and Bulacan only. Snails dredged by

Laguna operators, on the other hand, are sold in Bulacan, Quezon, and Laguna.

Figure 6. Geographic flow of snails dredged from Laguna Lake, 2012


The province of Laguna still accounts for 47% of the total volume of snails dredged from Laguna Lake. More than one third of the snails dredged in the province

is sold within the province, especially to duck farms that have proliferated in the area, while 17.4% of this volume is brought to the neighboring province of Quezon. Only a

very small percentage (1%) goes to Bulacan. Meanwhile the snails dredged in Rizal

are largely brought to Pampanga, accounting for about 88% of the total volume, and

the rest to Bulacan. About 62% of the volume dredged in Pasig is sold in Pampanga.

The demand for snails in the province of Pampanga is high due to the various

duck and prawn farms present in the area. Pampanga is known to be the top producer of prawns in the country; the province produced a total of 18,653 MT in 2012. On the

other hand, the total number of ducks in Pampanga is 916,043 while there are

818,316 ducks in Bulacan, 91,890 in Laguna and 148,675 in Quezon.

It can be observed that a large percentage of snails is sold in Laguna despite a

relatively lower number of ducks in the province. This is because majority of the snail dredgers in Laguna prefer to supply nearby farms due to lower transfer cost incurred

and the long-established business relationship with their regular customers (suki).

Profitability of Snail Dredging

Snail dredging is not a lucrative business that can provide the snail dredgers with

a high income. As shown in Table 4, the Pasig dredgers have the highest weekly gross sales of PhP26,400 but a gross margin of PhP 3,639 given their volume of operations

(15,750 kg/week). The dredgers in Laguna and Rizal have relatively higher returns

than those in Pasig since their operating costs are much lower. Pasig snail dredgers have higher operating costs (e.g., fuel, food) since they are the farthest from the lake

and employ more workers than their counterparts in Laguna and Rizal.

Table 4 also shows the gross margin ratio that amounts to 0.15, on the average.

This represents the proportion of each peso that a snail dredger retains as his profit. It

implies that a snail dredger would retain PhP0.15 from each peso generated for paying off his operating expenses. The larger the volume dredged, the higher the price

received and the lower the cost of operation, the higher would be the profit that can be retained by a snail dredger. A gross margin ratio of around 0.15 implies that snail

dredging operations are becoming more costly.

Despite the high cost of operation, snail dredgers are likely to remain in the business for the following reasons. First, these operators and hired workers lack

alternative high paying jobs or lucrative sources of livelihood and sure sources of

income. The workers interviewed are largely dependent on this business. Second, there remains a high demand for snails from duck and prawn farms. Third, operators

that have already incurred high investment costs are not likely to get out of the

business.


Table 4. Gross margins of owner-operators in Rizal, Laguna and Pasig, 2012

Conclusion

Despite the ban on snail dredging in the Laguna in 1995, it continues to this day.

This is due to the continuing demand for snails coming from duck and prawn farms.

Majority of duck farm owners prefer to use snails as feeds for their ducks because the

protein provided by these shells is necessary for the ducks to produce hard egg shells and thicker egg yolks. Snails are also preferred not only by duck farm owners because

it is cheaper compared with the other commercial feeds available in the market.

The ban on snail dredging in Laguna Lake has led to a reduction in the number

of snail dredgers and the supply of snails. These in turn led to the reduction in volume

of production or closure of a number of industries dependent on the supply of snails from the lake. Those dependent on the supply of snails include the duck farms,

balutan and salted eggs businesses in Laguna and Rizal. As a consequence, many

families have lost a major source of livelihood. Generation of alternative sources of livelihood for these families is needed to reduce the incentive to engage in snail

dredging in Laguna Lake. On the other hand, lifting the ban on snail dredging to protect these industries and the snail dredgers would mean continuous deterioration of

the lake‘s water quality. Snail dredgers must be made to understand that snail

dredging is inimical to the lake‘s current condition. As of now, they attempt to justify their activity by arguing that dredging helps remove the accumulated garbage in the

lake bed.

One possible means of protecting the lake from snail dredging while supporting

the duck and prawn industry is to undertake research that would determine where and

how snails can be profitably cultured. Snail farming, also known as heliculture, has been done in Europe, America and Southeast Asia. However, snail culture in other

countries is being practiced to provide a steady supply of snails for human

consumption. Studies on the possibility of applying this existing technology in the

Philippines, for snails used as duck and prawn feeds, are needed.

Location Item

Rizal Laguna Pasig

Gross sales 15,518 13,432 26,400

Variable cost 13,215 11,386 22,761

Gross margin 2,302 2,046 3,639

GM/GS ratio 0.15 0.15 0.14


References

Atienza, L. M., M. Rodriguez, D. Babilonia, M. Guimary and C. Rebancos. 2015.

―Food Safety Study of Duck Eggs Produced Along Laguna Lake Areas,

Philippines.”Journal of Food Nutrition and Food Sciences S3 (005).

Colting-Pulumbarit, C. and K. Lapitan. 2009. ―Weighing the costs of dredging the

lake.‖Philippine Daily Inquirer.Retrieved October 6, 2015, from http://

showbizandstyle.inquirer.net/lifestyle/lifestyle/view/20091014230071/Weighing

_the_costs_of_dredging_the_lake

CountrySTAT Philippines. 2015. Retrieved August 14, 2015, from http://

countrystat.psa.gov.ph

Delmendo, M. N. 1987. ―Milkfish Culture in Pens: An Assessment of its Contribution

to Overall Fishery Production of Laguna de Bay.‖ Makati, Metro Manila, Philippines. Retrieved August 9, 2015, from http://www.fao.org/documents/

card/en/c/577ba47a-d31f-54d2-bb09-00cbb55c315b/

Ecological Footprint Report: ―Restoring Balance in Laguna Lake.‖ 2013. Retrieved May 18, 2014, from http://www.footprintnetwork.org/images/article_uploads/

Philippines_2013_Ecological_Footprint.pdf

BFAR. Fisheries Administrative Order No. 35. 1952. Retrieved August 8, 2015, from

http://region7.bfar.da.gov.ph/Downloads/Fisheries%20Administrative%

20Order/3630/FAO%2030,%20s.%201952.pdf

Fisheries Administrative Order No. 41. 1955. Retrieved August 8, 2015, from

www.bfar.da.gov.ph/bfar/download/fao/FAO41.pdf

Manacop, P., and S. Capco. 1953. ―The Goby Dredge-Trawl Fishery of Laguna de Bay, with Notes on the Composition of the Commercial Catches.‖The Philippine

Journal of Fisheries2 (2):121-157. Retrieved August 10, 2015, fromhttp://

scinet.dost.gov.ph/unio n/Downloads/STII-S-354_A2353_121-157_313134.pdf

Palma, A., E. Mercene and M. Goss. 2005. ―Fish.‖ In R. D. Lasco, ed. Ecosystems

and People: The Philippine Millennium Ecosystem Assessment (MA) Sub-global Assessment. Los Baños, Laguna: Environmental Forestry Programme, College

of Forestry and Natural Resources, UPLB, pp. 116-132.

Republic Act No. 8550: The Philippine Fisheries Code of 1998. 1998. Retrieved

August 9, 2015, fromhttp://repository.gov.ph/customs/Regulated%20Imports%

20Supporting%20Documents%20%20by%20Individual%20Regulating%20Agency/Bureau%20of%20Fisheries%20and%20Aquatic%20Resources%20

(Dept.%20of%20Agriculture)/Legal%20Basis%20and%20Reference%20Ma

Santos-Borja, A. C. 2014. ―Dynamics of Laguna de Bay and Her Stakeholders.‖

Working paper, Laguna Lake Development Authority, Quezon City.

Santos-Borja, A. and D. Nepomuceno. 2006. Laguna de Bay: Experience and Lessons Learned Brief. Retrieved May 16, 2012, from World Lake Database: Interna-

tional Lake Environment Committee Foundation (ILEC): http://wldb.ilec.or.jp/

data/gef_reports /15_Laguna_de_Bay_27February2006.pdf


1 BS Agricultural Economics Graduate, College of Economics and Management, University of the Philippines

Los Baños 2 College of Economics and Management, University of the Philippines Los Baños

Email Addresses: [email protected], [email protected]

Effects of Extension Services on the Technical Efficiency of

Rice Farmers in Albay, 2014-2015

Lyndon A. Peña1 and Bates M. Bathan2

Abstract

Through the Farmers Information and Technology Services (FITS) program,

extension services in the form of trainings and seminars, and education and

communication materials were provided to rice farmers in Albay. Using farm-level

input-output data covering the wet season of 2014 and the socio-economic and

farm-specific characteristics of randomly sampled 30 farmer beneficiaries and 30

non-beneficiaries, a stochastic production function and technical inefficiency model were

estimated. Results of the study reveal that beneficiaries have higher input utilization, yield

and technical efficiency than non-beneficiaries. Furthermore, the amount of seed and

fertilizer significantly and positively affects the yield of rice. In addition, the number of

trainings and seminars attended, number of information, education and communication

(IEC) materials availed, and distance of farms to FITS center are the significant

determinants of technical inefficiency. The problems encountered in rice production are

the occurrences of typhoons and floods, high cost of inputs, insufficient capital, pest

infestation and limited knowledge of new farming technologies. It is recommended that

non-beneficiaries be encouraged to participate in FITS program and form farmers‘

association, irrigators‘ association or cooperative, as well as monitor the performance of

farmer beneficiaries.

Keywords: extension services, technical efficiency, FITS program

Introduction

The Department of Agriculture (DA), Department of Science and Technology

(DOST) and local government units (LGUs) have implemented various interventions such as provision of extension services, distribution of high-yielding seed varieties,

construction and rehabilitation of irrigation facilities, and delivery of credit and crop insurance aimed at increasing the productivity of rice farmers. The Techno Gabay

Program (TGP) of the DOST-Philippine Council for Agriculture, Aquatic and

Natural Resources Research and Development (PCAARRD) is one of the platforms where interventions are provided to farmers. It seeks to distribute science-based

information and bring technology services to the farmers, entrepreneurs, researchers and other stakeholders. Its primary objectives are to improve input-enhancing

technology in the agricultural sector and to provide support services extension to

LGUs (Aquino, Brown and Cardenas 2011). It has four components: (1) Farmers Information and Technology Services (FITS) or Techno Pinoy Centers, which

provide information and training services to farmers in order to improve their

agricultural production; (2) Information, Education and Communication (IEC), which provides its clients access to various forms of information through trade fairs,

print-ads, radio and online; (3) Information and Communication Technology (ICT), which involves the storage and delivery of data clients through short message service

(SMS) or electronic mail; and (4) Farmers-Scientist Bureau (FSB) or Magsasaka-

Siyentista, which enhances technology transfers to clients (Imperial 2011).

84 Lyndon A. Peña and Bates M. Bathan

In Albay, FITS centers were first established in 2006. From 2006 to 2011, the Bicol Consortium for Agriculture Resources, Research and Development (BCARRD)

was the one which implemented the program. In 2012, pursuant to Republic Act 7160 or the Local Government Code of the Philippines which devolves power,

responsibility and resources to LGUs, the program was transferred to the Agricultural

Training Institute (ATI) (Imperial 2011). After years of implementation, it is imperative to determine whether or not the extension services provided under the

program improved the level of productivity among rice farmers in Albay.

Literature Review

A study of Aveno et al. (2011), a chapter in Aquino, Brown, and Cardenas

(2011), assessed the effect of TGP on rice farmers in Central Luzon. Using 181

farmer beneficiaries and 30 non-beneficiaries, results showed that the yield and

technical efficiency of farmer beneficiaries were higher as compared to

non-beneficiaries. Gabunada et al. (2011), another chapter in Aquino, Brown, and Cardenas (2011), also studied the impact of TGP on rice farmers in Eastern Visayas.

Thirty rice farmers were interviewed and farm-level data before and after participation in the program were collected from the respondents. Results revealed

that yield and technical efficiency increased after program participation. Brown

(2011), still another chapter in Aquino, Brown, and Cardenas (2011) analyzed the impact of TGP on semi-temperate vegetable farmers by comparing input-output data

before (2003) and after (2009) the program. Based on the data from 30 respondents,

the net income and yield were significantly higher after the program.

Stochastic production frontier analysis has been widely used to determine the

factors affecting production and technical efficiency of rice farmers. Socio-economic and farm-specific characteristics of rice farmers are the variables which affect the

technical efficiency in rice production. In the study of Quilloy (2000), the use of

shallow tubewell has improved technical efficiency in rice production in Laguna. Tijani (2006) identified the application of traditional preparations, amount of off-farm

income, and contact with extension officers as significant determinants of technical inefficiency in rice production in Iljesha Land of Osun State in Nigeria. Kadiri et al.

(2014) revealed that marital status, education and farm size influenced the technical

efficiency in rice farming in the Niger Delta region of Nigeria.

Research Methodology

Sources of Data and Methods of Data Collection

Primary data on rice output, area and inputs of production (i.e., seeds, fertilizers, family labor, hired labor, man-animal labor, man-machine labor, pesticide and

herbicide) in the wet season of 2014, socio-economic and farm-specific characteristics (i.e., age, sex, educational attainment, years in farming, topography and type of soil),

number of FITS trainings and seminars attended, number of IEC materials availed,

and distance of farms to FITS center were collected from rice farmer respondents through face-to-face interviews using a pre-tested questionnaire. The municipalities of

Oas and Polangui in Albay were selected because of the presence of many rice farmer

beneficiaries. Using simple sample matching with age, sex, educational attainment, topography and number of years in farming as auxiliary variables, a total of 30 sample

matches (i.e., 30 beneficiaries and 30 non-beneficiaries) were drawn from the sampling frame consisting of 4,356 farmer beneficiaries and 1,421 non-beneficiaries

in Oas and Polangui.


Descriptive Analysis and Test of Two Means

Descriptive analysis was employed to describe the primary data with the aid of

tables. Test of two means was conducted to compare the input utilization, production, area, yield, and technical efficiency between farmer beneficiaries and

non-beneficiaries.

Stochastic Production Frontier Analysis

Using Frontier 4.1, a production function and a technical inefficiency function

were simultaneously estimated to determine the factors affecting yield and technical

efficiency of rice farmers (Coelli 1996). A Cobb-Douglas production function was

estimated which is expressed as follows:

ln Yi = β0 + β1lnX1i + β2lnX2i + β3lnX3i+ β4lnX4i + β5lnX5i + β6lnX6i+ β7lnX7i

+ β8lnX8i + (vi - ui)

where :ln = natural logarithm

Yi = yield of the ith farm (mt/ha)

X1i = amount of seed used by the ith farm per hectare (kg/ha)

X2i = amount of fertilizer used by the ith farm per hectare (kg of N/ha)

X3i = amount of family labor used by the ith farm per hectare (man day/ha)

X4i = amount of hired labor used by the ith farm per hectare (man day/ha)

X5i = amount of animal labor used by the ith farm per hectare (man-animal

day/ha)

X6i = amount of machine labor used by the ith farm per hectare (man-machine

day/ha)

X7i = amount of pesticide used by the ith farm per hectare (liters/ha)

X8i = amount of herbicide used by the ith farm per hectare (liters/ha)

vi = random variable assumed to be independently and identically

distributed with mean zero and variance σ2 and independent of ui (i.e.,

error term for statistical noise, weather disturbances and other factors

out of rice farmer‘s control)

ui = non-negative random variable assumed to account for technical

inefficiency in production

β0 = intercept

β‘s = parameter estimates

To identify the determinants of technical inefficiency, the following model was

specified:

ui= δ0+ δ1Z1i + δ2Z2i + δ3Z3i + δ4Z4i


where:

Z1 = number of FITS trainings and seminars attended

Z2 = number of IEC materials availed

Z3 = distance of farms to FITS center (kilometers)

Z4 = dummy variable for soil type (1 = loamy, 0 = otherwise)

δ0 = intercept

δ‘s = parameter estimates

ξi = error term

Mathematically, to compute for technical efficiency, the average production

function (Y) can be expressed in terms of frontier production function (Y*):

Yi = Yi* exp (-ui)

The ratio of Y and Y* provides the estimate of technical efficiency (TE).

TEi = Yi/Yi* = exp (-ui)

Results of the Study

Socio-economic and Farm-specific Characteristics of Rice Farmer Respondents

The average age of rice farmer respondents from Oas and Polangui in Albay is 50 and 80% of the respondents are male. They are high school graduate and have

been engaged in rice farming for 20 years. The average level of income of

beneficiaries is PhP 39,534, which is higher than the average income of non-beneficiaries (PhP 32,700). The difference is significant at 5% level of probability.

All of the respondents cultivate palay in a flat topography. It is the most ideal type of

topography for rice cultivation because it is easier to cultivate and manage and to retain water. Twenty-four farmer beneficiaries and 26 non-beneficiaries plant palay

on loamy soil. According to the USDA textural soil classification study guide (1987), loamy soil is ideal and commonly used in rice production as it contains more

nutrients, moisture and humus as compared to sandy and silty soils. The distance of

farm to FITS center is significantly higher for farmer beneficiaries (6.02 km) than non

-beneficiaries (5.31) at 5% level of probability.

Input Utilization of Rice Farmer Respondents and Area and Yield of Sample Rice

Farms

As shown in Table 1, the amounts of seed, fertilizer, hired labor, machine labor

and pesticide are significantly higher for farmer beneficiaries than non-beneficiaries. For the type of seeds used, almost half of the beneficiaries and non-beneficiaries use

hybrid seeds. In terms of fertilizer applied, both types of rice farmers used urea with

price ranging from PhP 1,300 to 1,500 per 20-kilogram bag and ammonium sulfate and diammonium phosphate with price ranging from PhP 1,200 to 1,400 per 20-

kilogram bag. In terms of labor utilization, family labor is only minimal for both types of rice farmers as it is only used during manual transplanting and while overseeing the

farm during harvesting. Hired labor is devoted to transplanting and application of

fertilizer, pesticide and herbicide. Hired laborers are paid an amount ranging from PhP 150 to 300 per manday. Animal labor is utilized for plowing operation. Animal

labor is also low at it is only utilized for plowing operation.


The animals and implements used in this operation are either owned or rented by the rice farmers. The cost ranges from PhP 400 to 600 per man-animal day. Machine

labor is used for plowing and harrowing operations with cost ranging from PhP 1,100 to 1,300 per man-machine day. The machines are either rented, and payment includes

the salary of the laborer, or owned by the operator.

Table 1. Average input utilization of rice farmer respondents and area and yield

of sample rice farms by type of respondent, Oas and Polangui, Albay,

2014 (N = 60)

** and *** = significant at 5% and 1% level of probability, respectively

ns = not significant at 5% level of probability

Rats and snails are the most common pests among rice farms. Pesticides with price ranging from PhP 1,300 to 1,600 per liter are used by both types of rice farmers.

Weeds also compete with the nutrients intended for palay and thus, applied by both types of rice farmers on their rice farms. The price of herbicide ranges from PhP

1,100 to PhP 1,500 per liter.

The average farm size of farmer non-beneficiaries is 1.77 hectares in comparison with 1.71 hectares estimated for farmer beneficiaries. The difference in average farm

sizes between the two types of rice farmers is not statistically significant at 5% level

of probability. However, yield is significantly higher for farmer beneficiaries as compared with non-beneficiaries. At 5% level of probability, the yield difference of

0.56 hectare is statistically significant. This could be explained by the higher input utilization in terms of seed, fertilizer, hired labor, machine power and pesticide as

well as the better cultural management practices of farmer beneficiaries relative to

non-beneficiaries.

Item

Farmer

Beneficiaries

(n=30)

Farmer Non-

beneficiaries

(n=30)

Mean

Difference

Standard

Error

Seed (kg/ha) 50.79 40.81 9.98** 3.189

Fertilizer (kg of N/ha) 90.48 66.33 24.15*** 5.902

Family labor (man day/ha) 1.07 1.03 0.04ns 0.139

Hired labor (man day/ha) 10.33 9.6 0.73*** 0.412

Animal labor

(man-animal day/ha) 1.8 1.53 0.27ns 0.165

Machine labor

(man-machine day/ha) 1.87 1.2 0.67** 0.110

Pesticide (liter/ha) 1.55 1.01 0.54** 0.145

Herbicide (liter/ha) 1.4 1.62 -0.22ns 0.155

Area (ha) 1.71 1.77 -0.06ns 0.026

Yield (mt/ha) 3.00 2.44 0.56*** 0.170


Although all of them perform seedbed preparation, plowing, transplanting, application of fertilizer, application of pesticide and herbicide, and harvesting as

cultural management practices, 40% and 30% of the beneficiaries mechanized their plowing and harvesting operations, respectively. This is in contrast with 30% and

20% of the non-beneficiaries who used machines in their plowing and harvesting

operations, respectively.

Results of the Stochastic Production Frontier Analysis

Table 2 summarizes the results of the estimation of the OLS and MLE

models. The presence or absence of technical inefficiency was determined using the t-test of gamma and the generalized likelihood ratio (GLR) test. Based on the

results of the t-test of gamma, the value of gamma which is equal to 0.50 is found to

be significant at 5% level of probability. This suggests that 50% of the residual

variation in the model is due to technical inefficiency effects while the rest is due to

random error. Since the value of sigma-squared is also significant at 5% level of probability; this indicates the correctness of the specified assumption of the

distribution of the error term. On the other hand, the GLR test shows that since the LR test statistic of 27.06 is greater than the chi-square value of 17.192 from Kodde

and Palm, the null hypothesis that technical inefficiency is absent in the model

(i.e., accept OLS model) is rejected. Therefore, the MLE model better fits the data

of the 60 sample rice farmer respondents than the OLS model.

Table 2. Results of the Ordinary Least Squares (OLS) and Maximum

Likelihood Estimates (MLE) of the stochastic production frontier

model for rice farmers, Oas and Polangui, Albay, 2014 (n = 60)

Item Parameter OLS Coefficient MLE Coefficient

Production function

Constant β0 0.4182ns -2.1269**

Seed β1 0.3056ns 0.8003***

Fertilizer β2 0.2481** 0.5007***

Family labor β3 0.0001ns -0.0001ns

Hired labor β4 -0.4609** -0.0636ns

Animal labor β5 0.0730ns -0.0234ns

Machine labor β6 -0.0002ns -0.0001ns

Pesticide β7 0.1160ns 0.0562ns

Herbicide β8 -0.0541ns 0.0998ns

Technical inefficiency

function

Constant δ0 0.0909ns

Number of FITS

trainings and seminars

attended

δ1 -0.0003**


Number of IEC

materials availed δ2 -0.0007**

Distance of farm to

FITS center δ3 -0.2788**

Soil type δ4 -0.1488ns

Variance parameters

Sigma-squared σ2 0.0820**

Gamma γ 0.0500**

Log-likelihood value -9.1272 4.4022

LR test statistic 27.0587**

** and *** = significant at 5% and 1% level of probability, respectively

ns = not significant at 5% level of probability

Among the factors of production, the amounts of seed and fertilizer

significantly affect yield at 1% level of probability. A one percent increase in the use of seeds per hectare, holding other factors constant, would result in a 0.80%

increase in rice yield (Table 2). For fertilizer usage, all things remain unchanged, a 1% increase in kilogram of nitrogen per hectare would lead to a 0.50% increase in

yield (Table 2). The average amounts of seed and fertilizer utilized by the rice

farmer respondents, especially the non-beneficiaries, could be below the recommended amounts, and thus increasing the utilization of these inputs could still

improve rice yield.

The number of trainings and seminars attended, number of IEC materials availed, and distance of farm to FITS center significantly influenced technical

inefficiency at 5% level of probability (Table 2). As rice farmers attend more

trainings and seminars, they become less technically inefficient as shown the negative coefficient of this program intervention (Table 2).These trainings and

seminars are conducted by the municipal agriculturists and agricultural technicians

and the topics included use of hybrid seeds, demonstration on the use of mechanized farm equipment, relationship between climate change and agriculture,

and preparation for drought or heavy rainfall. This program intervention has improved the productivity of rice farmers through an increase in farm-level

technical efficiency.

As rice farmers avail of more IEC materials, their level of technical

inefficiency decreases, and thus, they become more technically efficient in rice

farming. These materials use research-based communication procedures which facilitate technology adoption, and these include the recommended cultural

management practices. They are disseminated in the form of flyers and booklets.

The negative coefficient of IEC materials means that this intervention has improved the technical efficiency of rice farmers, particularly through adoption of better

cultural management practices (Table 2).


Lastly, as the distance between the farm and FITS center increases, the level of technical inefficiency decreases. This seems contrary to expectation that farther

distance, since it means higher cost on the part of the farmers, should negatively affect technical efficiency. However, those who incur more financial costs and higher

opportunity cost of time due to the long distance travelled may have appreciated and

valued what they have learned from the trainings and seminars held at FITS centers

relative to those with lower money and time allotted for the trainings and seminars.

Technical Efficiency in Rice Production of Farmer Respondents

The mean level of technical efficiency of rice farmer respondents was 76.10% (Table 3). This denotes that the rice farmers can still improve yield by 23.90%.

Extension services in the form of trainings and seminars and IEC materials can be

used as interventions which could significantly increase yield and move the rice

farmers near the production frontier. The difference in mean technical efficiency

levels between rice farmer beneficiaries and non-beneficiaries was significant at 5% level of probability (Table 3). Yield obtained by rice farmer beneficiaries can only be

improved by 6.62% but non-beneficiaries can still improved yield by 41.18% (Table 3). Majority of rice farmer beneficiaries had a technical efficiency level between 91 to

100% while most non-beneficiaries lie in the 61 to 70.99% range (Table 3).

Table 3. Frequency and distribution of technical efficiency levels of rice farmer

respondents by type of respondent, Oas and Polangui, Albay, 2014

(n = 60)

Technical Efficiency

Level (%)

Farmer Beneficiaries Farmer Non-Beneficiaries

Frequency % Frequency %

Below 30.99 0 0 0 0

31 to 40.99 0 0 1 3.33

41 to 50.99 0 0 1 3.33

51 to 60.99 0 0 17 56.67

61 to 70.99 0 0 11 36.67

71 to 80.99 6 20 0 0

81 to 90.99 1 3.33 0 0

91 to 100.00 23 76.67 0 0

Total 30 100 30 100

Mean by type 93.38 58.82

Mean difference 34.56**

Standard error mean 0.0194

Mean of all types 76.10

** = significant at 5% level of probability


Problems Encountered in Rice Production by Farmer Respondents

In general, the most cited problem in rice production by the farmer

respondents was the occurrences of typhoons and floods (Table 4). This was followed by high cost of inputs, insufficient capital, pest infestation and limited

knowledge and skills on new farming technologies. High cost of inputs and

insufficient capital were reported by more farmer non-beneficiaries than beneficiaries as problems in rice production. Limited knowledge and skills on new

farming technologies was cited by more farmer non-beneficiaries as a problem

encountered in rice production as compared to beneficiaries. This could be addressed if they have access to extension services in the form of trainings and

seminars and IEC materials. On the other hand, although more rice farmer beneficiaries report pest infestation as a problem, they managed to minimize the

negative effect of this problem more effectively than non-beneficiaries. This is by

means of utilizing more pesticides as shown by the higher level of pesticide use by

rice farmer beneficiaries relative to non-beneficiaries.

Table 4. Frequency and distribution of rice farmer respondents by type of

problem encountered in rice production in Oas and Polangui, Albay,

2014 (N = 60)

Conclusions and Policy Implications

The FITS program provides extension services in the form of trainings and

seminars and IEC materials to the rice farmer beneficiaries in Oas and Polangui in Albay with the aim of increasing the productivity of rice farms. As rice farmer

beneficiaries learn and adopt better cultural management practices, yield is expected

to be higher than that of non-beneficiaries. Results show that the yield of rice farmer beneficiaries is higher than that of non-beneficiaries. They also use more

inputs on a per hectare basis, particularly seed, fertilizer, hired labor, machine labor

and pesticide, relative to non-beneficiaries.

Problem Farmer Beneficiaries

Farmer Non-

Beneficiaries All Types

Frequency % Frequency % Frequency %

Occurrences

of typhoons

and floods

30 100 30 100 60 100

High cost of

inputs 20 66.67 21 70 41 68.33

Insufficient

capital 14 46.67 17 56.67 31 51.67

Pest

infestation 16 53.33 14 46.67 30 50

Limited

knowledge

and skills on

new farming

technologies

7 23.33 13 43.33 20 33.33

Note: Percentages do not add up to 100% as multiple responses are allowed.


Ultimately, farmer beneficiaries have higher levels of technical efficiency than non-beneficiaries. In order to increase yield, higher amounts of seed and fertilizer

per hectare are needed to be applied on rice farms. The number of trainings and seminars attended, number of IEC materials availed and distance of farm to FITS

center are also found to positively affect the technical efficiency of rice farmers.

The problems encountered in rice production which require attention are the occurrences of typhoons and floods, high cost of inputs, insufficient capital, pest

infestation and limited knowledge on new farming technologies.

Drawing from the results of the study, the following policy implications are

identified:

Access to timely and sufficient credit. Since there is a need to improve the

utilization of seed and fertilizer in order to increase yield and to address the high

cost of inputs and insufficient capital which were cited as problems, access to timely

and sufficient credit is necessary.

Participation in the FITS program. Rice farmer non-beneficiaries should be

encouraged to participate in the FITS program as it has been proven to increase farm-level technical efficiency. The limited knowledge about new farming

technologies can be addressed through the extension services provided under the

program.

References

Aquino, A. P., E. O. Brown and D. C. Cardenas. 2011. Impact Assessment of the

Techno Gabay Program in Selected Municipalities - The Farm-level Impact of the Techno Gabay Program: Assessment, Evidences and Implications. Book

Series No. 182/2011, DOST-PCAARRD, Los Baños, Laguna.

Aveno, J. L., L. M. Galang, T. T. Battad, M. E. M. Orden and J. R. Suyat. 2011.

"Impact of the Techno Gabay Program on Palay Farms in Selected

Municipalities of Central Luzon" in Aquino, Brown and Cardenas (2011), pp.

74-100.

Brown, E. O. 2011. "Impact Assessment of the Techno Gabay Program in Selected

Municipalities" in Aquino, Brown and Cardenas (2011), pp. 29-40.

Coelli, T. J. 1996. A Guide to FRONTIER Version 4.1: A Computer Program for

Stochastic Frontier Production and Cost Estimation. Centre for Efficiency and Productivity Analysis (CEPA) Working Papers, Department of Econometrics,

University of New England, Armidale, Australia.

Gabunada, F. M., S. B. Anwar, A. P. Aquino and P. A. B. Ani. 2011. "Impact of the Techno Gabay Program on Palay Farms in Selected Municipalities of Leyte,

Eastern Visayas" in Aquino, Brown and Cardenas (2011), pp. 43-56.

Imperial, P. L. B. 2011. Techno Gabay Program: From Outside, Looking In.

Retrieved from http://ati.da.gov.ph/bicol/feature/techno-gabay-program-

outside-looking.


Kadiri, F. A., C. C. Eze, J.S. Orebiyi, J. I. Lemchi and D.O. Ohajianya. 2014. ―Technical Efficiency in Niger Delta Region of Nigeria. ‖European Center for

Research Training and Development UK. Global Journal of Agricultural

Research, Vol. 2, No. 2, pp. 33-43.

Quilloy, A. J. A. 2000. The Economics of Shallow Tubewell Irrigated Rice

Production, Laguna. November 1998-February 1999. Department of Agricultural Economics, University of the Philippines Los Baños, College,

Laguna.

Tijani, A. A. 2006. ―Analysis of the Technical Efficiency of Rice Farms,‖ in Ijesha

Land of Osun State, Nigeria. Agekon, Vol. 45, No. 2.

United States Department of Agriculture. 1987. USDA Textural Soil Classification

Study Guide. Retrieved at http://www.wcc.nrcs.usda.gov/ftpref/wntsc/H&H/

training/soilsOther/soil-USDA-textural-class.pdf <23 December 2015>


1 B.S. Economics Graduate, College of Economics and Management, University of the Philippines Los Baños


Email: [email protected]

Decoy Effect and Student Preference with regard to USB Flash Drives

Mark Angelo R. Alcaide1 and Jefferson A. Arapoc2

Abstract

Standard economic theory assumes that individuals make decisions that are

consistent across different choice situations. Behavioral economics challenges the

standard economic theory account of individual behavior. This budding field suggests that

consumer preference for one option over another changes as a result of adding another

option. This study aims to determine the effect of a decoy option on consumer preference

for USB flash drives. A quasi-experiment was conducted among 100 UPLB students

wherein each respondent was initially asked to choose between two kinds of flash drive.

The researcher then adds a decoy option to test if the respondent will change his initial

preference. To investigate whether a decoy option causes preference reversal,

Independence of Irrelevant Alternatives was verified through Hausman specification test.

Results show that the introduction of a decoy option results in inconsistencies in the

choice of several respondents. This implies that the standard economic theory account of

individual behavior may not always hold given that consumer‘s preferences are based on

what is on offer rather than on absolute preferences.

Introduction

―Homo Economicus‖ is the concept in standard theory that portrays individual

consumers as self-interested decision makers trying to maximize personal advantage. Given that consumers are faced with different sets of alternatives during market

transactions, their decision in choosing an option will always be based on the concept

of utility maximization (Foka-Kavalieraki and Hatzis 2011).

Rational choice theory, as one of the core foundations of neoclassical

economics, provides more concrete explanation about consumer rationality. Axioms of rational choice allow economics to explain unambiguous rankings of bundles.

Consumers undergo different processes of evaluating options of bundles before selecting a specific bundle. The selected commodity bundle should conform to the

assumptions of the theory in order for the consumer choice to be considered rational.

However, the assumption that individuals conform to the concept of Homo economicus has been subjected to criticisms through the years. The rationality

assumption, as the foundation idea of neoclassical economics, results in models that

are unable to explain real-world scenarios. Behavioral and experimental economists were able to provide experimental results that challenge the assumptions of

neoclassical economics regarding individual choice.

Dan Ariely (2008), a well-known behavioral economist, highlighted the concept

of relativity in consumer rationality. According to him, an individual‘s decision

making capability in choosing between a set of options is highly affected by the advantage of an option relative to the others. He added that humans rarely choose

things in absolute terms because they tend to compare each option available to them,

and humans are not capable of immediately assessing the worth of things.

96 Mark Angelo R. Alcaide and Jefferson A. Arapoc

Individuals cannot instantly determine what specific item or option they prefer unless they see it in a clear context. When a consumer is faced with several

alternatives, comparison of each alternative takes place, which allows the consumer to rank these alternatives. The combination of options has a significant role in the

decision making of individuals in choosing among different sets of choices.

Inconsistencies in individual preferences and irrational behavior were observed in different situations which raises some questions regarding the validity

of the rational choice theory. Some examples of choice inconsistencies are evident

using the concept of ―decoy effect‖ presented by Ariely (2008), Wedell and Petibone (1996) and Park and Kim (2005). These studies provided evidences that

proved that adding an additional option which is relatively higher than the inferior alternative will tend to increase the attractiveness of the inferior alternative, causing

consumers to change their preference. An additional alternative in an existing set of

choices that may cause some irregularities on consumer preference is known as the ―decoy option.‖ The power of relativity in the context of choice plays a significant

role in explaining consumer behavior. Since traditional economic theories lack the ability to describe violations of basic economic assumptions, behavioral economists

attempt to create better descriptive models and explanations about human decision

making process. The study aims to provide empirical evidence on the effect of a decoy option on consumer choice. The results of the study provides important

insights in understanding the concept of relativity and its effect on consumer

choice.

Analytical Framework

Rational Choice Theory

Rational choice theory assumes that individual preferences satisfy the axioms

of preference relation, which means that an individual is able to rank goods with

consistent preference ordering (Read 2008). In any market transactions, consumers are faced with ―choice set‖ or ―consumption set‖ that contains different alternatives

or options with different characteristics. Given two equally attractive alternatives with specific attributes, selection of an option will always depend on the

consumer‘s judgement. The decision process of a consumer can be explained by

two models of consumer behavior: 1) the preference-based approach and 2) the

choice-based approach.

Preference-based choice simply suggests that a consumer facing a choice set

should have a preference that will satisfy the rationality axioms of preference

relation. It means that a rational individual should have a preference that is

complete, reflexive and transitive. Completeness refers to a situation where a consumer has the ability to rank options. Reflexivity assumption, on the other hand,

states that an option will be ―always as good as itself‖. Lastly, transitivity refers to

consistency of an individual‘s preference (Foka-Kavalieraki and Hatzis 2011).

Meanwhile, choice-based preference is grounded on the concept of the Weak

Axiom of Revealed Preference (WARP). WARP simply states that if an individual

prefers a certain option over another, say alternative A over alternative B from a given set of alternatives, then option B should never be chosen over A—unless A is

not available or affordable (Chapman 2003).


Violation of Rationality Choice Theory and Decoy Effect

Decoy effect was first identified by Huber, Payne, and Puto (1982). Applications

of this idea in different fields follow through the years and varieties of studies were conducted that generated interesting results. A decoy by definition is an alternative

choice added in a set of options which basically aims to alter or change the

attractiveness of a certain alternative relative to others (Wedell and Pettibone 1996). Inclusion of decoy alternative in the choice is presumed to have an effect on the

consumers‘ behavior, following the idea that the decoy is biased towards a specific

alternative; it can cause preference reversals once it is added in the choice set. To further understand how the decoy effect works, assume that a consumer is faced with

two alternatives, Option A and B. If the consumer initially chooses option A, the individual is likely to choose option B after employing a decoy option. This situation

is clearly inconsistent with rational choice theory since, in the case presented above,

an irrelevant alternative—a decoy option—affects the initial rankings of alternatives.

Analytical Procedure

Data Collection and Survey Design

This study used primary data which were gathered through a survey among 100

UPLB undergraduate students. The researcher requested the list of registered students in the first semester of academic year 2013-2014 from the Office of the University

Registrar. From the requested list, a simple random sampling was employed. The data

were collected using an online survey for convenience purposes.

Respondents were asked to rank, based on their ―subjective‖ preferences, a

number of Portable storage devices (more commonly known as USBs) with different

features. A portable storage device was chosen to be the commodity in this study

since the target respondents are well accustomed to this kind of device.

Figure 1: Consumer preferences and the decoy effect

INITIAL CHOICE SET

OPTION

A

CHOICE SET WITH DECOY

OPTION

B

Preference Reversal

OPTION

A

DECOY

OPTION

OPTION

B


The respondents were subjected to two rounds of choice selection process. Three alternatives were available in each round. For the first round, the respondents were

made to face three initial options. The first option is a portable storage device with a storage capacity of 16 gigabytes and a price of 500 pesos. The second option, on the

other hand, is a device with a storage capacity of 8 gigabytes and a price of 250 pesos.

Meanwhile, the third option was set to be a strictly dominated option with a storage capacity of 4 gigabytes and a price of 700 pesos. Clearly, the third option is

dominated by the first two options and was presumed to be not chosen by any

respondents. The rationale for employing the third option in the first round was to

make the available options equal for both rounds.

For the second round of the selection process, the first and second options were both retained. The third option was replaced with a decoy option which is biased

against the initial decision made by the respondent in the first round (Figure 2). For

example, a respondent choosing the first option in the first round will be facing the same set of options (excluding the third option) plus a decoy option with a storage

capacity of 2 gigabytes and a price of 250 pesos. Noticeably, the decoy option makes

the second option relatively more attractive to the respondents.

Figure 2: Options available to the respondents in the 2-round selection process

Portable Storage Device Storage Capacity Price

Option 1 16 GB PhP 500.00


Option 3 ( Strictly Dominated Option) 4 GB PhP 700.00




Option 3 ( Decoy Option) 4 GB PhP 700.00




Option 3 ( Decoy Option) 4 GB PhP 700.00

ROUND

1

ROUND

2

If Option 1 is chosen in Round 1

If Option 2 is chosen in Round 1


Test of Independence of Irrelevant Alternatives (IIA)

Upon the collection of information from the online survey, a test of

independence of irrelevant alternatives (IIA) was conducted. This test examines the independence of each alternative from one another, and it also tests whether an

inclusion or removal of a new or existing alternative affects the probability of the

remaining options to be chosen. A Hausman specification test was used to test for

IIA in order to evaluate the impacts of a decoy option on consumer‘s choice.

Hausman specification test requires two estimates from two models. The

unrestricted model is a multinomial logistic model where the first estimates were taken. The independent variable for this model is based on the consistency of choice

made by the respondents in the first and second round of survey, while the

dependent variables include their socio-demographic characteristics and other

characteristics that might affect their decision-making process. The coefficients of

the unrestricted model were consistent and efficient under null and alternative hypotheses. On the other hand, the second estimates came from the restricted

model. Unlike the first model, the estimates in the second model were generated by employing logistic regression estimation. The independent variable for the restricted

model was based on the decision made by the respondents in the first round of

survey; whether the respondents opted to choose option 1 or option 2. The coefficients in the restricted model were found to be inconsistent under alternative

hypothesis, but efficient under null hypothesis.


Socio-demographic Characteristics of Respondents

A total of 102 students were randomly selected across the nine colleges. Most of the randomly selected students are from CAS (27%) and CEAT (22%). Table 1

shows the summary of the distribution of sex and classification of the respondents.

More than half of the respondents in the sample are female (59.8%). The oldest respondent of the sample was from batch 2009, while the youngest came from batch

2013.

Table 1. Distribution of respondents by sex and class standing or

“classification”

College Sex Classification

Male Female Freshman Sophomore Junior Senior

CA 4 6 0 4 3 3

CACAS 1 1 0 1 0 1

CAS 16 12 3 12 7 6

CDC 4 0 1 1 2 0

CEAT 9 14 3 4 10 6

CEM 12 6 3 2 9 4

CFNR 5 1 0 3 1 2

CHE 6 0 0 1 2 3

CVM 4 1 0 3 0 2

Total 61 41 10 31 34 27


Preference with Regard to Portable Storage Device (USB)

Table 2 shows the distribution of choices made by the 102 UPLB undergraduate

students in the first round of survey. Most respondents (about 63.73%) preferred Option 2, a portable storage device with a storage capacity of 8GB and a price of 250

pesos. One-third (33.33%) of the respondents, on the other hand, favored option 1, a

portable storage device with a storage capacity of 16GB and a price of 500 pesos. Interestingly, three (3) respondents selected the strictly dominated alternative. These

respondents were immediately dropped from the sample.

Table 2. Distribution of respondents’ choice in the first round of the survey

Application of Decoy Effect

The decoy option presented to the respondent in the second round depends on their initial choice in the first round. The decoy option added in the choice set favored

the option which was not chosen in the first round of survey. Surprisingly, more than

25% of the respondents changed their initial preference in the second round of survey (Table 3). It was apparent that most of the respondents who experience preference

reversal are those who initially chose Option 2 in the first round (Table 4). This can

possibly be explained by the fact that respondents who initially prefer an 8G USB find it more appealing to shift 16G USB given that the decoy option employed gave

them an illusion of relatively more savings in reversing their initial preference. On the other hand, respondents who initially choose 16G USB might not be that affected by

the Decoy Option since they might really prefer to have a flash drive with higher

capacity without even considering its price.

Table 3. Distribution of preference reversals in the second round of survey

Table 4. Preference reversals

Portable Storage Device (USB) Number of Selections

16 GB with price 500 34

8GB with price 250 65

4GB with price 700 3

Total 102

Preference Reversal Frequency Percent Cumulative

No 74 74.75 74.75

Yes 25 25.25 100.00

Total 99 100

Movement of Choice Number of Respondents

16 GB to 8GB 3

8GB to 16GB 22

Total 25


Impact of a Decoy Option

In order to determine whether the presence of decoy affects the respondent‘s

choice, Hausman specification test was done to test the IIA. A multiple logistic regression was initially employed in order to have the estimates from the unrestricted

model.

The dependent variable for this model includes three categories: 1) Consistent option 1(C1) was used to show respondents that consistently chose 16GB USB from

both rounds of survey, 2) Consistent option 2 (C2) was used to show respondents that

consistently chose 8GB USB, and 3) Preference reversal (PR) was used to show respondents who experienced preference reversal. The first model was found to be

significant at α = 5%. For the restricted model, first round choices of the respondents

were used. The model was also found to be significant at α = 5%.

Table 5. Hausman specification test

b = consistent under Ho and Ha; obtained from mlogit

B = inconsistent under Ha, efficient under Ho; obtained from logit

Test: Ho: difference in coefficiencts not systematic.

chi2(11) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 570.90

The Hausman specification test was conducted by using the estimates from the two models mentioned above. Based on the results, the estimates from two models

were found to violate the IIA assumption. The estimates in the unrestricted model are

consistent under the null and alternative hypothesis. On the other hand, the estimates in the restricted are inconsistent under the alternative hypothesis but efficient under

the null hypothesis. Based on the results of the Hausman specification test, the p-value was found to be significant; thus, the differences in the coefficients of the

estimates in the two models are systematic, meaning alternatives are not independent

from each other. The results showed that the presence of decoy alternative in the

second round of the choice set has a significant effect in the respondents‘ choice.

Coefficient

Variable (b)

1st Estimate

(B)

2nd Estimate

(b-B)

Difference

Sqrt (Diag (V_b – V_B))

S.E

Age 1.1517 -.9524 2.1041 .1600

Allowance -.20905 .2210 -.4301 .1942

Female .79320 -.0336 .8268 .3009

Sophomore -1.5922 1.9658 -3.5580 .5220

Junior -4.4048 4.2656 -8.6703 .6003

Senior -4.2999 4.0961 -8.3960 .6260

Public .6777 -.61085 1.2885 .2691

No computer

course

1.9312 -.87415 2.8054 1.1773

Printing -1.1095 .2807 -1.3902 .3236

Back-up -.31986 -.5941 .2743 .5118

Storage of

other files

-.6472 -.7870 .1398 .4398


Summary and Conclusion

The study aims to determine the impact of a decoy option on the consistency of

consumers‘ choice across different choice situations. In order to achieve the objectives of this study, data were gathered through an online survey. The survey

consists of two rounds to allow examination of the consistency of respondents‘

choice.

In general, the introduction of a decoy option was found to affect consumers‘

choice under an experimental setup. In fact, more than 25% of the respondents shifted

their preference upon the introduction of a decoy option. Moreover, a test of independence of irrelevant alternatives (IIA) was used to test whether the presence of

a decoy option in a choice set affects the consistency of consumers‘ preference. By

using the Hausman specification test, the study was able to provide empirical

evidence that the introduction of a decoy option might affect consumer‘s choice

across different choice situations. The results of the study provide relevant insights on consumer behaviour and rationality, particularly on its effectiveness in nudging

consumer selection process. One important application of decoy effect is in business and marketing. The possibility that consumer choices could be swayed opens up

opportunities for business owners to use the decoy effect to their advantage ― for

example, offering options to consumers that will make the business firm‘s products

relatively more attractive than their counterparts in the market.

References

Ariely, D. 2008. Predictable Irrational, The Hidden Forces That Shape Our

Decisions. Harper Collins Publisher.

Camerer, C. and G. Loewenstein. 2002. Behavioral Economics: Past, Present, Future. Unpublished, Department of Social and Decision Sciences Carnegie-

Mellon University.

Chapman, B. 2003. ―Rational Choice and Categorical Reason‖. University of

Pennsylvania Law Review 151: 169–210.

Conolly, T., J. Reb and E. E. Kausel. 2013. ―Regret Salience and Accountability in Decoy Effect‖. Journal of Judgment and Decision Making Vol. 8 No. 2: 136-

149.

Foka-Kavalieraki, Y. and A. N. Hatzis. 2011. ―Rational After All: Toward an Improved Theory of Rationality in Economics.‖ Paper presented at the First Pan

-Hellenic Conference in Philosophy of Science Athens, October 2010.

Huber, J., J.W. Payne and C. Puto. 1982. ―Adding Asymmetrically Dominated Alternatives: Violations of Regularity and the Similarity Hypothesis.‖ The

Journal of Consumer Research Vol. 9 No.1: 90-98.

Merriam-Webster. 2013. Attractiveness. Merriam-Webster Dictionary.

Park, J. and J. Kim. 2005. ―The Effects of Decoys on Preference Shifts: The Role of

Attractiveness and Providing Justification‖ Journal of Consumer Psychology

Vol. 15 No. 2: 94-107.


Read, D. 2008. Experimental Tests of Rationality. Unpublished, University of

Warwick-Warwick Business School.

Schley, D. 2005. ―Minimized Regret is Sufficient to Model the Asymmetrically

Dominated Decoy Effect.‖ The Marketing Bulletin Vol. 16 Article 2.

Selod, H. 2007. Lecture 1: Preference and choice. Unpublished, Paris School of

Economics.

Office of the University Registrar. 2013. List of registered students – 1st semester A.Y

2013-2014.

Wedell, D. H. and J. C. Pettibone. 1996. ―Using Judgments to Understand Decoy Effects in Choice.‖ The Journal of Organizational Behavior and Human

Decision Process Vol. 67 No.3: 326-344.


1 Indonesian Oil Palm Research Institute, 51 Brigjend Katamso Medan North Sumatera Indonesia 20158 Phone:

062-061-7862477 email : [email protected].


[email protected]


Email Address: [email protected]

4 School of Environmental Science and Management, University of the Philippines Los Baños

Abstract

Oil palm smallholders on peatlands have contributed significantly to economic

development in rural areas by augmenting income and reducing poverty. However, these

plantations also cause adverse environmental impacts such as carbon emission, haze and

peat fires, deforestation, water supply disruption and biodiversity loss. The objective of

the study is to determine the economic impacts of developing smallholder oil palm

plantations on peatlands in Siak District Riau Province, Indonesia. Cochran sampling

technique was employed to select the respondents. Cost benefit analysis was used to

determine the economic impacts of the smallholder oil palm plantations. Results reveal

that the development of 94,726 ha oil palm smallholder plantation on peatlands in 2014

had generated an estimated 37,326 jobs and increased the average total income of

smallholder households to US$ 4,556 per year with a multiplier effect estimated at 3.01

for the Siak economy. Total benefit from the 94,726 ha oil palm smallholder plantation

was computed at US$ 2,152 million per year. However, the unsustainable oil palm

cultural practices of smallholders have led to negative environmental effects. It was

estimated that approximately US$ 1,116 million is lost per year due to the adverse

environmental impacts such as carbon emission, deforestation, water supply disruption

and biodiversity loss, among others. The results of the economic analysis show NPV,

BCR and EIRR to be equal to US$ 1,036 million, 1.93 and 21.91%, respectively. These

results indicate that smallholder oil plantations on peatland in Siak provide net economic

benefits for Siak‘s economy. Proposed policies include the encouragement of sustainable

oil palm plantations characterized by a synergistic relationship among legal, social and

financial aspects in order to provide optimal economic impacts to communities and

minimize adverse effects on the environment.

Keyword: oil palm, smallholders, economic impacts, Indonesia

Introduction

Oil palm is one of the strategic agricultural commodities of Indonesia, serving as

one of its economic pillars (Goenadi 2008). It is a major contributor to job generation, increasing income and promoting economic development and reducing

poverty incidence in the rural areas (Syahza 2012; Wahyunto et al. 2013).

Increased demand for palm oil in the world market has attracted big companies and smallholders alike to invest in oil palm plantations. This has increased the

demand for land for such purpose. Peatlands can be an alternative site for oil palm plantations as long as technical conditions are met and are financially feasible for oil

palm cultivation (Rahutomo et al. 2008).

Economic Impacts of Smallholder Oil Palm (Elaeis guineensis Jacq.)

Plantations on Peatlands in Indonesia

Muhammad Akmal Agustira1, Roberto F. Rañola Jr.2, Asa Jose U. Sajise3 and

Leonardo M. Florece4

106 Agustira, Rañola, Sajise and Florece

The development of oil palm plantations on peatlands, however, has various adverse effects on the environment. Currently, there is an ongoing debate as regards

its main impacts on the environment and the economy, which in turn affect the welfare of the communities. The primary issues against oil palm plantation on

peatlands include the significant carbon stock and greenhouse gas emissions, tropical

peatland deforestation, biodiversity loss, and fire, air, and water pollution (Norwana et al. 2011). Among these issues, greenhouse gas (GHG) emission is considered the

main concern, as peatlands are capable of storing large quantities of carbon and thus

can potentially emit large amounts of GHG, which contribute to global warming and climate change (Schrier et al. 2013). The deforestation of peatland forests in

Indonesia is also being blamed on the development of oil palm plantations (Hooijer et al. 2006). In addition, the conversion of peatlands to oil palm plantations can affect

hydrology and water storage such as soil subsidence, flood, and salt water intrusion

(Page et al. 2010; Silvius et al. 2000). It can also lead to the loss of ecosystem services and biodiversity (Koh and Wilcove 2009). It also causes air pollution from

haze, resulting from forest and peat burning during land preparation, that affect

human health (Tacconi 2008).

There are, however, concerns regarding the development of oil palm plantations

considering that the economic benefits to the country in general and the communities in particular might be attained at great environmental costs from carbon emissions,

deforestation, water supply disturbance, floods, air pollution (haze) and biodiversity

losses (Obizinki et al. 2012). Moreover, these adverse impacts are not limited to the locality but also have effects at the regional and global levels (Schrier et al. 2013).

Peatland conversion loss is likely to cause greater losses than gains from oil palm plantation development (Obizinki et al. 2012). Hence, oil palm development should

consider the environmental aspects in order to minimize environmental impacts and

achieve sustainable economic development. This study attempts to determine the economic costs and benefits of smallholder oil palm plantations on peatlands in

Indonesia.

Methods of Analysis

The study was conducted in the province of Riau, which has the largest peatland

area in Indonesia. Using purposive sampling, Siak district was chosen because it has the largest area devoted to smallholder oil palm plantations on peatland in Riau. Both

primary and secondary data were used in this study. Based on the data from the

Forestry and Estate Agency of Siak, there were 134,178 ha of oil palm plantations on

peatlands in Siak in 2013, of which 70.6% (94,726 ha) were smallholders‘

plantations (Table 1). The two kinds of oil palm developers on peatlands in Siak are the dependent/plasma/supported and independent smallholders. Dependent or

supported smallholders are those who participate in the government‘s oil palm

plantation development programs that may be implemented through a system of partnerships with plantation companies. Independent smallholders are those who

develop their plantations through their own efforts; they self-finance, manage, and

equip their plantations and do not transact with any of the palm oil milling

companies.


5 Planting year is the year oil palm is planted by farmer/smallholder respondents (they planted during the period

1998-2010).

Cochran Sampling Technique was employed in selecting the smallholder-respondents:

where:

Table 1. Area and number of oil palm smallholder plantations on peatlands in

Siak, 2013

Based on the Cochran sampling technique, a total of 273 respondents were

selected for the study. Cost Benefit Analysis (CBA) was used to evaluate the gains

and losses from oil palm plantations in peatlands. Stratified sampling was used to choose smallholder respondents that planted oil palm over different periods5 and

others who have immature crops (2011-2014).

Economic Gains

Production Value

Economic gains generated from oil palm production (fresh fruit bunch) value

was measured using:

PVffb: ((AQ x P) –TC) x A

where :

N : Sample size

P : Proportion of P independent smallholders

Q : Proportion of Q group dependent smallholders

D : Acceptable samples error (5%)

No. Sub District Area

(ha)

Number of Smallholders

Dependent Independent Total

1 Siak 2,398 374 293 667

2 Sungai Apid 2,484 165 684 849

3 Bunga Raya 13,903 934 3,634 4,568

4 Tualang 16,696 - 4,540 4,540

5 Dayub 18,012 2,299 3,825 6,124

6 Mempura 28,049 966 4,540 6,343

7 Sungai Manday 5,508 143 1,289 1,432

8 Lubuk Dalam 5,853 - 420 2,066

9 Sabah Auh 550 - 196 196

10 Pusako 1,273 1,494 572 420

Total 94,726 6,375 20,830 27,205

PVffb = Production value

AQ = Annual productivity (ton/ha)

P = Price ( US$/ha)

TC = Total Cost (US$)

A = Area ( ha)


Regional Multiplier Effects

Economic Impact Analysis (EIA) examines the effect of an event on the

economy in regional specified area. It measure changes in business revenue, profit, personal wages and jobs. It is applied to estimate all of the impacts including direct,

indirect and induced effect in terms of regional multiplier effect (Weisbrod and

Weisbrod 1997). The impact on regional development can be measured (Syahza

2012) as:

k = (1/((1-(MPC x PSY)))

Where:

K = Economic multiplier effect in the area

MPC = Marginal propensity to consume represents income spent by

smallholders in the local the area

PSY = Percentage of farm input required by smallholders oil palm

plantation that be served from local area

Economic Losses

Carbon Emissions

Estimates of economic losses due to carbon emissions are based on the condition

of the technical culture used by smallholders (Table 2). The condition of the technical

culture determines the amount of carbon emissions released.

Table 2. Amount of carbon emission released based on the condition of the

technical culture (ton CO2/ha/year)

Source: Rahutomo et al. (2008)

The benefit transfer method was used to estimate the economic losses from

carbon emissions. In this study, the valuation of economic losses is based on the price of CO2 emission equal to US$ 4.9 per ton, in accordance with the ecosystem market-

place in 2014 (Bloomberg Business 2014),

CEV = CER x SCP

Sampling

Time

Water Table

(cm) Type of Cover Crop

Depth of

Peatland Compaction

50-60 >100 Mucuna

Bracteata

Mixed

Cover

Crop

Without

Cover

Crop

Deep Shallow With Without

Morning 34.50 66.35 32.90 26.16 65.77 65.77 39.62 26.16 37.46

Aftenoon 39.88 79.20 12.60 46.96 60.70 60.70 30.19 46.96 74.94

Evening 26.67 54.59 43.20 28.55 65.38 65.38 31.89 28.55 85.21

Average 33.68 66.71 29.57 33.89 63.95 63.95 33.90 33.89 65.87


where:

CEV = Carbon value

CER = Carbon release (ton CO2/ha/year)

SCP = Price (US$ /ha)

Deforestation

The economic losses due to deforestation based on the potential stumpage value

of peat forests were estimated using the formula:

StV = VDS x V x P

V = ½. Π. d2.h. Cf

StV = Stumpage value (US$)

VDS = Vegetation density structure (population)

V = Volume (m3)

P = Standard price in the market

d = Diameter (m)

h = Height (m)

Cf = Coefficient factor

Water supply

Degradation of peatlands‘ environment causes a disruption in the hydrological

system as manifested in the decreasing availability of water during the dry season and floods during the rainy season. The equation for estimating decreasing water

availability and flooding is (Widodo and Bambang 2010):

LEV = ([ETcop – ETcf] x P x A)

ETc = Kc x ETP

where:

LEV = Loss of environmental economic value due to hydrological system

disruption

ETcop = Oil palm evapotranspiration coefficient

ETcf = Forest evapotranspiration coefficient

P = Price of water (US$/m3)

A = Area

Kc = Crop coefficient


Air Pollution Index Diagnosis

0 - 50 Good

51 – 100 Moderate

101 – 200 Unhealthy

201 – 300 Very unhealthy

301 – 500 Dangerous

Health (haze)

Exposure to haze has an impact on health, such as upper respiratory tract

infection (URTI), asthma, bronchitis, painful and watery eyes, chest pains and skin allergies. The health costs were estimated based on the Cost of Illness (COI) that

included the treatment cost and estimated workday lost. Losses from illnesses caused

by haze were estimated using dose response function methods employing the Air Pollution Index (API) taken from Air Monitoring Service data covering the periods

from January – December 2014.

Table 3. Air pollution index indicators

The total treatment cost was estimated using the formula:

TCTST = (NT x PT) + (NST x PST) (1)

NT = ∑i CHLi x DRC1 x HDi x POPi /10,000 (2)

NST = ∑i CHLi x (DRC1 + DRC2 ) x POPi /10,000 HDi x F1 x F2 (3)

where :

Source: Othman and Shahwahid (1999)

NT = incremental number seeking treatment in the area (person)

NST = incremental number seeking self-treatment or directly buying medicine in the area (person)

CHLi = difference between the average haze index in state I and the normal haze index of 25

DRC1 = dose response coefficient per 10,000 population for the number of hospitalized cases in public hospitals

DRC2 = dose-response coefficient per 10,000 population for the number of

outpatient treatment cases in public hospitals

HDi = number of hazy days in area (days)

F1 = factor of those seeking outpatient treatment in the area

F2 = factor of those seeking self-treatment in area

POPi = population of those seeking self-treatment in area i

TCTST = total cost of treatment and self-treatment (US$)

PT = price of outpatient treatment and medication (US$)

PST = shadow cost of self-treatment (US$)


A loss in productivity was estimated using the formula:

where :

TPLI = total productivity losses from illness (US$)

TNWDL = total workday lost (days)

W = average wage per employee (US$)

NWDL = incremental number of workdays (days)

NSL = incremental number of days of sick leave to adult outpatient (days)

NRAD = incremental reduced activity (days)

NA = incremental number of patients hospitalized (person)

AAR = percentage of adult patients admitted to hospital (%)

LH = average length of stay in hospital (days)

ATR = proportion of adults seeking treatment (%)

MCR = proportion of the proportion of outpatients seeking treat-ment and obtaining sick leave (days)

NRAD = number of workdays lost by workers at risk ( days)

LRA = number of reduced productivity days experienced by indi-viduals at risk (days)

F3 = factor for reduced productivity for individuals at risk but still working

NDA = total number of days of hospital admission throughout the country (days)

CA = incremental cost of hospitalization (US$)

PH = price of hospitalization per day (US$)

Biodiversity

The estimated value of biodiversity loss using the benefit transfer method was

US$ 30 per hectare (ISAS cited in Tuccony et al. 2003). This value, however, is not

fully reflective of the real loss due to the difference in local conditions.

TPLI = T N W D L x W (4)

TNWDL = TNWDL = NWDL + NSL + NRAD (5)

NWDL = NWDL = NA x AAR x LH (6)

NSL = ATR x NT x LMC x MCR (7)

NRAD = (NT + NST) x ATR x LRA - NWDL - NSL) x F3 (8)

NA = ∑iCHL + DRC2 + HDi x F2 x Popi / 10,000 (9)

NDA = NA x LH (10)

CA = NDA x PH (11)


Estimated Total Economic Value

TEV = { (∑OP + ME ∑OP} - (∑EC + ∑DF + ∑PF +∑WS + ∑BD)

The assumptions used in this analysis are:

1. Economic analysis of developing oil palm plantation on peatlands covered a

25-year period.

2. The land area used in the analysis is the whole of oil palm smallholders‘ plantation

areas on peatlands in Siak covering 94,726 ha.

3. The quantifiable benefit was based on the net benefit value from developing oil

palm plantation on peatland and its multiplier effect.

4. The economic cost include the social cost of carbon emission, peatland fire on

health, loss in farm productivity due to illness, deforestation, water supply, and

biodiversity loss.

5. The official exchange rate in 2014 was approximately IDR 12,000 per US $ while

the foreign exchange premium was IDR 20% .

6. The social opportunity cost of capital in Indonesia is 12%.


The oil palm development program of the Siak government aims to boost

economic growth and improve the welfare of society, especially in rural areas. Loss of

natural forest resources caused by illegal logging contributes to the increase in poverty.

The Poverty, Ignorance Eradication and Infrastructure programs (PIEI) aims to

alleviate poverty in rural areas through the development of oil palm plantation.

The average area of land owned by the 273 smallholder-respondents was 3.04 ha, with values ranging from 1.5 to 9 ha. Smallholders acquire peatland areas for oil palm

plantations through various means, most of them by purchase (56.41%). The information provided by key informants indicates that the land acquired through

purchase include lands with expired concessions and industrial timber plantation, as

well as degraded peatlands due to illegal logging.

TEV = Total economic value (US$)

OP = Oil palm production value (US$)

EC = Emission carbon loss (US$)

ME = Multiplier effect

DF = Deforestation loss (US$)

PF = Peat fires and haze loss (US$)

WS = Water supply disruption loss (US$)

BS = Biodiversity loss (US$)


Ambiguity and obscurity in the government policy for Regional Spatial Plan allowed the unauthorized sale of peatland forest areas to the smallholders. The second form of

acquisition is through land conversion of paddy fields, rubber plantations, and other crops (12.82%). Other forms of acquisition include forest clearings (9.89%), expired

concession and industrial timber plantations (6.96%) and public forest area (5.49%)

(Figure 1).

Average production was 40.14 tons per smallholder. Hence, the average

productivity (planting year 4-14 years) was 13.60 tons per ha per year.

Economic Gains

Oil palm development programs provide economic benefits to the communities and surrounding areas. The economic gains from developing oil palm plantation on

peatlands include generation of new employment, improvement in income and well-being of rural communities, and hastening of the multiplier effects of the additional

economic activities.

Potential Employment

Activities related to oil palm development involve a lot of labor. Employment is

possible because oil palm smallholders generally carry out their activities manually.

Oil palm smallholders in Siak do not perform all operational production activities but

hire laborers from outside the plantation.

Results of the employment requirement analysis show that peatlands in Siak generated employment for 37 persons (Table 4) with an employment coefficient of

0.44/ha. It means that 2.27 ha of oil palm plantation will generate employment for 1

person. Hence, oil palm smallholders provided the largest share of employment in

Siak at 20.50%.

Figure 1. Acquisition of peatlands for oil palm cultivation of smallholders in Siak, 2014


Table 4. Estimate of employment generated in smallholder oil palm plantations

on Peatland in Siak, Indonesia, 2014

Income of Smallholder Households

The average total income of smallholder households was approximately US$

4,556 per year in 2014. Income from oil palm constitutes a very large percentage of

the total family income. Based on the analysis of the structure of smallholders' income, the average contribution of income from oil palm to the total household

income is 74.40%. Average smallholder income from oil palm plantations was estimated to be US$ 3,452 per year, which is 72.03% of the 2014 per capita Gross

Regional Domestic Product (GRDP) of Siak amounting to US$ 4,793 per year. It is

higher than the GRDP, however, regardless of whether it comes from oil and gas, by

119.03% or approximately US$ 2,900 per year.

Multiplier Effects

The development of oil palm plantations on peatlands has had enormous economic impacts on rural development in Siak due to the multiplier effects of the

additional income. The development of oil palm plantations has generated a lot of jobs for the surrounding communities and the emergence of business opportunities

such as eateries, convenience stores, transportation, workshops, household industries,

banking services and other services. All these have eventually led to the emergence of the market in residential and rural areas, thus increasing income and improving social

welfare.

Item

Manpower Requirement

(ha/year)

Employment

Mandays

(Person/year/ha)

Manpower

(Person/year/ha)

Smallholders

Nursery

Land clearing

Immature crops

Mature crops

Harvesting

Transportation

-

3.161,00

66

83

34

-

-

-

11.00

0.22

0.28

0.11

0.075

-

27.205

446

322

530

2.018

5.778

1.027

Total (person)

Oil Palm Smallholders on Peatland Area (ha)

Employment Coefficient

Total Employment in Siak, 2013

Contribution of Oil Palm Smallholders to Siak Employment

37.326

94.726

0.44

182.059

20.50%


The computation of the multiplier effect is based on household income spent by smallholders in the local region (MPC) and the needs of the oil palm plantation

activities that can be met in the local area (PSY). Smallholders generally used the proceeds of fresh fruit branch (FFB) for household expenses, while revenues from

other business sources were used as savings or investments in oil palm cultivation.

The average expenditure of smallholder households was pegged at US$ 2,644 per year. The value of the multiplier was computed at 3.01, which means that every US$

100 spent by oil palm smallholders will generate an additional amount US$ 301 from

auxiliary services.

Problems of Smallholder Oil Palm Plantations in Siak

While the development of 94,726 ha smallholder oil palm plantations on

peatlands in Siak may provide enormous economic benefit for Siak‘s economy, the

results also revealed that there are accompanying problems related to their

development. These are as follows:

1. Most oil palm smallholder plantations are on peatlands.

Peatland area in Siak comprises 53.94% (461,527 ha) of total area and is still available for oil palm development. However, among others, there are many

environmental challenges in developing them such as carbon stock and greenhouse

gas emissions, tropical peatland deforestation, biodiversity loss, and fire, air, and

water pollution.

2. Lack of suitable peatlands for oil palm cultivation.

There are 159,890 ha, which is 34.64% of total peatland areas in Siak, with a peat depth of less than 3 meters that is suitable for oil palm cultivation. The rest of the

area with a peat depth of at least 3 meters are no longer suitable.

3. Lack of knowledge and low adoption of appropriate cultural practices, as well as

lack of funding.

The application of Best Management Practices for sustainable palm oil production on peatlands is crucial for reducing its negative environmental impacts. In

this study, 10 indicators were used to determine whether farmers were adopting

sustainable palm oil management practices. This are shown in Table 5.

4. Low productivity

Lack of knowledge, low adoption of recommended cultural practices and lack of

funding are the major reasons for the low productivity in smallholder farms. Results

show that the average farm productivity was 13.60 tons per ha per year which is only

55.93% of the potential standard productivity. Despite this, smallholders perceive that oil palm production on peatlands is still profitable, thus they continue to expand

peatland areas for oil palm cultivation.

Under these conditions the major challenge in the production of smallholder

plantations in Siak is how to address the adverse impacts on environments. Results

show that only 44.69% of smallholder-respondents applied the best management

practices (BMPs) (Table 5).


Table 5. Number of smallholder respondents who implemented the sustainable

oil palm plantations in Siak, 2014

Economic Losses

While there are economic gains from smallholder oil palm plantations on

peatlands, there are also economic losses related to the degradation of environment

and losses in social welfare.

Carbon Emissions

Economic losses from carbon emissions depend on the cultural practices. Results show that the estimated carbon emissions released by smallholder oil palm

production is 48 – 66 tons CO2/ha/year. Based on information from ―ecosystem

marketplace‖ (Blomberg Business 2014), the price of CO2 emission is US$ 4.9 per ton. The estimated economic costs of carbon emission based on the ecosystem age of

the oil palm crops are listed in Table 6.

Table 6. Estimated economic losses from carbon emission released by oil palm in

Siak, 2014

Indicator Applied Did Not Apply Total

No. % No. % No. %

Identification of land suitability

zero burning

Using high yield planting

Material

Compaction

Water management

Balance fertilization

Integrated pest management

Using cover crop

Road maintenance

Prevention and control fires

135

144

122

118

95

25

21

24

269

267

49.45

52.75

44.69

43.22

34.80

9.16

7.69

8.79

98.53

97.80

138

129

151

155

178

248

252

249

4

6

50.55

47.25

55.31

56.78

65.20

90.84

92.31

91.21

1.47

2.20

273

273

273

273

273

273

273

273

273

273

100

100

100

100

100

100

100

100

100

100

Average Implementation 44.69 55.31 100

Age of Oil Palm

Estimation of Emission Carbon

Released

(CO2/ha/year)

Economic Cost of

Carbon Emission

(US$/ha/year)

0-3 48.69 238.60

4-9 56.19 275.34

10-15 62.73 307.36

16-25 66.30 324.87


Deforestation

The condition of vegetation density structure of plants in the peat forest in

research areas were determined by the Regional Planning Agency in 2013 (Table 7). Standard Price of Mix Forest Provision, is USD 80.25 per m3 (Regulation No. 22,

Series of 2012 of Indonesia‘s Minister of Trade). This standard refers to the

calculation basis of the forest resources provision that surcharges are imposed as a substitute for the intrinsic value of forest products harvested from state forests. Thus,

the economic value is US$ 162.57 per ha.

Table 7. Plant vegetation density structure for four research areas in Siak, 2014

Palm oil cultivation accounts for 9.89% of total peatland deforestation in Siak (Figure 1). Hence, during the 15 years of oil palm development (1998-2014), the

economic losses due to deforestation was estimated at US$ 101,535,477 per year

(Table 8).

Table 8. Estimated stumpage value per hectare of deforestation in Siak, 2014

Water supply

The decline in water supply is one of the economic losses that are attributed to

the development of oil palm plantations, particularly, during the dry season (June,

July, August) when there is a deficit of 50 mm/ha/year. (Widodo and Bambang 2010) (Figure 2). This means that as much as 500 m3/ha/year of water is lost for every 1 ha

of oil palm plantation developed.

Research Area Population/ha

Tree Pole Pile Tiller

Merempan Hulu 100 250 383 225

Dayun 65 180 245 245

Bunga Raya 27 187 262 187

Sungai Mandau 120 195 180 190

Average 78 203 268 212

Plant Vegetation

Structure Volume/ Tree Population/ha Total Volume

Tree 44.31 78 3,456

Pole 16.88 203 3,427

Pile 4.40 268 1,178

Tiller 0.20 212 42

Total Volume (m3) 2,026

Price (US$/m3) 80.25

Stumpage Value (US$/ha) 162,569

Estimates of Deforestation due to Oil Palm (ha) 9,368.51

Economic Value (US$) 1,523,032,155

Economic Value Per year (US$) 101,535,477


Based on the data obtained from the Regional Water Company Siak Tirta, the price per m3 of water is US$ 0.402. Thus, given a 94,726-ha of peatland areas

developed to oil palm plantations by smallholders, the estimated value from the

reduction in water availability is US$ 19,028,085 per year.

Peat Fire

The economic losses from peatland fires consist of the cost of treating

haze-related illnesses, the relief fund from government to assist victims of these fires and loss of productivity of these smallholder farmers. Based on the data from the

Health Agency of Siak Government (2014), the number of persons who contracted

haze-related sickness was 76,570 persons.

Table 9. Number of persons who suffered from illnesses related to peatland fires

The total economic costs from peatland fires is estimated at US$ 3,952,714 per year. This consists of treatment costs valued at US$ 2,647,271, the relief fund valued

at US$ 833,333 per year that is provided to the Budget of Regional Disaster Board for

Disaster Management for Haze Catastrophe by the Anggaran Belanja Pendapatan Belanja Daerah (APBD)/Regional Government‘s Revenue and loss in farm

productivity valued at US$ 1,305,444 per year.

Source: Widodo and Bambang (2010)

Figure 2. Water table (a) before oil palm plantation and (b) after oil palm plantation in the research site

in Siak, 2010

Illness January February March April May –

December Total

URTI 5,030 32,861 28,354 534 - 66,779

Pneumonia 242 563 617 12 - 1,434

Asthma 342 1,169 1,193 32 - 2,736

Eye irritation 175 929 1,234 9 - 2,347

Skin irritation 110 1,564 1,562 38 - 3,274

Total 5,899 37,086 32,960 625 - 76,570 Source: Public Health Agency of Siak District


Biodiversity

Using the benefit transfer method, the estimated value of the biodiversity loss

from oil palm development is estimated at US$ 30 per hectare. This value is based on the study of ISAS (cited in Tuccony et al. 2003). Considering the difference in

biodiversity, the value may not be exactly accurate but it can be reflective of potential

value of the biodiversity lost. Given therefore a 94,726 ha peatland area, the estimated economic loss due to the conversion of these areas into oil palm plantations is

US$ 2,841,780 per year.

Table 10. Estimated total economic losses caused by peat fires in Siak, 2014

Economic Viability of Smallholder Oil Palm Plantation on Peatlands

Table 11 shows that with an NPV of approximately US$ 1,036, 303,250, a B/C Ratio of 1.93 and an EIR of 21.91, the social benefit of smallholder oil palm

plantations on peatlands is higher than the social cost of the adverse environmental

impacts from developing smallholder oil palm farms on peatlands. This indicates that smallholder oil plantations on peatland in Siak leads to a positive net economic

benefit.

Table 11. Economic analysis of gains and losses of smallholder oil palm

plantations on peatlands in Siak, 2014

Item Economic Losses

(US$/Year)

Social cost per hectare

(US$/Year)

Treatment Cost 2,647,271 27.95

Cost of treating illness 1813,937 19.15

Disaster relief fund 833,333 8.80

Productivity 1,305,443 13.78

Total 3,952,714 41.73

Item NPV

(US$)

NPV

(US$/ha)

Benefit

Net benefit (oil palm) 326,057,298 3,442

Multiplier effect 1,825,940,193 19,276

Total Benefit 2,151,997,492 22,718

Cost

Carbon emission 203,700,926 2,150

Health 20,901,948 221

Incremental productivity from illness 10,307,339 109

Deforestation 708,107,101 7,475

Biodiversity losses 22,437,728 237

Water supply 150,239,290 1,586

Total Cost 1,115,694,242 11,778

NPV 1,036,303,250 10,940

B/C 1.93

EIR (%) 21.91


Proposed Development Policies for Sustainable Smallholder Palm Oil

Plantations on Peatlands

Sustainable methods of production on peatlands should be adopted by small-holder farmers to mitigate the adverse effects of oil palm plantation development such

as carbon emissions, soil subsidence, peatland fire, biodiversity and deforestation.

However, results show that only 44.69% of smallholder-respondents apply best management practices (BMPs). The main reasons as mentioned earlier are the lack of

technical information and awareness of the appropriate cultural practices as well as

lack of funding. In addition, there are issues related to uncertainty in regional spatial plan (Rencana Tata Ruang Wilayah (RTRW), lack of law enforcement, the slash and

burn method of land clearing which often trigger peatland fires, and cultivation of

forest area reserves for palm oil plantations.

Given these concerns, there is a need for a policy on sustainable oil palm

development on peatlands that will consider the legal, social and financial issues that will enhance the economic benefits to the communities while minimizing the adverse

impacts on the environment (Figure 3). This will include the establishment of RTRW and strengthening of law enforcement to encourage oil palm plantation development

on peatlands. It would be best also to divide the oil palm plantations into zones where

some agricultural commodities with good market prospects can be established to support the region's economy. Furthermore, prior information is needed to ensure that

the smallholder oil palm plantations are not in the peat forest areas. In addition, there

is a need to raise social awareness as well as capacitate smallholder institutions to ensure compliance with sustainable development of oil palm plantations. This will

involve capacity building programs based on Indonesian Sustainable Palm Oil (ISPO) and Best Management Practices (BMP). Lastly, there is a need to provide incentives

for smallholders on the implementation of sustainable oil palm plantations and

preservation of ecosystems. In addition, financial incentives such as a premium price for FFB can be granted to smallholders who implement sustainable oil palm

production based on the standards set by the Indonesian Sustainable Palm Oil (ISPO).

Figure 3. Policy for developing smallholder oil palm plantations on peatlands


Funding of these activities can be performed by various sources including the government, companies, organizations/donor countries, and market mechanisms

through a premium price. Funding can be conducted by a direct aid system and payment for ecosystem services. Direct funding is provided by the government to

stimulate the development of palm oil industry. Funding is directly sourced from

Regional/National Government Revenue and Expenditure (Dana Anggaran Belanja Pendapatan Belanja Daerah/Negara (APBD/APBN), share from export tax, grants

from donor countries for preservation of peatlands and development of sustainable oil

palm production and Corporate Social Responsibility programs that support the

preservation of peatlands and development of sustainable oil palm plantations.

In addition, funding can be sourced from payments for ecosystem services, which serve as compensation to landholders for their opportunity cost (the forgone

earnings from land use). These externalities are likely to persist in time, so long-term

compensation is required (Wunder et al. 2008).

Design and implementation of ecosystem payments can be conducted through

the following:

1. Incentives in the form of withholding tax (income tax, CPO export tax, property

taxes) and incentives that ease the process of licensing can be done for plantation

companies that help fund peatland conservation programs and partnerships to support the development of sustainable oil palm smallholder plantations on

peatlands through CSR programs.

2. Incentives to smallholders for the implementation of sustainable oil palm planta-tions and preservation of ecosystems. The incentives may include fertilizer subsi-

dies, priority construction of access roads and garden infrastructure and interest

subsidies.

3. Market mechanism through increased value of the price (premium price) for FFB

produced by planters who implement sustainable palm as standardized by ISPO.

4. Payment system of ecosystem services through carbon fund. This policy can be

initiated through the implementation of programs that can successfully reduce the amount of carbon emissions through the development of sustainable palm oil

cultivation. However, there are obstacles to this, which include the following:

a) It requires commitment from organizations and donor countries.

b) Strict rules on the international mechanisms system of ecosystem service

payment. This is an obstacle faced by prospective recipient countries with

donor funds due to differences in the perception of rules and mechanisms of

carbon payment system (Sterner and Jessica 2002).

c) NPV of oil palm plantation is much more favorable than the NPV of carbon fund. Based on financial analysis, NPV of smallholders oil palm plantation

gives US$ 3,442.11 ha/year, while the NPV from carbon emission fund

receives US$ 1,269.48 ha/year.


Conclusions and Recommendations

Smallholder oil palm plantations on peatlands provide enormous net economic

benefits for Siak‘s economy. It is therefore important to reconsider the moratorium on oil palm plantations on peatlands especially for smallholder oil palm plantations

development. The considerations are follows:

1. The economic benefits from oil palm plantations are crucially important for indigenous peoples in remote areas that have limited sources of income. With the

degradation of peatlands and deforestation, these people are losing their source of

livelihood. Oil palm can be an alternative source of livelihood as well as the

agent of economic development in these areas.

2. As a result of the moratorium, there are illegal conversions of peatlands into oil

palm plantations particularly on the independent smallholdings which is causing

the widespread degradation of peatland areas with the consequent adverse

environmental impacts.

3. Attention should be given to the suitability and environmental aspect of

peatlands. The suitability of peatland areas where oil palm plantations will be established should be carefully considered since not all peatland areas are suitable

for such plantations. By so doing, the potential contribution to the economy can

be maximized while minimizing the adverse impacts on the environment from

peatland degradation.

References

Dixon, J.A. and M.M. Hufshmide. 1986. Economic Valuation Technique for The Environment A Case Study Workbook. London: The John Hopkins University

Press.

Goenadi. 2008. ―Prospective on Indonesian Palm Oil Production.‖ Paper presented at

the International Food & Agricultural Trade Policy Council‘s Spring 2008

Meeting, 12 May 2008, Bogor, Indonesia.

Hooijer, A., M. Silvius, H. Wosten and S. Page. 2006. ―Peat CO2 Assessment of CO2

Emission from Drained Peatlands in SE Asia,‖ 1st Edition. Kuala Lumpur. Delft

Hydraulics Report.

Koh, L.P. and Wilcove.D.S. 2009. ―Oil Palm: Disinformation Enables Deforestation‖.

Trends in Ecology & Evolution Vol. 24 (2):67-68.

Norwana, A. A. B. D., R. Kunjappan, M. Chin, G. Schoneveld, L. Potter and

R. Andriani. 2011. ―The Local Impacts of Oil Palm Expansion in Malaysia.‖ An

Assessment based on a Case Study in Sabah State. Bogor Indonesia: Center for

International Forestry Research.

Obidzinnski, K., R. Andriani, H. Komarudin and A. Andrianto. 2012. ―Environmental and Social Impacts of Oil Palm Plantation and their Implication for Biofuel in

Indonesia‖. Ecology and Society 17(1): 25.

Othman, J. and H. O. Shahwahid. 1999. ―Cost of Trans-Boundary Haze

Externalities‖. Malaysia Economics Journal 33 (1999): 3-19.


Page, S., R. Wust and C. Bank. 2010. ―Past And Present Carbon Accumulation And Loss In Southeast Asian Peatland‖. In Scientific Highlights: Pages News 18(1):

25-35.

Rahutomo, S., R. Winarna and H. Setiawan. 2008. ―The Eight Key Success Factor of

Oil Palm Management on Peat Soil.‖ Medan: Indonesia Oil Palm Research

Institute Journal.

Schrier-Uijl, A., P. M. Silvius, F. Parish, K. H. Lim, S. Rosediana and G. Anshari.

2013. ―Environmental and Social Impacts of Oil Palm Cultivation on Tropical

Peat- Scientific Review‖. Reports from the Technical Panels of the 2nd Greenhouse Gas Working Group of Roundtable on Sustainable Palm Oil

(RSPO).

Syahza, A. 2012. ―Development Impact of Oil Palm Oil Plantation Toward Rural

Economic Multiplier Effect in Riau Province‖. Economic Journal, Th.X/03/

November 2012. pp 12-22.

Tacconi, L., F. Jotzo, and R. Quentin Grafton. 2008. ―Local Cause, Regional

Co-Operation And Global Financing For Environment Problems: The Case Of Southeast Asian Haxe Pollution,‖ International Environmental Agreements;

Politics, Law And Economics‖. 8(1): 1-16

Wahyunto., AI Dariah, Djoko Pitono and Muhrizal Sarwani. 2013. ―Prospect of Peatland Utilization for Oil Palm Plantation in Indonesia‖. Perspective Journal

12(1): 11-22.

Weisbrod, G. and B. Weisbrod. 1997. ―Measuring Economic Impacts of Project and Program. Economic Development Research Group. Northwestern University

Boston, USA, pages 1-11.

Widodo, I. T. and D. D. Bambang. 2010. ―The Estimation of Oil Palm Plantation

Environmental Value Using Crop Evapotranspiration of Oil Palm‖. J. Agromet

24(1): 23-32.

Reviewers for JEMAD Volume 1 Number 2



Nora DM. Carambas, Department of Agricultural and Applied Economics,


Dinah Pura T. Depositario, Department of Agribusiness Management and Entrepreneurship,


Rowena A. Dorado, Department of Economics, University of the Philippines Los Baños

Arnold R. Elepaño, College of Engineering and Agro-Industrial Technology,


Flordeliza A. Lantican, Retired Professor, College of Economics and Management,


Ma. Eden S. Piadozo, Department of Agricultural and Applied Economics,


Zenaida M. Sumalde, Department of Economics, University of the Philippines Los Baños

Kevin F. Yaptengco, College of Engineering and Agro-Industrial Technology,


Rico C. Ancog, School of Environmental Science and Management,


Journal of Economics, Management & Agricultural Development (JEMAD)

The official journal of the College of Economics and Management, University of the Philippines

Los Baños (CEM-UPLB)

About the Journal

The Journal of Economics, Management & Agricultural Development (JEMAD) is published by

the College of Economics and Management, University of the Philippines Los Baños

(CEM-UPLB). The primary objective of JEMAD is to present and disseminate high-quality

research in a forum that is informative and useful to a broad audience of academicians,

practitioners, policy-makers and development workers in the fields of economics, applied

economics, management, entrepreneurship, cooperatives, and agricultural and rural

development. JEMAD is a peer-reviewed journal with the Editorial Board and reviewers coming

from alumni of CEM-UPLB and experts in the above-mentioned fields.

Instructions for authors wishing to submit articles are available online at http://www.

cem.uplb.edu.ph

Please direct all editorial correspondence to the Editor at [email protected] or at

[email protected]

© December 2015

Department of Agricultural and Applied

Economics

Department of Agribusiness Management and

Entrepreneurship

Department of Economics

Institute of Cooperatives and Bio-Enterprise

Development

Editor

Isabelita M. Pabuayon

Associate Editors

Nanette A. Aquino

Amelia L. Bello

Ma. Eden S. Piadozo

Dinah Pura T. Depositario

Marilyn M. Elauria

Agnes T. Banzon

Zenaida M. Sumalde

Rodger M. Valientes

Editorial Advisory Board

Narciso R. Deomampo

Gregmar I. Galinato

Flordeliza A. Lantican

Rodolfo M. Nayga, Jr.

Nguyen Thi Duong Nga

V. Bruce J. Tolentino

Jimmy B. Williams

Managing Editor

Gideon P. Carnaje

Cover and Layout

Marion M. Bueno



College, Laguna 4031

http://cem.uplb.edu.ph

Telefax: (049) 536-4750

Email Address: [email protected]