do macro economic conditions matter for agriculture

8/8/2019 Do Macro Economic Conditions Matter for Agriculture

1/27

Do Macroeconomic Conditions Matter for Agriculture?

The Indian Experience

by

Shashanka Bhide, B.P. Vani and Meenakshi Rajeev*

Abstract

Macroeconomic instability, characterised by high inflation, a fragile foreign exchangeposition and high rates of interest, increases uncertainty for any investor or producer andhence slows down economic growth. While this is generally accepted, the usual perceptionabout the agricultural sector, particularly in India, is that it is immune to general

macroeconomic shocks. In this paper, we intend to examine this perception using a vectorauto regressive model. By studying the significance of macroeconomic conditions to theagricultural sector, we observe that the sector is not insulated from macroeconomicshocks.

1. Introduction

A conducive macroeconomic environment is necessary for rapid economic growth.Macroeconomic instability, characterised by high rates of inflation, a fragile foreignexchange position, high rates of interest, increases uncertainty for any investor and

producer and hence slows down economic growth. Besides these direct indicators,macroeconomic instability may also be indicated by overall imbalances such as the fiscal

balance and external current account balance, especially when prices are underadministrative controls. An underlying assumption in these arguments is that productionsectors are influenced by macroeconomic conditions. Any attempt to examine this

proposition will require identification of the indicators of macroeconomic conditions andof the performance of the sectors.

In the Indian context, the agricultural sector has been important from a policy perspectivefor several reasons. Even from the point of view of accelerating economic growth,

transition from an agrarian economy to an industrial or modern economy would depend onhow well the agricultural sector enables this transition. Therefore, besides the concernsrelating to employment and poverty alleviation, the performance of agriculture is of policyinterest from the viewpoint of accelerating economic growth as well. In this context, thegeneral belief is that overall macroeconomic policies have little effect on the agriculturalsector and that we need sector-specific policies to boost this primary sector of theeconomy. While the latter may be true, the macroeconomic environment may also havenon-trivial effects on the agricultural sector. It is necessary to examine this hypothesismore rigorously as if it does influence agricultural performance, it would be important tounderstand the significance of this impact in designing policies for agriculture.

*Shashanka Bhide is Senior Research Counsellor at National Council of Applied Economic Research, New Delhi; Meenakshi Rajeev isAssociate Professor and B.P. Vani is Assistant Professor at the Institute for Social and Economic Change, Bangalore. The authors wishto acknowledge the research assistance of Ms. G. Aparna in the preparation of this paper.

1


2/27

The key to understanding the impact of changes in macroeconomic parameters at thesectoral level is the transmission mechanism of these policy impulses to the varioussectors both directly and indirectly through inter-sectoral relationships. Such impactassessment is often made in the framework of macroeconomic models in which agricultureand other sectors are featured in detail. Rangarajan (1982) provides one of the early

attempts at estimating the inter-linkages between agriculture and industry. There are otherstudies, such as those of Narayana, Parikh and Srinivasan (1991), Strom (1993) andKalirajan and Bhide (2003), where the impact of macroeconomic policies on agriculture issimulated using economy-wide models for India. Shand and Kalirajan (1999) provideanother approach at capturing these linkages where they examine inter-sectoraldependence through Granger causality tests. Both these approaches capture two-way links:from agriculture to the non-agricultural sectors and vice-versa within an implicit orexplicit specification of a macroeconomic environment. In one of its recent reports, theReserve Bank of India (2002) draws attention to the impact of rising food subsidies onmacroeconomic conditions. However, the analysis is limited to the one-way linkage.

India saw very wide-ranging changes in macroeconomic policies in the 1990s. Thechanges at the macroeconomic level were changes in the fiscal and financial sectors, tradeand investment policies. It has been argued that agriculture was only indirectly affected bythese reforms. The industrial sector was most affected directly by the removal of the

production licensing system enforced through control over new investments. Changes inthe financial sector including exchange rate policies and attempts to stabilize fiscalimbalance, however, can be expected to have an impact across all the sectors. Howimportant was this impact to agriculture?

The economy-wide models, of macroeconomic variety or the CGE type, provide acomprehensive analytical structure for analysis. One limitation of the macroeconomicmodels is of course that much effort is needed to build the structural equations thatincorporate the various inter-relationships and it is often difficult to check for the impactof changes in polices due to changes in the structure. A more flexible approach to theassessment of the impact of impulses emanating from the macroeconomic factors toagriculture is the framework of time-series analysis. We should note at the outset that theVAR approach essentially captures the reduced form relationships among the selectedvariables. Interpreting the estimated linkages in a theoretical framework is not easy

because the theoretical specification is not complete in a VAR.

This paper is an attempt to assess the nature of the inter-relationship between selected

macroeconomic factors and agriculture using the vector auto regression (VAR) approach.The VAR approach provides a general framework for assessing the impact of inter-relatedvariablesi. The general framework of analysis we adopt here is to first specify a set ofvariables that capture the performance of agriculture and a set of variables that specify themacroeconomic environment. The agricultural variables we consider are real agriculturalGDP, agricultural exports and fixed investment in agriculture. The last mentioned variableis measured through the gross fixed capital formation (both public and private) in theagriculture sector. The macroeconomic variables are interest rate, foreign exchange rate ofthe rupee and fiscal deficit of the Central Government. We then use two methods ofquantifying the impact of macroeconomic factors on agriculture. One approach is toestimate the VAR and the impulse response functions. The second approach is the

variance decomposition to quantify the impact of the macroeconomic factors onagriculture. We have used annual data for the period 1970-71 to 2003-04 for the analysis.

2


3/27

This period covers a variety of experiences both in macroeconomic conditions,agricultural performance and policies.

The rest of this paper is devoted to the presentation of the results of analysis anddiscussion of the findings. To provide a context for the analysis that follows, we briefly

review the trends in some of the major macroeconomic variables and discuss the likelyimpact of these changes on agriculture. We then present a discussion of the trends inselected macroeconomic variables, followed by methodology and results of the VARanalysis respectively. A concluding section follows at the end.

2. Conceptual Framework

Indias macroeconomic policies have generally attempted to ensure adequate resources forthe investment programmes of the public sector while maintaining adequate supplies ofessential commodities for mass consumption. Sectoral policies, whether in agriculture orindustry, were cast within the overall framework of macroeconomic goals. Indiasmacroeconomic stabilisation programme of the 1990s that preceded and then overlappedthe structural adjustment reforms aimed at reducing the fiscal imbalance, moderatinginflation, correcting the overvalued foreign exchange rate and bringing down interest rates.The emphasis on public sector investment has changed to investment that is commerciallyviable. While food security and economic growth remain critical objectives, there isgreater stress on maintaining the conducive macroeconomic environment rather than ondirect public investment at the micro level. Do these changes have an impact onagriculture? To attempt an assessment of this question, we will need to identify the factorsthat describe the macroeconomic conditions and the variables that describe performance of

agriculture. In this section, we identify these factors and variables, examine the trends inthese variables over time and provide a discussion of the potential mechanisms by whichthe impact of changes in the macroeconomic conditions is transmitted to the agriculturalsector.

Selection of Variables for Analysis: Of the many indicators of macro-economicconditions, we focus on aggregate market imbalances and aggregate prices to discern theimpact of macro-economic conditions on agriculture. The imbalance that we have chosento reflect macro-economic conditions in this study is the fiscal imbalance1. The threeaggregate prices chosen for the analysis are interest rate, exchange rate and terms of trade.Prices would reflect market conditions fully only if policy measures are not used to control

these prices in the divergence of market forces. In the Indian context, although controlsover prices existed, at the aggregate level, the controlled prices also saw gradualadjustments with respect to market conditions. Besides the potential for direct impact,these variables capturing the macro level imbalances and price conditions also trigger

policies that may have a direct impact on specific sectors. For example, high rates ofinflation bring (captured by terms of trade) to focus the need for ensuring adequatesupplies of essential goods of consumption and therefore greater attention to policies thatraise agricultural growth.

In this sense, the choice of variables to reflect macro-economic conditions should includethe indicators that not only have a direct impact on the performance of agriculture but may

1 We have also considered imbalances that arose in the external current account and did a similar study.Qualitatively results remain the same (see Bhide, Vani, Rajeev,2005) .

3


4/27

also influence agriculture indirectly through policies resulting from macro-economicconditions.

With these considerations, four broad measures of macro-economic conditions selected forfurther analysis in this study are: gross fiscal deficit of the central government (GFDC),

and interest (PLRX), real effective exchange rate (REER) andterms of trade (TT).

To assess the impact of macroeconomic factors or conditions on agriculture, we also needto identify the variables that reflect the performance of agriculture. In this study, weconsider the variables that capture different dimensions of agricultural sector. Agriculturalinvestment, agricultural exports and agricultural GDP are the three variables selected inthe study to reflect the performance of agricultural sector. They reflect the overall output

performance of agriculture and also relate more directly to interest rate, inflation rate andexchange rate changes. Interest rate conditions affect agricultural investment and changesin exchange rate influence exports. The overall agricultural GDP is inter-linked with

prices, measured through terms of trade, and all other factors that influence either the

demand or the supply of farm products.

In most cases, of the above macro-economic indicators, alternative measures areavailable. In the case of interest rate, a number of interest rates are available indicating thewide range of financial markets. We have selected a rate that is a benchmark forinvestment lending by the commercial banks. A combination of the minimum lending rateof IDBI and the Prime Lending Rate of commercial banks (PLRX) was chosen as theinterest rate for analysis in this study. In the case of exchange rate, we have selected realeffective exchange rate of the rupee (REER), which is a trade weighted real exchange rageof 36 major trading partners of India. We have used the wholesale price index (WPI) foragricultural commodities and manufactured products to define a standard measure of termsof trade (TT) as (WPI for agriculture/WPI for manufacturing)*100.

(a) The Macroeconomic Trends:

The key variables that are tracked in this section (Table A.2 in Appendix presents the dataused for the study) include fiscal deficit and the price related variables. We have focusedhere on the fiscal deficit of the Central Government, as the initial correction under thestabilisation program was at this level of government. In addition, we also present thetrends in the terms of trade, real exchange rate (REER) and interest rateii.

The intense pressures of fiscal and external sector imbalances at the time of themacroeconomic crisis of 1990-1991 are well known. The macroeconomic crisis triggeredmany economic policy changes. Two of the key indicators that reflected the crisis were thefiscal and external imbalances.

The macroeconomic adjustment process focused attention on the gross fiscal deficit of theCentral Government although the overall fiscal imbalance is known to be much higherthan this deficit. There were similarities in the two measures in terms of direction ofchange although the relative changes have been different. In the Appendix we present thedata only for the Central government deficit.

4


5/27

Turning to the three price variables in the aggregate markets, the nominal interest rate,PLRX, increased sharply in 1991-92 and remained about 15% up to 1996-97. Since 1996-97 there has been a decline in PLRX. In nominal terms, the decline is by almost 4

percentage points since the high levels of 1995-96. The decline is less marked in terms ofreal interest rate. However, it must be pointed out that there has been some lending below

the PLR by the commercial banks indicating that trends in PLR can only be a crude proxyfor the trends in interest rates in the economy. The drop in interest rates has been a featureof the economy since the mid-1990s.

The real effective exchange rate, REER, saw a major correction in 1991-92 and 1992-93after a steady depreciation for about a decade. Since then there has been a relatively stable

period marked by a tendency towards appreciation. Although controls on external capitalaccount transactions remain, the rupee is sensitive to supply-demand pressures in foreignexchange markets. The large levels of forex reserves moving closer to $90 billion have

led to the strengthening of the rupee.

Indicators of terms of trade (calculated as a ratio of wholesale price index for agriculturalcommodities to wholesale price index of manufacturing products) reflect a relativelystable behaviour during our period of study. The index rose sharply between 1975-76 and1982-83 after which there has again been stagnation up to 1995-96 thereafter we observesome decline. The period of stable terms of trade from the mid-1980s to mid-1990sincludes the period when fiscal imbalances were growing and were high, as well as periodof high rates of inflation. We also note that this has been a period when private sectorcapital formation in agriculture was rising. The years since 1995-96 up to 2002-03include a period when non-agricultural investment was also on the decline and overallinflation rate decreased especially after 1998-99. In other words, agricultural prices havekept pace with the overall inflation rate especially when the inflation rate has beenrelatively high. Macro-economic instability, which included high rates of inflation, wasalso characterised by higher growth in agricultural prices.

(b) Performance of Agriculture:

The trends in key agricultural variables in this analysis are illustrated in Figures 1, 2 andTable A.2. Figure 1 shows the gross fixed capital formation (GFCF) in agriculture both

public and private sector combined. A disaggregated analysis of this data reveals thatprivate agricultural investment (gross fixed capital formation in constant prices), indeedrose sharply between 1987-88 and 1990-91, became stagnant for the next three years till1993-94. It rose steadily again till 1998-99 after which it remained at the same level in thesubsequent year. The years 1991-92 to 1993-94 formed a period when the macro-economic parameters were unstable reflecting the adjustments in policy. The subsequent

period was marked by a few years of strong growth in industry and a climate favourablefor investment. This period also appears to have influenced private sector investment inagriculture. However, the public sector capital formation in agriculture has continued tostagnate for well over a decade. Fiscal pressures on the one hand and preference forsubsidies have led to stagnation in Government spending on investment in agriculture.

The impact of adverse macro-economic conditions on investment may have also affectedin particular public investment in agriculture.

5


6/27

Figure 1. Investment in Agriculture (GFCF) (Rs. Crore 1993-94 prices)

Note: Rs 1 Crore= Rs 10 million.

Agricultural exports increased sharply during the period 1988-89 to 1996-97 (Figure 2).Exports in value terms declined to some extent from 1996-97 onwards but again increased

in the current decade. The latter decline is attributed to a decline in the unit value ofexports during the period when global commodity prices also experienced a decline. iii

Macro-economic factors primarily relating to the real exchange rate would have an impacton exports, including agricultural exports. The decline in exports has occurred during a

period when the real exchange rate has been relatively stable or slightly appreciating.

Figure 2. Agricultural Exports (in Rs crores, 1993-94 prices)

6

0.0

5000.0

10000.0

15000.0

20000.0

25000.0

Year

0.0

1000.0

2000.0

3000.0

4000.0

5000.0

6000.0

year


7/27

T rends in agricultural GDP trend at constant 1993-94 prices show a steady increase overthe years, except for a slight fall in a few years (Table A.3 in Appendix).

3. Methodology for Assessing the Impact of the Macroeconomic Factors

The broad trends in the macroeconomic factors and measures of agricultural performanceindicate fluctuations and changes in pattern over a long period of over three decades(1970-71 to 2004-05). In the case of macroeconomic variables, the trends reveal upwardand downward movements in fiscal deficit, downward movement of terms of trade,correction in exchange rate and drop in nominal interest rate. How would these changes

have an influence on agriculture?

The policy channel: The mechanisms through which changes in the macroeconomicvariables are transmitted to agriculture are several. At a general level, macroeconomicimbalances reflected in the levels of fiscal are characterised by their composition as well.Lower fiscal deficit may be achieved by expenditure compression or revenue expansion.The manner in which imbalances are realised may also have an impact of its own. Besidesthese composition effects, the twin deficits have an impact on aggregate prices: terms oftrade, interest and exchange rates. More importantly, the imbalances also lead to policyresponses such as controls on credit availability, access to markets and quality ofgovernment services each of which affects all producers, including agriculture.

The investment channel: The market signals induced by macroeconomic changes have animpact on investment. Changes in nominal interest rate and output prices affect realinterest rate, which in turn influences investment decisions of farmers. Further, Changesin terms of trade may imply differential changes in sectoral price indices, which in turneffect investment decision of a farmer. In other words, there may be changes in the termsof trade not only due to the changes in prices resulting from structural factors such as thechanges in tariff rates but also due to differences in the speed with which different pricesadjust to the macroeconomic shocks. In addition, poor fiscal conditions also affectgovernment spending on investment projects.

The export channel: The third mechanism is the impact of changes in exchange rate onagricultural exports. Changes in real exchange rate influence the competitiveness ofexports in general and hence agricultural exports as well. Changes in the performance ofexports influences total demand for farm output and hence overall agricultural output.

There are, thus, potentially individual or direct effects of changes in macroeconomicconditions on agriculture and these effects are further influenced by inter-relationshipswithin agriculture. However, we should also point out that the macroeconomic variablesthemselves are inter-related and the impact of one change influences the others and the

impact on agriculture is not limited to just the change in one factor. The Vector AutoRegressive method is a suitable econometric tool to analyse such inter-relationships.

7


8/27

Methodology

The vector auto regression (VAR) representation ( or a Vector Error Correction (VEC)Model , if the variables are cointegrated) of variables allows an assessment of the inter-

relationship in a dynamic framework. Although this representation has often been termeda-theoretic, choice of the variables in the VAR can be guided by theory.

The VAR (or, VEC) framework permits us to examine the impulse response functionsof one of the variables in the VAR to shocks in the other variables. In other words, we areable to assess the response of say agricultural investment to macroeconomic shocks in aVAR framework. However, in one of the approaches, the results of VAR analysis would

be sensitive to the ordering of variables in the VAR. To overcome the ambiguity, thegeneralised impulse response functions have been developed (Pesaran and Shin, 1998).

The variance decomposition of the VAR allow us to quantify the contribution ofdifferent variables in a VAR to the variability of a selected variable. For example, thecontribution of macroeconomic variables to the variance of agricultural investment

provides an assessment of the impact of the macroeconomic factors on agriculture.

Though the impulse responses measured through generalized impulse response approachare independent of the ordering of the variables, to arrive at variance decomposition oneneeds to specify an ordering. Thus it is not possible to completely get rid of the orderingissue. Further an ordering based on theoretical justification is expected to provide betterindication of the effect of different factors. Keeping this in mind we have considered twoalternatives of sequencing of the variables2. The macroeconomic variables considered in

the present analysis are:

i. (Gross fiscal deficit/ GDP at market prices)ii. Terms of trade: (WPI agriculture/WPI manufacturing)*100iii. Interest rate (PLRX: Minimum lending rate of IDBI and PLR of commercial banks)iv. Real exchange rate (REER based on 36 country trade weighted bilateral rates)

The agricultural variables are:i. Agricultural investment (Gross fixed capital formation in constant prices)ii. Agricultural exports (in constant prices)iii. Agricultural GDP (in constant prices)

The VAR model under consideration is

VAR based on (GFDC, REER, PLRX, GDP_AG, EXP_AG, TT, GFCF_AG)

Proceeding with the analysis,. first for the data under consideration unit root test is carriedout to identity the order of integration using the following test. Three different tests areapplied to attain robustness, viz.,

Augmented Dickay- fuller test

2 One in the main text and other in Appendix.

8


9/27

Schmidt Phillips test: two alternatives i.e Z(Rho) and Z ( Tau)

KPSS test ( kwiatkowski, Phillips, Schmidt & shin(1992) which tests the nullhypothesis of stationarity, level stationarity and trend stationarity.

Results are presented in table1 and they indicate that most of the variables are I(1) series3.

Table 1 Unit Root test result

Schmidt Philips test KPSS

ADF Rho test Tau test Levelstationarity

TrendStationarity

AG_GDP I(1) I(1) I(0) I(1) I(1)

AG_EXP I(1) I(1) I(1) I(1) I(1)

AG_INV I(1) I(1) I(1) I(1) I(1)

REER I(1) I(1) I(1) I(1) I(1)PLR I(1) I(0) I(1) I(1) I(1)

TT I(1) I(1) I(1) I(1) I(1)

The next step is to check whether or not the variables are cointegrated. If the they are, thenformulating a pure VAR model will be inappropriate. Johansen trace and Saikkonen andlitkepohl tests are used to examine this. Here it should be noted that the number ofcointegration structure, depends on the number of lags in the model. To identify the lagstructure AIC, HQ and Schwatz criteria were used. These tests indicate 3 as the optimallag length implying VEC model will have 2 lag structure (results in table 2).

Table 2: Cointegration test result

Johansen trace test Saikkonen and Lutkepohl test

R0 LR p-value R0 LR p-value

0 202.14 0.0000 0 152.52 0.0001

1 109.90 0.1396 1 81.07 0.2137

2 80.53 0.1698 2 53.97 0.3490

3 55.72 0.2009 3 30.94 0.5925

4 34.70 0.2645 4 13.63 0.8507

5 18.72 0.3040 5 10.67 0.2805

6 5.39 0.5506 6 1.30 0.7124

Both the tests suggest the presence of one cointegration vector. Thus VEC model wasconstructed with one cointegrating vector and with 2 lags.

The general form of the VEC model under consideration is as follows.

3 Except for a few variables like GDP_AG and PLR follows I(0) under Schmidt Philips test.

9


10/27

0 Yt = [ n ]

1

1

....

t

co

t

D

Y

+ 1 Yt-1 + ..+ p Yt-p +c Dt + Ut

where, Yt = (Y1t..Ykt) is a vector of K endogenous variables; cotD contains alldeterministic terms ( like Constant, trend etc) included in the co integration relations andDt correspondsto the deterministic terms of the VAR component. The parameter and

have dimensions (k r), where r is the cointegration rank ; represents the cointegration

parameter and the loading co-efficients; s are the co efficient of the VAR part.

With this specification a VEC model was run and the estimated cointegration relationturned out follows.

10


11/27

1.0 GFDCt-10.146 REERt-1 0.022 PLR 0.690 TT + 0.0001 GDP_AG + 0.002EXP_AG + 0.004 GFCF_AG + 120.30 = 0

Thus we observe that GDP_AG, EXP_AG and GFCF_AG show positive responses tochange in REER, PLR and TT and negative response to GFDC in the long run. Theagriculture related variables are also positively associated with each other. Thus increasein agriculture investment will have positive impact on agriculture output. Similarly,increase in agriculture export will also have positive impact on agriculture output. Furtherimprovement in terms of trade for agriculture impacts all the agriculture related variables

positively.

The co integrated vector was further checked for stationarity. All the three tests indicatedthat co-integrated vector is I(0) series. The next step is to make sure that the errors are freefrom auto correlation and the model is stable. The LM test indicated that the residual are

free from auto-correlation. The roots of the reverse characteristic polynomial was found tobe

(1.1242 1.1883 1.1883 1.1693 1.3100 1.9672 2.4676 1.4327 1.4327 1.3140 2.7640

2.7640 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 )

For the condition of the stable VEC model with endogenous variables and 1 co

integrating equation the companion matrix should have exactly ( -1) i.e., in this case weshould have 6 unit roots and the remaining ones should be strictly greater than one. Fromthe above numbers we can see that we have exactly 6 ones the rests are all greater than one

indicating that the model being estimated is a stable one. Policy responses furthergenerated from the model are as follows.

Impulse Response Functions:

VAR methodology provides an estimate of the impact of a shock in terms of change in onecomponent of VAR on all the components over time. The impact is measured by theimpulse response coefficients (Greene, 1997). We discuss below the results of the impulseresponse analysis4. In each of the cases described below, the impact on agriculture relatedvariables is a result of a one-standard error increase in a selected macroeconomic variable.

Impulse response analysis provides an estimate of the impact of a shock over timebeginning with the year (period) in which the shock is administered. For the purposes ofthe present analysis, we have presented the results over a 25-year period from the year ofthe shock, however, one can see that substantial impact is only for five years or so. Wenote that although the initial shock to the system is in terms of one macroeconomicvariable, in the years following the initial shock, the other macroeconomic variablesrespond to the initial shock and in turn influence the agricultural variable. Thus, thesubsequent impact on an agricultural variable, after the initial year, comprises the impactof the initial shock, secondary impacts from the other macroeconomic variables and thedynamics of the agricultural variable itself. In the discussion, we attribute the impact to the

4 The estimated VAR equations are not reported in the paper to conserve space. They are available with theauthors.

11


12/27

initial shock but it should be borne in mind that secondary influences are at work, besidesthe primary or initial shock. Given this analytical background we begin with the firstmodel specified above:

Agricultural Investment and the Macro-economic Factors: Figure A-1 in Appendixpresents the estimated impact of the one-standard error shock of each macroeconomicfactor (one at a time) on agricultural investment. Some general observations can be madefrom the impulse response patterns. First, The impact of shocks given to most of thevariables is maximum during the first 5 years but some impacts can be seen in the lateryears as well. However, after 10 years or so the effects become minimal and the responsesstabilizes. Second, fiscal deficit appear to have negative impact on agricultural investmentin the long run even though some of these deficits may have been incurred in order tosubsidise agriculture. As expected an increase in interest rate measured through PLRX hasnegative impact ; while increase in terms of trade has a positive cumulative impact. Theimpulse responses are also summarised in Table 2 below for the impact in the year 1 of the

shock (Year 1 in the table), in two subsequent years and cumulative impact for 5, 10 and25 years.

Table 2. Results from Impulse Response Analysis: Impact of Macroeconomic Shocks

on Agricultural Investment*

Cumulative impact

1st year 2nd year 3rd year 5 yrs 10 yrs 25 yrs

GFDC -34.90 16.42 -124.14 -80.67 -87.43 -83.50

REER -61.01 -114.42 -285.93 -280.24 -198.44 -231.58

PLR -0.29 -139.35 46.08 -150.85 -160.14 -142.46TT 251.35 -80.90 122.11 170.09 147.00 100.71

*measured in crores of rupees (10 million rupees) at constant prices.A rise in the real value of the rupee (rise in REER) leads to a decline in agricultural

investment in terms of machineries.

Agricultural Exports and Macro-economic Factors: The impulse responses ofagricultural exports are presented in Figure A.1-2 in the Appendix. Although fiscal deficithas fluctuating impact in the initial years (see the first three years in Fig. A.1-2) the overall

long run impact is negative. The initial (first period) impact on agricultural exports ispositive for an increase in the interest rate (PLRX). In the later period, there is a largernegative impact, although the subsequent impacts are not uniform over time but much lessfluctuating. The impulse response is summarised in Table 3 below for agricultural exportsand also in Figure A.1-2 in Appendix. An increase in agricultural price (TT) has positiveimpact on agricultural export. We have seen earlier that TT also has positive impact onagricultural investment. Thus favourable terms of trade are having positive impacts on thesector. Given the impact of price increase on agricultural export, an increase in REER maycapture an appreciation of the value of rupee. Usually this should result in negative impacton export. However, an appreciating rupee may in turn imply an overall goodmacroeconomic condition and hence a positive impact on export, which we observe in this

situation.

12


13/27

Table 3. Results from Impulse Response Analysis: Impact of Macroeconomic Shocks

on Agricultural Exports*

Cumulative Impact


GFDC -111.51 -44.28 2.66 -135.59 -127.75 -133.48REER 51.38 42.08 14.52 124.96 121.15 119.75

PLR 5.58 52.96 -118.88 -111.95 -89.64 -80.60

TT 102.64 8.18 30.05 102.95 89.57 100.67

* Measured in U S million dollars

Agricultural GDP and Macroeconomic Factors: The overall performance measure foragriculture selected in this analysis is agricultural GDP. Impact of macro-economicshocks on agricultural GDP captured in the impulse responses is spread over relativelyshort periods of time (Figure A.1-3 in the Appendix). We observe that an increase infiscal deficit has a negative impact on agricultural GDP. Increase in fiscal deficit is earlierseen to effect both investment and export also adversely. This may be due to adverse termsof trade resulting from inflationary pressure created by fiscal deficit. The impacts ofREER and PLRX also have expected sign. The results are summarized in Table 4.

Table 4. Results from Impulse Response Analysis: Impact of Macroeconomic

Shocks on Agricultural GDP*

IRF: Agricultural GDP

Cumulative Impact


Fis_def -2453.28 -1046.98 2407.88 -2487.44 -2450.19 -2140.02

REER -2400.45 3430.56 -2337.67 -1581.32 -1077.04 -1219.86

PLR -1403.88 566.44 -443.39 -1090.63 -1476.31 -1346.92

TT 1101.25 1016.06 .2239.24 2457.82 1488.18 1112.97

* Measured in crores (10 million) of rupees at constant prices

Integrated View of the Results: Results from three different specifications of the VARs

show that the impact of different macroeconomic factors on the three selected agriculturalperformance variables are quite similar only except in the case of effect of REER onagricultural export. For example, an appreciation of REER is seen to affect agriculturalexports and agricultural investment differently. First is to note than exports constitute onlya small part of agricultural output. Secondly export decisions are based on entirelyseparate set of considerations. From variance decomposition also it can be seen that REERexplains only 2.8% of variability of agricultural export (see Table 6). To provide anintegrated view of the results, we summarise in Table 5 the estimated impact from theVAR by aggregating impulse responses over 5 year period. It should be noted here thatqualitatively the integrated picture does not change even when we consider impulseresponses over 25 or 10 year period.

13


14/27

Table 5. Impact of the Macroeconomic Factors on Agriculture: An Integrated

picture (Cumulative Impulse Response over 5 years)

Macroeconomicvariable

Agricultural GDP Agriculturalexports

Agriculturalinvestment

GFDC Negative Negative Negative

PLRX Negative Negative NegativeREER Negative Positive Negative

TT Positive Positive Positive

Note: The results that are different from the remaining two cells across the row are in boldfont.

Table 5 shows that in almost all cases we observe expected linkages, except in the case ofREER, which shows a positive impact on export.

Clearly, it is difficult to identify structural relationships through VARs. But the resultspresented above show that the macroeconomic factors do influence agriculture through anumber of interfaces. Thus, the agricultural sector is not immune to macroeconomic policychanges as generally perceived.(c) The Variance Decomposition:

VAR methodology provides a means of assessing the contribution of different variables ina system to changes in any given variable within the VAR. The variance decompositioncomponent provides an estimate of the contribution of a variable to the variability ofanother variable in the VAR. We apply variance decomposition technique to assess thecontribution of the macroeconomic variables to variability in the selected agricultural

variables. Table 6 summarises the results of the variance decomposition analysis.

The results indicate that many of the macroeconomic factors are important in explainingthe variability in agricultural performance indicators. In fact The VAR results show that40--50% of variability in agricultural performance indicators is captured by the macro-level factors. This is a strikingly significant impact and indicates that agricultural sector isfar from being insulated from the overall macroeconomic conditions of the economy. In

particular, the macroeconomic factors in different VARs also point to the fact that thefiscal deficit has significant contribution in explaining the variability in the case of allthree agricultural variables. Amongst the price related variables impact of interest rateappears to be minimal. As expected agricultural GDP has significant impact on

agricultural investment and export. Further past values of the variables explains about50% (or more) of the variability in all the cases.

Table 6. The Impact of Macroeconomic Factors on Agriculture: Percentage of

Variance in Agricultural Performance Variables Explained by Macro-economic

Factors

GFDC REER PLAR TT -80.90 AG_EXP AG_INV

AG_GDP 4.6 8.1 1.1 5.4 66.9 3.3 10.0

AG_EXP 7.6 2.8 9.0 7.2 11.0 51.3 11.0

AG_INV 1.0 6.2 1.8 4.9 8.8 7.5 69.7

14


15/27

5. Policy Implications and Concluding Remarks

In this paper, we set out to assess the impact of macroeconomic factors on agricultureusing the time-series approach. There are clearly several mechanisms by which the shocksto macroeconomic variables would be transmitted to decision variables in agriculture and

finally affect the performance of the sector. For the purposes of the present analysis, pricevariables, viz., exchange rate, terms of trade and interest rate and the macro imbalances,captured through gross fiscal deficit of the centre to GDP are considered. We also chosethree agriculture-related variables: agricultural investment (real), agricultural exports (real)and real agricultural GDP for measuring the sensitivity of agricultural sector to themacroeconomic conditions.

Results of impulse response analysis capture the direction of the impact of macroeconomicshocks to agriculture. The direction of the impact is generally along the expected linesindicated by structural relationships, suggesting significant inter-relationships of variables.First it indicates that fiscal deficit incurred by providing subsidies to agricultural sector inthe long run does not generate either higher production or export of the sector. This showsthat incurring fiscal deficit to provide subsidy or other stimuli to the sector may not bevery effective in the long run. On the other hand relatively higher price for agriculturaloutput may induce the farmers to invest, export and produce more. Thus, providingnecessary security to the farmers through remunerative procurement prices and ensuring

better access to the market (without incurring fiscal deficit) are essential for the sector.

Interestingly, improvement in the value of rupee does not have an adverse effect onexport. Improvement in the value of rupee witnessed in recent years is due to the general

opening up of Indias markets resulting in greater inflow of foreign exchange. Liberalizedmarkets have also helped improve export potential of all sectors including agriculture.Therefore, even with an appreciating REER one can observe improvement in exports.However, this increase in exports is not accompanied by an increase in agricultural outputor investment when the source of export is appreciating rupee. This could be due to thefact that appreciating rupee also induces imports, which may meet some of the domesticconsumption needs. Furthermore, there may also be other inter-relationships that are notconsidered here explicitly.

Another interesting finding is that interest rate has comparatively much less impact on theagricultural variables especially investment and output. The presence of excess demand

for formal credit and the demand- rationed nature of the credit market are possibleexplanation for this outcome. This is clear from the wide spread existence of informalsector lenders. If a change in the formal interest rate is accompanied by a correspondingchange in the supply of credit then one may expect to see a significant impact. However,given the rigidity of supply of credit from the formal credit supply channels, the impact ofa change in the interest rate on the use of credit is also muted. Thus, it is the availabilityand delivery mechanism of credit that plays a more important role the than interest rate.

The present analysis has pointed to the substantial impact of macroeconomic factors onagriculture. The analysis does not track all the transmission mechanisms but points to thelikely links. While the results need to be qualified by the underlying assumptions, it is

difficult to ignore the need to keep in view the macroeconomic factors in understandingthe changes in the agricultural sector.

15


16/27

References

Bhide, S, M Rajeev and B P Vani, 2005, Do Marcoeconomic Conditions Matter for

Agriculture? The Indian Experience, Working Paper Number 162, Institute for Social and

Economic Change, Bangalore, India

Economic Research Foundation, 2002,National Accounts Statistics, CD, Mumbai.

Enders, W. 1995,Applied Econometric Time Series, John Wiley and Sons, Inc., NY.

Greene, W.H. 1997,Econometric Analysis, Third Edition, Prentice Hall, Englewood Cliffs

NJ.

Kalirajan, K.P. and S. Bhide, 2003, A Disequilibrium Macroeconometric Model for the

Indian Economy, Ahgate, London.

Narayana, N. S. S., Parikh K. S. and T. N. Srinivasan, 1991, Agricultural Growth and

Redistribution in Income: Policy Analysis with a General Equilibrium Model of India,

Elsevier Science Publishers, Amsterdam.

Pesaran, H.H. and Y. Shin, 1998, Generalized Impulse Response Analysis in Linear

Multivariate Models,Economics Letters, 58, pp. 17-29.

Rangarajan, C, 1982, Agricultural Growth and Industrial Performance in India,

International Food Policy Research Institute, Washington, D.C.

Reserve Bank of India, 2002,Report on Currency and Finance, 2001-02, Mumbai.

Reserve Bank of India, 2003, Handbook of Statistics on Indian Economy, 2002-03,

Mumbai.

Shand, R. and K.P. Kalirajan, 1999, The Agriculture-Manufacturing Nexus and

Sequencing in Indias Reform Process, in editor,Economic Liberalisation in South Asia,

16


17/27

R. Shand (ed.) Macmillan India Limited, Delhi.

Sims, Christopher, 1980, Macroeconomics and Reality,Econometrica, 48, pp.1-49.

Strom, S., 1993. Macroeconomic Considerations in the Choice of an Agricultural policy:

A Study into Sectoral Interdependence with Reference to India, Ashgate Publishing group,

London.

17


18/27

Appendices

Appendix A.1: Details of VAR analysis

Table A.1.1 Unit Root Tests of Selected VariablesVariable ADF Statistic for Variable in Order of

IntegrationLevel form

lag1

Level form

lag2

First

Difference

form lag1

First

Difference

form lag2

Second

Differenc

e form

lag1

Second

Difference

form lag2

REER -1.510 -1.254 -3.886** -3.037 I(1)

(DW 1.936) (DW 1.818) (DW

1.836)

(DW

1.451)

PLR -1.112 -0.687 -4.916*** -4.431*** I(1)

(DW 2.008) (DW 2.097) (DW2.143)

(DW1.938)

GDP_AG -2.618 -1.947 -5.816*** -3.799** I(1)

(DW 2.086) (DW 1.941) (DW

1.980)

(DW

1.957)

EXP_AG -1.623 -1.463 -2.1792 -1.508 -4.303 ** -4.026** I(2)

(DW 1.552) (DW 1.546) (DW

1.563)

(DW

1.635)

(DW

1.871)

(DW 1.908)

GFCF_AG -1.797 -2.055 -3.748** -3.061 I(1)

(DW 1.852) (DW 2.030) (DW

1.970)

(DW

1.984)

GFDC/GDP -1.372 -0.814 -5.298*** -4.308*** I(1)(DW 2.028) (DW 1.882) (DW

1.925)

(DW

2.018)

TT -3.161 -1.246 -7.567*** -4.466*** I(1)

(DW 1.644) (DW 2.135) (DW

2.108)

(DW

1.895)

Note: See text for details of variables

Figure A.1.1 Impulse Response Analysis for Agricultural Investment

Response of agriculture investment to one s.d. real effective exchange rate innovation

18


19/27

Fis_def toAG_INV

-150

-100

-50

0

50

100

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Response of agriculture investment to one s.d. real effective exchange rate innovation

REERtoAG_INV

-400

-300

-200

-100

0

100

200

300

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

_

Figure A.1.1 Impulse Response Analysis for Agricultural Investment (Continued)

Response of agriculture investment to one s.d. interest rate innovation

19


20/27

PLRtoAG_INV

-200

-150

-100

-50

0

50

100

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Response of agriculture investment to one s.d. terms of trade innovation

TTtoAG_INV

-150

-100

-50

0

50

100

150

200

250

300

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Figure A.1.2 Impulse Response Analysis for Agricultural Exports

Response of agriculture exports to one s.d. real effective exchange rate innovation

20


21/27

REERto AG_EXP

-30

-20

-10

0

10

20

30

40

50

60

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Response of agriculture exports to one s.d. fiscal deficit innovation

Figure A.1.2 Impulse Response Analysis for Agricultural Exports (Continued)

21

FIS_DEF to AG_EXP

-120

-100

-80

-60

-40

-20

0

20

40

1 2 3 4 5 6 7 8 9 10 1 1 1 2 1 3 14 1 5 1 6 17 1 8 19 2 0 2 1 22 2 3 24 2 5 2 6


22/27

Response of agriculture exports to one s.d. interest rate innovation

FIGURE A-2 (continued)

Response of agriculture exports to one s.d. terms of trade innovation

22

PLRtoAG_EXP

-140

-120

-100

-80

-60

-40

-20

0

20

40

60

80

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

TT to AG_EXP

-60

-40

-20

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26


23/27

Figure A.1.3 Impulse Response Analysis for Agricultural GDP

Response of agriculture GDP to one s.d. real effective exchange rate innovation

Response of agriculture GDP to one s.d. PLR innovation

FIS_DEF to AG_GDP

-3000

-2000

-1000

0

1000

2000

3000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

23

REER to AG_GDP

-3000

-2000

-1000

0

1000

2000

3000

4000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26


24/27

Figure A.1.3 Impulse Response Analysis for Agricultural GDP(Continued)

Response of agriculture GDP to one s.d. PLR innovation

Response of agriculture GDP to one s.d. terms of trade innovation

TTtoAG_GDP

-3000

-2000

-1000

0

1000

2000

3000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

24

PLRto AG_GDP

-2000

-1500

-1000

-500

0

500

1000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26


25/27

Appendix A.2

Table A .2 Basic Data Used in the study

Year

AgriculturalExports (RsCrores in1993-94

prices)

AgriculturalGDP (Rscrores in1993-94prices)

Investmentin

Agriculture(Rs Croresin 1993-94

prices)

CurrentAccountBalance(% ofGDP)

Fiscaldeficit of

the centralGovernment(% of GDP)

PrimeLendingRate (%)

RealEffectiveExchange

Rate(Index)

Terms oftrade

1970-71 1005.3 7902.0 137320.0 -0.97 3.08 8.5 125.36 97.55

1971-72 1062.9 8349.0 134742.0 -1.02 3.53 8.5 121.78 106.43

1972-73 1212.2 8831.0 127980.0 -0.58 4.04 8.5 119.56 107.791973-74 1257.1 8760.0 137197.0 1.73 2.64 9.0 127.50 97.74

1974-75 1456.3 8212.0 135107.0 -1.23 2.97 10.3 114.14 96.96

1975-76 1654.7 8924.0 152522.0 -0.21 3.64 11.0 106.27 106.13

1976-77 1628.7 11066.0 143709.0 1.00 4.24 11.0 101.34 107.79

1977-78 1692.4 11347.0 158132.0 1.11 3.62 11.0 100.12 99.99

1978-79 1582.5 12780.0 161773.0 -0.22 5.18 11.0 91.98 101.84

1979-80 1775.5 13344.0 141107.0 -0.46 5.29 11.0 97.08 111.60

1980-81 1719.6 13721.0 159293.0 -1.54 5.77 14.0 104.48 119.20

1981-82 1753.1 13407.0 167723.0 -1.68 5.14 14.0 104.48 111.60

1982-83 1645.4 13766.0 166577.0 -1.74 5.64 14.0 101.17 107.64

1983-84 1520.5 13926.0 182498.0 -1.51 5.94 14.0 104.24 100.93

1984-85 1578.8 13846.0 185186.0 -1.17 7.09 14.0 100.86 101.49

1985-86 1597.5 13061.0 186570.0 -2.14 7.86 14.0 98.27 107.53

1986-87 1672.6 12789.0 185363.0 -1.87 8.47 14.0 90.24 100.97

1987-88 1643.1 13375.0 182899.0 -1.78 7.63 14.0 85.36 95.53

1988-89 1498.4 14335.0 211184.0 -2.75 7.34 14.0 80.41 98.99

1989-90 1670.7 12728.0 214315.0 -2.34 7.33 14.0 78.44 107.88

1990-91 1768.5 15805.0 223114.0 -3.05 7.85 14.5 75.58 102.87

1991-92 2054.1 14546.0 219660.0 -0.34 5.56 19.0 64.20 95.86

1992-93 2111.5 15610.0 232386.0 -1.71 5.37 18.0 57.08 98.54

1993-94 2526.7 14749.0 241967.0 -0.42 7.01 16.0 61.59 100.00

1994-95 2654.6 15978.0 254090.0 -1.05 5.70 15.0 66.04 96.70

1995-96 4159.5 16824.0 251892.0 -1.65 5.07 16.5 63.62 96.70

1996-97 4515.2 17009.0 276091.0 -1.19 4.88 14.8 63.81 91.20

1997-98 4038.7 17035.0 269383.0 -1.37 5.84 14.0 67.02 91.20

1998-99 4114.4 16516.0 286094.0 -0.95 6.45 12.5 63.44 85.00

1999-00 4010.6 18082.0 286983.0 -1.04 5.35 12.3 63.29 86.20

2000-01 4337.7 18602.0 286666.0 -0.54 5.32 11.5 66.53 86.60

2001-02 4547.1 19402.0 304666.0 0.3 5.07 11.5 68.43 85.10

2002-03 5232.9 19280.0 283393.0 1.24 5.89 10.2 72.76 84.5

2003-04 5010.9 21628.0 310611.0 2.32 4.47 8.9 74.14 85.6

2004-05 5144.2 19428.0 314180.0 -0.8 4.03 8.5 74.53 88.5

25


26/27

26


27/27

i

ii

iii

do macro economic conditions matter for agriculture

Documents