aciar - ifpri - heterogeneity in male and female farmers’ preference for a climate-smart...

Heterogeneity in male and female farmers’ preference for a climate-smart technology: The case of direct seeded rice (DSR) in India

P. K Joshi, Md. Tajuddin Khan & Avinash Kishore

Objectives

1. To understand the preference heterogeneity between

men and women for a CSA technologies

2. To measure the willingness to pay for direct seeded rice

(Drum Seeder) for both men and women

3. To find out the factors that influence WTP for DSR

Direct Seeded Rice (Drum Seeder) Vs Transplanted Rice

How it works?• Pre‐germinated seeds are directly sown using a in a well puddled and levelled wet field.

• It’s a fibre material and easy to operate. It allows one person to sow one hectare in 5‐6 hours

compared to three to four days of transplanting by 20‐ 30 people in case of traditional

cultivation method.

Advantages

❖ Nursery preparation and tasks (pulling, transporting and transplanting seedlings) are avoided

❖ It requires less labor and water and emits less greenhouse gases into the environment than

transplanted rice

❖ It increases yield by significant amount(However it is not consistent)

Dis-advantage

❖ DSR has higher weed growth than the transplanted rice

❖ Fixed Investment to buy Machine

Sampling and Data: Stratified random sampling

• 2 districts (Thane and Palghar)

• 6 rice growing blocks from 2 district, four from Palghar (Palghar, Jawhar, Mokhada and Wada) and two from Thane (Shahapur and Morbad)

• Within each block, we randomly selected five villages (30 villages)

• Our sample consists of 666 respondents (329 men and 337 women) farmers from 400 households

1. 266 households (both men and women)

2. 134 households (63 male and 71 female separately)

Accounting for Women in Choice of Technologies

• Agricultural decisions are not undertaken by a unitary household

(Duflo and Udry, 2004)

• Women farmers may have different preferences from the men in

their families and adoption of a new technology may affect them

differently• true particularly for rice cultivation since women contribute 60-80% of labor in

rice in Asia

• Even greater share in transplanting labor

• Studies ignore women when sampling or compare FHHs &

MHHs• Leaves out a huge number (92%) of women who live in male-headed

households

• We sample one woman and one man from each HH in our sample• 266 pairs of respondents were surveyed independently

Women and technology adoption in agriculture

• Women have slower observed rates of adoption of a wide

range of technologies than men (Doss and Morris, 2002)

• and lower willingness to pay for new products like the

weather indexed insurance (Akter et al, 2016)

• Probably due to greater time and resource constraints

– lower human capital endowment (education and exposure to the

outer world) and

– poorer access to complementary inputs (Kamwamba-Mtethiwa,

et. al., 2012)

What we did

• We use choice experiment to estimate WTP for DSR Drum-seeder

Why Choice Experiment?

• Respondent choices are modelled using the random utility theory which assumes that the individual will choose the alternative that yields her the highest utility.– Random Parameter Logit Model (Mixlogit)

• More informative data, more variability, less collinearity among the variables, more degrees of freedom and more efficiency.

Selection of DSR-Drum Seeder Attributes and levels

• 41 FGD sessions were conducted in 6 villages.

• 30 male group and 11 female group FGDs (184 men and 68

women attended these FGDs) to get their independent inputs

on key attributes that may guide their choice of a drum-seeder

for DSR.

Attributes Definition Levels

Seed rate (Kgs) Seed (kg) is required for one acre of land 5, 10, 15, 20

Labour saved (mandays) Number of labour saving in one acre of land 8, 10, 12

Yield Increment (Kgs) Yield increment is considered in DSR 0, 50, 100, 150, 200

Weedicide cost (INR) Weedicide cost is must in DSR is considered 400, 600, 800, 1000

DSR Price (INR) Price of Direct seeded rice (drum seeder) 4000, 5000, 6000, 7000

Choice sets design

• We used the “dcreate” module in Stata to generate a D-optimal

design that takes into account all main effects

• We generated 36 unique choice sets which were randomly

allocated into four blocks of 9 choice sets (one respondent

faced only 9 question)

• Each choice set consisted of two alternatives and a status quo

option (see an example of a choice set in next page)

Example of choice set presented to survey respondents:

BLOCK 1: SET 3

Summary statistics from Women’s Empowerment in Agriculture Index (WEAI)

Men Women Diff. t-test

To what extent do you feel you can make decisions if you want(ed) to: (1-5)

Food crop farming 4.671 3.894 0.778 9.893***

Agricultural production 4.434 3.634 0.801 9.381***

Inputs to buy 4.521 3.691 0.829 10.224***

Crops to grow 4.407 3.732 0.675 8.108***

Crops to market 4.451 3.495 0.956 6.951***

Own wage salary 4.378 3.904 0.475 5.848***

Who, according to you, can decide whether to buy, sell, or rent/mortgage (self)?

Farm equipment (mechanized) 0.0703 0.002 0.067 4.713***

Leadership: Public speaking

Do you feel comfortable speaking up in

public? 0.406 0.274 0.132 3.545***

Time: Workload and leisure (1-10)

Did you work more than 10.5 hours in the

previous 24 hours? 0.539 0.762 -0.222 -6.142***

How would you rate your satisfaction with

your available time for leisure activities? 4.5 3.494 1.006 5.705***

Women have less say in agri. decision, negligible in mechanization and enjoy much less

leisure than their male counterparts.

Estimation of Random-parameter logit (RPLM) results

Note: Standard errors are in parenthesis and ***, ** and * denotes significance at 1, 5 and 10 percent level.

The top panel shows the posterior means values of marginal utility parameter and bottom panel shows the

heterogeneity in farmers’ preferences for each attribute.

Pooled Men Women

Random marginal utility parameters

Seed rate (kg) -0.01567** -0.01906** -0.01482

Labor saved (person-days) 0.20763*** 0.14591*** 0.2896***

Yield increment (quintals) 0.95118*** 1.30505*** 0.68066***

Weedicide cost (INR) -0.00004 -0.00015 -0.00007

Nonrandom marginal utility parameter

Price of DSR (INR) -0.00088*** -0.00093*** -0.00086***

Distribution parameters

Standard deviation (seed rate) 0.04658*** 0.027997* 0.07928***

Standard deviation (labor saved) 0.13533*** 0.07062** 0.13779***

Standard deviation (yield increment) 0.20774* -0.01035 0.54613***

Standard deviation (weedicide cost) 0.00018 0.00008 0.00037**

Log likelihood -4213.1788 -1788.1942 -2319.042

Estimated willingness to pay (WTP) for DSR-drum seeder

attributes

Note: Confidence interval derived using bootstrap procedure introduced by Krinsky and Robb (1986) based on 1000

random draws with mean and variance-covariance matrix of the estimated model parameter.

Mean Lower Upper

Pooled

Seed rate (kg) -17.814 -33.788 -2.920

Labor saved (person-days) 236.038 207.591 264.995

Yield increment (quintals) 1081.309 951.363 1202.904

Weedicide cost (INR) -0.042 -0.305 0.209

Men

Seed rate (kg) -20.561 -38.689 -2.928




Women

Seed rate (kg) -17.175 -44.165 7.086




T-test results of difference between Male and Female

MWTP for DSR-drum seeder attributes

Note: *** denotes significance at 1% level of significance,

Men Women Diff.

T-test of sig.

diff. in means

Seed rate (kg) -20.45 -16.48 -3.98 -1.27

Labor saved (person-days) 157.04 336.53 -179.49 26.39***

Yield increment (quintals) 1407.96 787.41 620.54 33.83***

Weedicide cost (INR) -0.16 -0.08 0.08 12.39***

More Money for Men, Less Drudgery for Women

Individual-level Marginal Willingness to pay (INR)

for DSR attributes

Note: Marginal willingness to pay (in Indian rupees) is shown on the horizontal axis.

Individual-level Total Willingness to pay (INR) for DSR

Note: Total willingness to pay (in Indian rupees) is shown on the horizontal axis

In our study, women were more interested in the new

technology: When shown the same choice-set, women

in the family are 20% less likely to choose the status-quo

option

Dependent variable: “Choose either of the two

DSR with drum seeder combinations” All households

Male -0.1950***

(0.0187)

Constant 0.9952***

(0.0931)

Card fixed effect Yes

Household fixed effect Yes

No. of observations 5867

R2 0.4025

Note: *** denotes significance at the 1 percent level. Figures in parentheses are

standard errors, which are clustered at the individual level.

What Determines the WTP behaviour for DSR-Drum seeder?

Note: Standard errors are in parenthesis and ***, ** and * denotes significance at 1, 5 and 10 percent level.

Dependent variable: mean WTP Full sample Households with 2 respondents

Land owned (hectares) 2.8128 2.5410

Average age of the respondent (years) 0.7084 1.0835

Gender (men = 1, else = 0) -93.1631*** -88.7659***

Scheduled tribe (yes = 1, else = 0) 11.5258 43.3328**

Primary education (%) 51.6877** 54.3138**

High school education (%) 40.5572* 52.1023*

Above high school education (%) 63.4655*** 73.9729***

Planning to use DSR drum seeder (%) 98.8973*** 111.9289***Migrate for off-farm employment opportunities?

(%) -18.0504 -2.5952

Member of self-help group? (%) 14.8968 23.6014

Like farming as a profession? (%) 30.7038** 39.1254**

Awareness of minimum support price? (%) 100.0034*** 101.6704***

Workload (work for more than 10.5 hours) 25.3472 14.3723

Working in MGNREGS 13.3335 8.9300

Access to input market (%) 9.8028 6.1369

Access to output market (%) 19.4466 34.4348

Access to credit facilities (%) 8.6710 12.7241

Access to weather advisory (%) 17.8597 15.5068

Constant 1095.2106*** 998.7731***

Number of observation 650 527

R-squared 0.2052 0.2120

Findings• Women have a higher WTP for DSR

• Age and landholding size, which represents do not affect adoption behaviour.

• Respondents who are planning to use DSR-drum seeder, aware about the MSP, like farming as a profession are willing to pay additional money for the adoption of technology.

• Access to credit, input as well as output market, weather advisory and who worked more than 10.5 hours in a day have a positive correlation for WTP for DSR drum-seeder

Conclusion• There are significant gender differences in marginal valuations

of different attributes of DSR, a climate smart technology.

• Men control cash. So, they have a higher WTP for attributes

that increase profits. Women have a higher WTP for labor

saving because they provide bulk of unpaid labor for rice

cultivation.

• Extension for promotion of DSR-drum seeder is likely to be

more successful if it also targets women farmers.

• Capital subsidy is needed to promote DSR-drum seeder

adoption.

Media coverage

Thanks !

aciar - ifpri - heterogeneity in male and female farmers’ preference for a climate-smart...

Education