solomon asfaw et al. fao of the united nations, rome, italy epic program diversification, climate...
Post on 19-Jan-2016
221 Views
Preview:
TRANSCRIPT
Solomon Asfaw et al.
FAO of the United Nations, Rome, ItalyEPIC Program
Diversification, Climate Risk and Vulnerability to Poverty: Evidence
from Rural Malawi
Dakar, Senegal 26 November 2015
Motivation• Climate variability (CV) will reduce agricultural productivity
and increase uncertainty
• The vast majority of the poor is concentrated in developing countries and live from agriculture
• Adaptation of the agricultural sector to CV is a priority to protect the poor and ensure food security
• At the household level, adaption can happen on-farm or off-farm livelihood diversification strategies are:- crucial to adapt or cope with climatic and other
risks/hazards;- important determinant of household welfare
Climate patterns in Malawi
Average rainfall 1983-2010 (mm/year)
Coefficient of variation of rainfall 1983-2010
Research questions1. Does climatic variability affects household welfare
measured by vulnerability to poverty (i.e., expected consumption and variance of consumption)?
2. Which policy mechanisms are effective in reducing the negative welfare effects of climate variability?
3. What are the drivers of labour and cropland diversification in rural Malawi?
4. What are the implications of diversification choices for household welfare?
Why households diversify livelihood?
Push factors
• Managing risk & income variability
• Adapting to changing weather condition
• Diversification provides safety-net
Pull factors
• Off-farm opportunities• Higher wage rates & higher
returns to entrepreneurial activities
• Economies of scope
Lead to lower, though more stable, welfare levels
Impact on welfare
Increase welfare, but not necessarily more stability
Data sources• World Bank Living Standard Measurement Survey (LSMS-IHS)
2010-11 – about 8,000 rural households
• Rainfall & temperature data linked to IHS at EA level:- Rainfall (1983-2010): Dekadal (10 days) rainfall data from Africa
Rainfall Climatology v2 (ARC2) of the National Oceanic and Atmospheric Administration’s Climate Prediction Center (NOAA-CPC)
- Temperature (1989-2010): Dekadal surface temperatures from the European Centre for Medium-Range Weather Forecasts (ECMWF)
• Institutional surveys at district level:- Credit, extension and other information sources, agricultural input
and output markets, public safety nets programs, donor/NGO programs & projects
Diversification indices: concepts• We use the Margalef, Berger-Parker and the Shannon-Weaver
indices to measure household-livelihood diversification
• On-farm diversification captured by:- Cropland diversification index: based on number of crop
types planted and the area allocated in the 2009-10 agricultural season
• Off-farm diversification captured by:- Labour diversification index: based on number of work
performed by household members and the time allocated [computed for males, females & total household members]
Diversification patterns in Malawi
Labour diversification Cropland diversification
Descriptive statistics: selected variablesMean Std. Dev. Min Max
Climate variablesCoV of Nov-May rainfall 1983-2010 0.211 0.035 0.123 0.288Average of Nov-May rainfall 1983-2010 (mm) 850 106.5 609.6 1,265.4Anomaly of Nov-May rainfall 2009-10 -0.086 0.092 -0.369 0.2Household wealthWealth index -0.502 1.37 -1.45 12.053Agricultural implements access index 0.374 1.378 -3.272 8.265GPS based land size (acre) 2.479 2.571 0 44.35InstitutionsNumber of agricultural extension and development officers in the district 9.546 3.9 0 22Number of microfinance institutions in the district 2.813 1.639 0 6Fertilizers distributed per household in the district (MT) 1.269 0.518 0.305 2.249ln(MASAF wages paid in 2008-09 season (million MKW/household)) 0.004 0.002 0.001 0.013Welfare indicators ln(total real consumption expenditure per household) 10.713 0.652 8.556 13.564Diversification indicesMargalef index of labour diversification, all adults 0.043 0.072 0 0.372Margalef index of labour diversification, male adults 0.039 0.071 0 0.721Margalef index of labour diversification, female adults 0.018 0.053 0 0.379Margalef index of cropland diversification 0.148 0.121 0 0.826
Effect of climate variability on welfare – components of vulnerability to poverty
• Asset-based-vulnerability estimation with cross-sectional data (Chaudhuri et al., 2002; Christiaensenand Subbarrao, 2005):
(1)
= log of per adult equivalent consumption, , , and = climatic, wealth, demographic and institutions = interaction between climatic and institutional variables
• Allowing for from (1) to be heteroskedastic, we use athree-step FGLS to recover unbiased estimates of:- = conditional expected consumption
Without interaction Interaction with policy
variables
Variance Expected
consumption Variance Expected consumption
Coefficient of variation of rainfall, 1983-2010-
0.349*** -0.984*** -0.622 -2.854***
Long term mean rainfall, 1983-2010 (mm) 0.001 0.050*** 0.005 0.040***
Rainfall anomaly, 2009-10 rainy season 0.238*** -0.608***
0.229*** -0.577***
Wealth index 0.003 0.207*** 0.003 0.207***
Agricultural implements access index-
0.015*** 0.047***
-0.016**
* 0.048***GPS based land size (acre) -0.000 0.012*** -0.000 0.012***
Agricultural extension and development officers in the district 0.003*** 0.014***
-0.014** 0.010
Number of microfinance institution 0.005 0.028*** -0.019 0.003
Fertilizers distributed per household (MT)-
0.025*** 0.074*** 0.105* -0.147*
ln(MASAF wages paid in 2008-09 season (million MKW/HH)) -0.880 4.297 -3.760 16.097
Extension service*CV rainfall 0.083*** 0.019
Safety-net*CV rainfall 11.891 -43.975Microfinance*CV rainfall 0.120 0.133
Fertilizer distributed*CV rainfall -0.617** 1.054**
Observations 8,009 8,009 8,009 8,009R-squared 0.019 0.451 0.021 0.452
Note: Robust standard errors in parentheses based on EA level clusters. *** p<0.01, ** p<0.05, * p<0.1.
Results (1)Effect of climate risk on vulnerability components: variance of consumption () and expected consumption per capita ()
What are determinants of diversification and the role of climate variability?
Diversification as a function of push and pull factors:
with
= Margalef index for j = labour and cropland = push & pull factors – climatic, demographic, wealth, institutional etc.
Seemingly-Unrelated Regression model (SUR)
Determinants of Diversification – Margalef Index
Results (2)
Margalef index
Crop diversification
Labor diversification
Coefficient of variation of rainfall, 1983-2010 3.962*** 1.715***Long term mean rainfall, 1983-2010 (mm) 0.005 0.006Rainfall anomaly, 2009-10 rainy season 0.349** -0.025Wealth index -0.048*** 0.118***Number of extension and development officers per household in the district 0.018*** 0.008***
Number of microfinance institutions -0.104*** 0.021***
ln(MASAF wages paid in 2008-09 season (million MKW/household)) -27.12*** -1.701
Agricultural implements access index 0.133***GPS based land size (acre) 0.189***Irrigation scheme in the community (1=yes) -0.137***
Fertilizers distributed per household ict (MT) 0.142***
Chi2 2541.18 795.11R-squared 0.259 0.098Observation 7242 7242Note:. *** p<0.01, ** p<0.05, * p<0.1.
Effect of diversification on household welfare
►Reverse causality between the welfare outcome variables and the diversification variable – potential endogeneity
►Instrumental Variable (IV) strategy, using variables on climate shocks (both drought and rainfall shocks) together with the variable indicating road density and distance to main facilities as our instruments
►Estimate OLS – no casual claim rather correlation.
Results (3)
Labor Cropland Both
Variance Expected consumption Variance Expected
consumption Variance Expected consumption
Margalef index of labor diversification, all adults 0.013 0.718*** 0.054 0.824***
(0.055) (0.079) (0.056) (0.081)Margalef index of cropland diversification -0.048 0.246*** -0.030 0.236***
(0.036) (0.053) (0.036) (0.053)Observations 7,862 7,862 7,255 7,255 7,023 7,023R-squared 0.019 0.455 0.018 0.434 0.019 0.443
Effect of diversification on household welfare
Results (3)
Note: Robust standard errors in parentheses based on EA level clusters. *** p<0.01, ** p<0.05, * p<0.1.
►Climate risk reduces consumption per capita and increases variances of consumption
►Labour and cropland diversification is higher in environments with greater climate variability [push factor]
►Wealthier households have also greater levels of diversification, supporting the pull factors hypothesis
►Diversification positively associated with welfare – pull factor outweighs push factor
►Access fertilizer subsidy seems to play prominent role in mitigating the negative effect of climate risk
Conclusion
Solomon Asfaw
THANK YOU!
Solomon Asfaw
Summary: determinants of diversification• Labour, cropland and income diversification is higher in environments with
greater climate variability [push factor]• Higher mean rainfall also is associated with greater diversification of income • A higher rainfall anomaly experienced in the last season reduces income
diversification income diversification not pursued to manage moderate shocks
• Wealthier households have greater levels of diversification, supporting the pull factors hypothesis
• Extension information leads to greater diversification acting as a pull factor by enabling farmers to take advantage of both on and off-farm opportunities
• Availability of fertilizer subsidies increases cropland and income diversification• Access to credit increases labour diversification, helping farmers to secure off-
farm income sources, but reduces cropland and income diversification• Female and male labour diversification perform similarly, though female
labour seems to be less responsive than male labour
top related