Introduction Universal fertilizer subsidies common in post- - - PDF document

4/11/13 ¡ 1 ¡

INDABA AGRICULTURAL POLICY RESEARCH INSTITUTE

Nicole M. Mason (MSU/IAPRI), T.S. Jayne (MSU), & Nicolas van de Walle (Cornell)

Presentation at the Department of Agricultural, Food, & Resource Economics Michigan State University

Fe Fert rtil iliz izer su r subsidie bsidies & v s & votin ing be behavio ior: r: Political economy dimensions of input subsidy programs

11 April 2013

Introduction

§ Universal fertilizer subsidies common in post-

independence SSA (dictatorships or one party rule)

§ Scaled back/eliminated in 1980s/1990s § Today: targeted subsidies (multi-party democracy)

§ 7 countries, US$2 billion in 2012 (Ricker-Gilbert et al., 2013)

§ Stated objectives:

§ Increase access to inputs, productivity, & production § Raise incomes, improve food security

§ Other objectives:

§ “Do something” for rural poor (Jayne et al., 2010) § (Re-) Election: garner and maintain rural votes

4/11/13 ¡ 2 ¡

“But there is no doubt that this Farmer Input Support Programme, which is supposed to be an economic activity, has sadly been abused or mismanaged by politicians and those seeking patronage and turned into a political tool for their election campaigns… And in this election year things will be worse – it will be nothing but a campaign tool; fertiliser bought with taxpayers’ money will be exchanged for votes.”

–Editorial, The Post, Zambia, March 13, 2011

Some evidence of input subsidy program – elections/voting links

• I. Past election outcomes  subsidized input targeting?

§ Tanzania: HHs w/ elected officials more likely to get input subsidy voucher (Pan & Christiaensen, 2012) § Ghana: fertilizer vouchers targeted to opposition strongholds

(Banful, 2011)

§ Malawi, Zambia: subsidized fertilizer targeted to supporters

(Mason & Ricker-Gilbert, 2013)

¨ Political economy not focus (election outcomes used as IVs)

• II. Targeted input subsidies  election outcomes?

§ Qualitative: input subsidies instrumental in Mutharika’s 2009 landslide victory in Malawi (Chinsinga, 2012; Mpesi & Muriaas, 2012) § Little (no?) quantitative empirical evidence to date: Do targeted input subsidies win votes ceteris paribus?

4/11/13 ¡ 3 ¡

Objectives

• I. Effects of election outcomes on HH-level

subsidized fertilizer targeting

• a. Swing voters, base, and/or opposition?
• b. Presidential vs. parliamentary election results?
• II. Effects of targeted fertilizer subsidies on

presidential election outcomes (share of votes won by incumbent)

• a. Do fertilizer subsidies win votes?
• b. If yes, to what extent? If not, what does?
• III. Link to poli. sci. debates, evidence
• IV. Policy implications

Subsidized fertilizer Election

• utcomes

Zambia

Contributions

• 1. Past election effects on subsidized fertilizer targeting

§ Panel data à control for unobserved heterogeneity (c.f. Banful, 2011) § More detailed examination – Zambia (c.f. Mason & Ricker- Gilbert, 2013)

• 2. Ceteris paribus effects of fertilizer subsidies on voting

patterns (incumbent’s vote share)

• 3. Fractional response w/ CRE & control function
• 4. Inform political science debates

§ Which voters do states target w/ clientelistic strategies? § Electoral effectiveness of targeted state expenditures?

4/11/13 ¡ 4 ¡

Outline

§

Zambia background

§

Fertilizer subsidies and elections in Zambia

§

Part I: Election outcomes è subsidized fertilizer?

§

Part II: Subsidized fertilizer è election outcomes?

§

Conclusions & policy implications

§

Research interests

Zambia

Indicator Year Size > Texas Population 14.2 mil 2013 % Rural 64% 2010

• Agric. % of labor force

85% 2004 GDP/capita (PPP) $1,700 2012 Poverty rate: 61% 2010 Rural 78% Urban 28%

Maize

• 84% of smallholders grow it; 60% of nat’l calorie consumption
• >85% of government ag spending = maize incentives

(fertilizer subsidy, Food Reserve Agency (FRA))

Sources: CSO (2011), CIA (2013)

4/11/13 ¡ 5 ¡

Elections & major political parties

Movement for Multi-Party Democracy (MMD)

Ruling party 1991-2011

§ First multi-party elections in 1991 § Presidential & parliamentary elections every 5 years

Patriotic Front (PF)

Defeated MMD in 2011

Fertilizer subsidies & timing of elections

20,000 40,000 60,000 80,000 100,000 120,000 140,000 160,000 180,000 200,000 MT of fertilizer

Source: MAL (2012)

MMD

• Nov. 1996:

Chiluba à FCP (200-800 kg)

• Dec. 2001 & Sep. 2006:

Mwanawasa à FSP (cash, 400 kg)

• Oct. 2008:

Banda à FISP 2009 (200 kg). Expands.

PF

Sep. 2011: Sata

4/11/13 ¡ 6 ¡

Part II: Effects of fertilizer subsidies

• n district-level share of votes won

by the incumbent

Part I: Effects of past election outcomes

• n HH-level

subsidized fertilizer targeting

Source: STR / Reuters

Part I: Effects of past election outcomes

• n HH-level subsidized fertilizer targeting

Conceptual framework § GRZ subsidy targeting criteria vague

§ Capacity to cultivate 1-5 ha of maize § Ability to pay farmer share of input costs (40-50%) § Cooperative membership

§ Theories of redistributive politics

§ “Core supporter”/turnout model (Cox & McCubbin, 1986) § “Swing voter” model (Lindbeck & Weibull, 1993; Dixit & Londregan, 1996, 1998)

§  Reduced form model of GRZ behavior

§ Kg of subsidized fertilizer allocated to HH function of farmer/ HH characteristics, election outcomes, other factors

4/11/13 ¡ 7 ¡

Part I: Effects of past election outcomes

• n HH-level subsidized fertilizer targeting

Empirical model: Unobserved effects Tobit § Corner solution (11% receive), qty differs, panel data

govtfertit = max(0,α + electktβ + zitδ +θwit +γ pit−1 + c1i + e1t + u1it) D(u1it | electkt,zit,wit, pit−1,c1i,e1t) = Normal(0,σ u1

2 )

§ govtfert: kg of subsidized fertilizer allocated to HH § i = 1,…,4285, t = 1999/2000, 2002/03, 2006/07 § elect: past election outcomes (Banful, 2011)

§ (a) =1 if ruling party (MMD) won constituency § (b) | Percentage point spread MMD – lead opposition | § (a) × (b)

Part I: Effects of past election outcomes

• n HH-level subsidized fertilizer targeting

Empirical model (cont’d)

govtfertit = max(0,α + electktβ + zitδ +θwit +γ pit−1 + c1i + e1t + u1it) D(u1it | electkt,zit,wit, pit−1,c1i,e1t) = Normal(0,σ u1

2 )

§ z: HH, community, region characteristics – targeting

§ Landholding, assets (farm equipment, livestock) § Age, education, sex of HH head; HH members by age group § Agro-ecological conditions; distances to nearest town, roads

§ w: market price of fertilizer pit-1 : maize price

§ c1i : unobserved heterogeneity e1t : year effects § u1it : idiosyncratic error

4/11/13 ¡ 8 ¡

Part I: Effects of past election outcomes

• n HH-level subsidized fertilizer targeting

Estimation strategy § Correlated random effects (CRE) Tobit § Not concerned about endogeneity of elect

§ Constituency level: outcome of 10,000s of votes (150 const.) § Part II: No feedback (govtfert à elect)

Data § 3-wave HH-level panel, nationally rep. (12,855 obs.) § Election results: Electoral Commission of Zambia

Part I: Results

• Dep. variable: kg of subsidized fertilizer to HH

Key explanatory variables APE Sig. p-value (a) MMD won (=1) 23.21 *** 0.00 (b) | PP spread MMD – lead opp. |

• 0.09

0.30 Interaction effect: (a) × (b) 0.54 *** 0.00

Note: *** p<0.01, **p<0.05, *p<0.10. APE = average partial effect.

MMD wins APE slightly smaller if use parliamentary (18.7 kg) Interaction effect example

• Compare predicted govtfert at 75th (59.1 pp) vs. 25th (18.9 pp)

percentile of MMD margin of victory in 2006

• 33.1 kg difference

4/11/13 ¡ 9 ¡

Part I: Effects of past election outcomes

• n HH-level subsidized fertilizer targeting

Discussion

§ Zambia: MMD rewards its base § Similar to Malawi findings (Mason & Ricker-Gilbert, 2013) § Consistent w/ “core supporter/turnout” model § Ghana: opposition strongholds targeted (Banful, 2011) § No evidence to date for “swing voter” model for

subsidized fertilizer

Part II: Effects of fertilizer subsidies

• n district-level share of votes won

by the incumbent

Do fertilizer subsides win votes in Zambia?

Source: STR / Reuters

No! But reducing poverty, inequality, & unemployment does.

4/11/13 ¡ 10 ¡

Part II: Effects of fertilizer subsidies on district- level share of votes won by the incumbent

Conceptual framework

§ Conventional wisdom in Zambia: fertilizer subsidies

(& FRA purchases) effective tools for gaining rural votes

§ Poli. sci. literature on voting in Zambia & SSA

(Bratton et al., 2011; Posner & Simon, 2002)

§ Ethnic voting § Economic voting (overall economy) § Own economic situation, private goods from government (e.g., subsidized fertilizer, high maize price from FRA) § Demographics

§  Reduced form model of voting behavior (Cerda & Vergara,

2008)

Part II: Effects of fertilizer subsidies on district- level share of votes won by the incumbent

E(sMMDdt | subfertdt−1,FRAdt−1,vdt,v pt,econpt,e2t,c2d ) = Φ(α + β1subfertdt−1 + β2FRAdt−1 + vdtδ1 + v ptδ 2 + econptγ + e2t + c2d )

Empirical model: Unobserved effects fractional response (district-level**) § Proportion dependent variable, panel data § sMMD: share of votes won by incumbent president § d (district)=1,…,72; p (province)=1,…,9; t = 2006, 2011 § subfert: scale/coverage of fertilizer subsidy program

§ % of smallholder HHs receiving § Mean kg/smallholder HH § Total district allocation (MT)

4/11/13 ¡ 11 ¡

Empirical model (cont’d)

Part II: Effects of fertilizer subsidies on district- level share of votes won by the incumbent

E(sMMDdt | subfertdt−1,FRAdt−1,vdt,v pt,econpt,e2t,c2d ) = Φ(α + β1subfertdt−1 + β2FRAdt−1 + vdtδ1 + v ptδ 2 + econptγ + e2t + c2d )

§ FRA: FRA maize purchases (MT) § vdt: total population, % female, % in various age groups,

registered voters, % female registered voters

§ vpt: % rural, province dummies (ethnicity), prov × year § econ: labor force, % unemployed, poverty rate, Gini

coefficient (income inequality)

§ e2t : year effects § c2d : unobserved heterogeneity

Estimation strategy § CRE pooled fractional probit (Papke & Wooldridge, 2008) § Concerned about endogeneity of subfert, FRA

§ Control function approach (Rivers & Vuong, 1988; Papke

& Wooldridge, 2008) § IVs: % of smallholder HHs cultivating 2+ ha (t-1 for subfert, t-2 for FRA)

¨ 2010/11: 27% of HHs, 55% of FISP fertilizer; 78% of FRA ¨ Strong IVs (p<0.03, positively correlated w/ subfert, FRA)

§ Fail to reject exogeneity (p>0.10)

Part II: Effects of fertilizer subsidies on district- level share of votes won by the incumbent

4/11/13 ¡ 12 ¡

Data

§ 2-wave district-level panel (144 obs.) § Election results: Electoral Commission of Zambia § Subsidized fertilizer: HH survey data, Min. of Ag. § Various secondary data sources (GRZ) Part II: Effects of fertilizer subsidies on district- level share of votes won by the incumbent

Part II: Results

• Dep. var.: share of district votes won by incumbent

Key explanatory variables APE Sig. p-value Subsidized fertilizer variables

• Not. sig.

p>0.50 FRA purchases (‘000 MT) 0.0025 * 0.06 Unemployment rate (%)

• 0.10 ***

0.00 Poverty rate (%)

• 0.03 ***

0.01 Gini coefficient (0-100 scale)

• 0.02 ***

0.00

Note: *** p<0.01, **p<0.05, *p<0.10. Bootstrapped standard errors (500 replications).

4/11/13 ¡ 13 ¡

Part II: Results (cont’d)

Magnitudes of FRA vs. economic conditions effects

Percentage point change in incumbent’s vote share given a: % Change from mean: 1% 50% é FRA purchases 0.02 1 ê Unemployment 0.5 33 ê Poverty 1.2 36 ê Gini coefficient 1.3 49

• Subsidized fertilizer: no stat. sig. effect
• FRA: stat. sig. effect but magnitude very small
• Economic variables: stat. sig. & large in magnitude
• Robustness check: no evidence of subsidized

fertilizer/FRA effects through economic variables

Why no subsidized fertilizer/FRA effects?

§ Relatively few farm HHs benefit

§ 2010/11: ~30% get FISP fertilizer and/or sell to FRA § 2006/07: ~10%

§ Benefits highly concentrated

§ 27% of HHs = 55% of FISP, 78% of FRA § 3% of HHs = 50% of FRA

4/11/13 ¡ 14 ¡

Conclusions

• 1. MMD used subsidized fertilizer to

reward loyalty

• 2. Fertilizer subsidies, FRA purchases had no

substantive effect on MMD’s share of votes in 2006 & 2011 elections

• 3.  poverty, inequal., & unemploy. wins votes

Policy implications

• 1. Is politically-motivated subsidy

allocation a problem? If so, how to  it? e.g., rules-based, transparent, & audited allocations

• 2. Politicization may be ê achievement of

stated objectives. Could depoliticizing  ‘more bang for the buck’ w.r.t. access to inputs, productivity, food security, incomes?

4/11/13 ¡ 15 ¡

Policy implications (cont’d)

• 3. ing effectiveness of subsidies as poverty- &

inequality-reduction, employment-creation tools = good politics! (e.g., target the poor, e-voucher to crowd-in private sector/create jobs)

• 4. >85% ag spending  fertilizer subsidies & FRA.

Shifting some funds to investments that  poverty, inequality, and/or unemployment = good politics! (e.g., roads, irrigation, electrification, ag R&D, improved extension, health, education, etc.)

Research interests

§ Broad themes:

• 1. Effects of gov’t policies/programs on S/H behavior/

well-being, market prices; political economy

• 2. Agriculture-health-nutrition linkages (HIV/AIDS)
• 3. Trends in & drivers of urban staple food consumption

patterns

§ Sustainable agricultural intensification; environment/

NRM; climate change resilience & adaptation

§ Asia-Latin America-Africa – lessons learned § Evidence-based agricultural policy processes § Applied econometrics; impact assessment

4/11/13 ¡ 16 ¡

Thank you for your attention! Questions?

Nicole M. Mason masonn@msu.edu IAPRI http://www.iapri.org.zm/ index.php Food Security Research Project http://fsg.afre.msu.edu/ zambia/index.htm

References

Banful, A. B., 2011. Old problems in the new solutions? Politically motivated allocation of program benefits and the “new” fertilizer subsidies. World Dev. 39, 1166-1176. Bratton, M., Bhavnani, Tse-Hsin, C., 2011. Voting intentions in Africa: Ethnic, economic, or partisan? Afrobarometer Working Paper No. 127. http://www.afrobarometer.org/component/content/article/68-english/working-papers/177-wp-127. Central Intelligence Agency (CIA), 2013. World Factbook – Zambia: https://www.cia.gov/library/publications/the-world-factbook/geos/za.html Central Statistical Office (CSO), 2011. Living Conditions Monitoring Survey Report: 2006 and 2010. CSO, Lusaka, Zambia. Cerda, R., Vergara, R., 2008. Government subsidies and presidential election outcomes: Evidence for a developing country. World Dev. 36, 2470-2488. Chinsinga, B., 2012. The political economy of agricultural policy processes in Malawi: A case study of the fertilizer subsidy programme. Working paper prepared for the Future Agricultures Consortium Political Economy of Agricultural Policy in Africa conference. http://www.future-agricultures.org/pp-conference-papers/doc_download/1645-the-political-economy-of-agricultural-policy-processes- in-malawi-a-case-study. Cox, G. W., McCubbins, M.D., 1986. Electoral politics as a redistributive game. Journal of Politics 48, 370-389. Dixit, A., Londregan, J., 1996. The determinants of success of special interests in redistributive politics. The Journal of Politics 58, 1132-1155. Dixit, A., Londregan, J., 1998. Ideology, tactics, and efficiency in redistributive politics. Quarterly Journal of Economics 113, 497–529. Jayne, T. S., Chapoto, A., Govereh, J., 2010. Grain marketing policy at the crossroads: Challenges for eastern and southern Africa, in A. Sarris and J. Morrison, eds., Food Security in Africa: Market and Trade Policy for Staple Foods in Eastern and Southern Africa. FAO, Rome. Lindbeck, A., Weibull, J., 1993. A model of political equilibrium in a representative democracy. Journal of Public Economics 51, 195–209. Mason, N. M., Ricker-Gilbert, J., 2013. Disrupting demand for commercial seed: Input subsidies in Malawi and Zambia. World Dev. 45, 75-91. Ministry of Agriculture and Livestock (MAL), 2012. Farmer Input Support Program implementation manual, 2012/13. Lusaka, Zambia, MAL. Mpesi, A. M., Muriaas, R. L., 2012. Food security as a political issue: the 2009 elections in Malawi. Journal of Contemporary African Studies 30, 377-393. Pan, L., Christiaensen, L., 2012. Who is vouching for the input voucher? Decentralized targeting and elite capture in Tanzania. World Dev. 40, 1619-1633. Papke, L. E., Wooldridge, J. M., 2008. Panel data methods for fractional response variables with an application to test pass rates. J. Econometrics 145, 121-133. Posner, D. N., Simon, D. J., 2002. Economic conditions and incumbent support in Africa’s new democracies: Evidence from Zambia. Comparative Political Studies 35, 313-336. Ricker-Gilbert, J., Jayne, T. S., G. Shively. 2013. Addressing the “wicked problem” of input subsidy programs in Africa. Applied Economic Perspectives and Policy, in press. Rivers, D., Vuong, Q. H., 1988. Limited information estimators and exogeneity tests for simultaneous probit models. J. Econometrics 39, 347–366.