Does the timing of Public Work payments matter?
- 1. ETHIOPIAN DEVELOPMENT RESEARCH INSTITUTE Does the timing of Public Work payments matter? Welfare effects of delays in the PSNP payments (2006 -2014) Berhane, G., Abay, M.H., Hoddinott, J., and Abay, K. A., Transformation and vulnerability in Ethiopia: New evidence to inform policy and investments GETFAM Hotel, Addis Ababa May 27, 2016 Addis Ababa 1
- 2. 2 Safety net-oriented Public Works (PW) programs - popular instruments of social protection schemes. Vast literature weighing successes of PW programs in view of design features and implementation capacities. targeting , size of transfer, delivery, sustainability Evidence on overall successes of PW programs is mixed. Often plagued by implementation problems related to the above design features. Little known about the optimal timing of payments in terms of maximizing welfare gains Motivation
- 3. 3 Key research questions? Study the welfare implications of delays in Public Works payments: whether and to what extent improving the timing of payments can improve household welfare. Insights to how PSNP payments can be best scheduled Should payments be made in small amounts (regularly)? … be delayed and made in lump-sums later in the hunger season? Do effects vary by welfare outcome, or objective?
- 4. 4 How would payment delays affect welfare? Effects of payment delays may depending on: Financial ‘flows’ and ‘needs’ of the receiver - managing basics vs. raising lump-sums; When (season!) and how (size!) the delayed payment will be paid Recent evidence suggesting the need for lump-sum transfers to bring about lasting changes in the lives of the poor. The poor are ‘saving constrained’ The poor ‘lack’ self-control & are ‘present-biased’
- 5. 5 Data and Method Five rounds of panel data covering 2006 – 2014, from four main regions that implemented the PSNP; Method: Pooled OLS and HH FE Exploiting variations in the timing of payments among participants; Control for drought shocks, payment levels, drought*delay Welfare outcomes considered: household consumption, income, assets, food gap, transfers; child nutrition & schooling
- 6. 6 Defining delay: timing of PSNP Public Works and Payments Public works, 2007 Public Works, 2008 Dec 2007 May 2008 Jan 2007 Jun 2007 Jan 2008 Jun 2008 Pre-survey year payment data Survey year payment data Public Works run in the first six months of the year (Jan –Jun). Payment schedule: end of public work month until mid next month, otherwise considered delayed!
- 8. 8 % reported … as primary source of food, by month and year 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% October November December January February March April May June July August September October November December January February March April May June July August September October November December January February March April May June July August September 2010 2012 2014 Own production Purchase Private transfers PSNP Others
- 9. 9 Average number of days worked, by month and round, 2007-2013 0 2 4 6 8 10 12 14 16 18 20 Average#daysworked 2007 2009 2011 2013 Public work months
- 10. 10 Average payments received for PW (in 2014 birr), by month and round, 2007-2013 0 50 100 150 200 250 300 350 400 450Averagepaymentsreceived 2007 2009 2011 2013 Public work months
- 11. 11 Timeliness of payment has improved over the years Lumpiness declines over time 1 = # months payment was received /# PW months done Lumpy payments in early years
- 12. 12 % that received PW payments during (Jan-Jul) and after (Aug-Dec) PW months Percent (%) year Jan-Jul Aug Sep Oct Nov Dec Total 2007 90.9 5.9 1.8 1.2 0.1 0.0 100 2009 92.0 3.4 2.9 1.2 0.3 0.2 100 2011 96.3 2.3 0.9 0.2 0.2 0.1 100 2013 95.3 2.2 0.8 0.6 0.6 0.6 100 Note: payments include both food and cash
- 14. 14 Using the monthly measure, we find that payment delay: decreases household consumption in March and April in the range of 6-8 percentage points. Effect driven by food rather than non-food. Delay has also negative effects on some child growth and schooling outcomes (using the aggregate measure): Increased wasting by 10 percentage points in March. Reduced the likelihood of grade achievement and increased the average age children started schooling. Effects more pronounced when delays are preceded by bad harvests. Key results – negative
- 15. 15 Key results – positive Using the aggregate payment delay measure, we find positive effects of delay on several outcomes; Increased livestock ownership (in TLU) and value of livestock, Increased value of production assets, & income from ‘crop and livestock’ sales, Increased private transfers, and borrowings, Decreased distress sales of livestock and production assets, Reduced food gap and the likelihood of facing acute food shortages during the hunger season.
- 16. 16 Key conclusion: delay is not necessarily bad thing 1) payment delays reduce outcomes that are needed to be bridged on daily basis (e.g., consumption). … more so towards the start of the lean season. 2) Payments delayed well into the lean season and made in lump-sums have strong positive effects … – ‘right timing’, or ‘investment effect of the lumpy transfer’. Consistent with recent evidence: “in addition to smaller periodical transfers, the poor need a one-off lump-sum transfer” Implication: back up small, frequent payments, with lump- sums. This supports the “livelihood transfers” in the livelihood component of PSNP 4.