Today I don't want to rehash the discussion of why do cash transfers (a recent discussion by some colleagues is here) or their myriad impacts (including among the legions, Berk's discussion here, David's here) but rather focus on one particular project. This project, the FAO-UNICEF from protection to production (PtoP) project is taking a number of cash transfer programs in Africa and evaluating their impact on productive activities such as agriculture and small businesses.
The emerging results are pushing me to the conclusion that in some cases cash is probably a good option for increasing households farm and business outcomes. But as this work is also showing (and as the literature shows elsewhere) it definitely isn't always a good option for these outcomes. And, of course, this raises the question: when is it? I am not going to answer this last question, but it is worth looking at the evidence coming from the PtoP project to see what's going on.
Let's start with Malawi. Here we have two papers which look at the impact of Malawi's Social Cash Transfer program. This is a relatively big unconditional transfer, clocking in around $14 a month which in these communities was enough to raise living standards from the bottom decile to above average. It's targeted at poor households, using a combination of the dependency ratio, having low food consumption and a lack of assets, combined with additional community targeting. Using a randomized phase in, two papers look at a range of productive outcomes (note: they have a small number of communities, so balance isn't great so these results are going to come from both the randomized diff-in-diff and matching estimates).
What do they find? Boone, et. al. (gated version here) show this transfer had significant and large increase in the ownership of farm tools (hoes, sickles, axes) and a fairly massive bump in the ownership of chickens (up by about 50 percentage points). Household adults shift to spending more time working on their own farm, with a concomitant drop in casual labor for others. In another paper, Covarrubias, et. al. show significant bumps in goat and cattle ownership. So for Malawi, we seem to be seeing a shift in working patterns and increases in productive assets.
Off to Zambia, where a paper by Seidenfeld and co. looks at the impact of the Child Grant Program. This is also an unconditional transfer, targeted to very poor areas and, within them, households with a child under age 5 (at the start of the program, they look for households with kids at 3 or below). Households get about $12 per month - a significant amount relative to household expenditure. Seidenfeld and co. use the randomized (community) assignment to the program to look at impacts.
And the impacts are fairly impressive. Food expenditure goes up and extreme poverty goes down by 5.4 percentage points. In terms of productive impacts, this program sweeps the boards in a rather strikingly consistent way. Operated land goes up by 18 percentage points (or 34%), with household spending 18 percentage points more on inputs and an increase in farm tool ownership. The value of harvest goes up by 50 percent relative to baseline (this is significant only at 10 percent). Households are more likely to be selling crops and they earn more from sales. Livestock ownership also goes up (particularly for chickens, goats and ducks). Off the farm, households are 17 percentage points more likely to have a non-farm enterprise, it's more likely to have operated for longer and they have way higher revenues and profits than the control group. The household labor story backs this up nicely with a shift from agricultural wage labor to family agriculture and non-farm work. With effects like this, one has to wonder about general equilibrium effects. Thome and co, as part of the Seidenfeld and co report, use a local economy general equilibrium model and seem to find positive spillovers.
The other two papers that have come out of this work so far show less striking and consistent effects. In an evaluation of the Kenya Cash Transfer Program for Orphans and Vulnerable Children, Asfaw et. al. find a bit on livestock ownership, but only when they push on heterogeneous impacts. And in an evaluation of the Livelihood Empowerment Against Poverty Program in Ghana, Handa, et. al. don't find much in the way of productive impacts (in this case, sporadic payments of the transfer may be part of the story).
So, clearly not a uniform win. But in some cases, like Zambia, we're seeing pretty large productive impacts that (as I have pointed out in an earlier post) dwarf the average agriculture or business development program with similar outlays. What we need to do next is start adding up lessons like this (particularly in Africa) to see where cash works well and where we might need something else. One place to start, is to consider the different populations targeted by these programs. Zambia aims for those in really poor areas, but clearly is getting households with a lot of working age adults by targeting on young kids. Programs like Kenya and Malawi on the other hand are specifically targeting households where working adults are likely to be missing. This gives us a start, but let’s see what we can learn about cash transfers for social protection versus (or and!) protection as we start to add things up.
The emerging results are pushing me to the conclusion that in some cases cash is probably a good option for increasing households farm and business outcomes. But as this work is also showing (and as the literature shows elsewhere) it definitely isn't always a good option for these outcomes. And, of course, this raises the question: when is it? I am not going to answer this last question, but it is worth looking at the evidence coming from the PtoP project to see what's going on.
Let's start with Malawi. Here we have two papers which look at the impact of Malawi's Social Cash Transfer program. This is a relatively big unconditional transfer, clocking in around $14 a month which in these communities was enough to raise living standards from the bottom decile to above average. It's targeted at poor households, using a combination of the dependency ratio, having low food consumption and a lack of assets, combined with additional community targeting. Using a randomized phase in, two papers look at a range of productive outcomes (note: they have a small number of communities, so balance isn't great so these results are going to come from both the randomized diff-in-diff and matching estimates).
What do they find? Boone, et. al. (gated version here) show this transfer had significant and large increase in the ownership of farm tools (hoes, sickles, axes) and a fairly massive bump in the ownership of chickens (up by about 50 percentage points). Household adults shift to spending more time working on their own farm, with a concomitant drop in casual labor for others. In another paper, Covarrubias, et. al. show significant bumps in goat and cattle ownership. So for Malawi, we seem to be seeing a shift in working patterns and increases in productive assets.
Off to Zambia, where a paper by Seidenfeld and co. looks at the impact of the Child Grant Program. This is also an unconditional transfer, targeted to very poor areas and, within them, households with a child under age 5 (at the start of the program, they look for households with kids at 3 or below). Households get about $12 per month - a significant amount relative to household expenditure. Seidenfeld and co. use the randomized (community) assignment to the program to look at impacts.
And the impacts are fairly impressive. Food expenditure goes up and extreme poverty goes down by 5.4 percentage points. In terms of productive impacts, this program sweeps the boards in a rather strikingly consistent way. Operated land goes up by 18 percentage points (or 34%), with household spending 18 percentage points more on inputs and an increase in farm tool ownership. The value of harvest goes up by 50 percent relative to baseline (this is significant only at 10 percent). Households are more likely to be selling crops and they earn more from sales. Livestock ownership also goes up (particularly for chickens, goats and ducks). Off the farm, households are 17 percentage points more likely to have a non-farm enterprise, it's more likely to have operated for longer and they have way higher revenues and profits than the control group. The household labor story backs this up nicely with a shift from agricultural wage labor to family agriculture and non-farm work. With effects like this, one has to wonder about general equilibrium effects. Thome and co, as part of the Seidenfeld and co report, use a local economy general equilibrium model and seem to find positive spillovers.
The other two papers that have come out of this work so far show less striking and consistent effects. In an evaluation of the Kenya Cash Transfer Program for Orphans and Vulnerable Children, Asfaw et. al. find a bit on livestock ownership, but only when they push on heterogeneous impacts. And in an evaluation of the Livelihood Empowerment Against Poverty Program in Ghana, Handa, et. al. don't find much in the way of productive impacts (in this case, sporadic payments of the transfer may be part of the story).
So, clearly not a uniform win. But in some cases, like Zambia, we're seeing pretty large productive impacts that (as I have pointed out in an earlier post) dwarf the average agriculture or business development program with similar outlays. What we need to do next is start adding up lessons like this (particularly in Africa) to see where cash works well and where we might need something else. One place to start, is to consider the different populations targeted by these programs. Zambia aims for those in really poor areas, but clearly is getting households with a lot of working age adults by targeting on young kids. Programs like Kenya and Malawi on the other hand are specifically targeting households where working adults are likely to be missing. This gives us a start, but let’s see what we can learn about cash transfers for social protection versus (or and!) protection as we start to add things up.
Join the Conversation