Hi everyone. I am excited to have Jed Friedman join us in this blog. Jed was one of the founders of Development Impact, way back in 2011.
Hi Jed, it is great to have you join this blog and give us your reflections on your recent working paper “Cash is Queen” with Sreelakshmi Papineni, Paula Gonzalez, and Markus Goldstein [another DI founder]! Your study presents the results on the direct and spillover effects of short run, unconditional cash transfers (CTs) paid to the foremost female household member in ultra-poor households in Northern Nigeria. CTs were distributed between September 2015 and March 2017, in total 75,000 Nigerian naira over fifteen months, which is equivalent to approximately USD 693 in 2015 PPP terms.
Thanks for joining!
There is lots to discuss in your paper. Let’s start with some main effects. Your study examines the impact of the CT on uptake of women operating an enterprise (small, often home-based, and nonfarm income generating activities which include a shop, cooking food to sell, oil-processing, or crop-processing). You find that beneficiary women started profitable non-farm enterprise activities. Moreover, in the same village, other women (non-recipients) also started businesses, albeit later (a year later) and the increase in rate of business ownership was not as large as with beneficiary women. So the gains from the CT went beyond the selected households, and to many of their neighbors.
The paper notes that markets for financial services (such as credit and insurance) are very sparse in these villages, and concludes that reliable cash enabled women to overcome credit constraints, and financed their new small business. Can you say something about the process by which their neighbors (non-recipients) overcame their credit constraints to start a business without CTs?
Hey Kathleen, it’s great to be discussing our paper with you on DI and thank you for reading our new working paper! Many years in the making (!), our paper engages the exciting (to our mind) possibility that social support, in this case a time-delimited cash transfer to ultra-poor households in Northern Nigeria, has longer-lasting beneficial impacts on both the beneficiaries as well their neighbors who did not receive any cash support. In this regard, our paper engages the “multiplier” literature looking at cash transfers, such as the great papers Egger et al. (2022) and Walker et al, (2024), that find powerful economic spillovers in certain contexts. For our context, the “slack productive factor” spurred to activity by the additional cash is female labor supply, specifically female micro-entrepreneurship.
As in many other settings, women in the study area are underemployed relative to their potential, partly due to restrictive norms around women's work. But when there’s a cash transfer in the community, the injection of resources seems to tip the scales: it becomes worthwhile for some women to start businesses. At first, the new entrants are cash recipients who likely use some of the newfound resources to finance business entry. Eventually, however, the increase in economic activity, and general increase in community purchasing power from the transfers, also induce women who aren’t directly receiving cash transfers to establish new businesses. For these non-beneficiaries, we see that the main source of business finance is simply drawing down their limited savings, but it appears the increased earnings potential from the now wealthier local market (as well as perhaps the demonstration effect of viable female-led micro-enterprises) draws more women into productive activity.
What’s especially striking is that the more cash transfer households there are nearby, the more likely it is that a non-recipient woman starts a business. This indicates that local market conditions – the available demand for locally produced goods– is a key factor. The businesses that emerge are often locally oriented enterprises that let women sell goods and services within their immediate area. The pattern of new entry hints that women create businesses that are complementary, such as locating oil-processing enterprises near to crop-processing ones, or co-locating cooking enterprises that can use the same capital inputs (e.g. cook stoves, pots and pans, etc). So in effect, the transfers do more than just raise individual incomes during the period they are received – the transfers seem to reshape local economies in a way that unlocks women's market-oriented productivity.
* * * * * *
Thanks for this excellent exposition! In the wider set of results, the study shows that in households with the CTs, women were more likely to work, agricultural yields and business profits were significantly higher, and the whole household consumed more and had a more diverse diet. Amazing impacts. Back to spillovers: non-recipients also experienced positive gains. Can you unpack a bit the pathway by which non-recipient households experienced improvements agricultural yields?
This is a great question — agriculture remains central to livelihoods in this setting, and it’s where we see some of the clearest gains both for beneficiaries and no-beneficiaries. Unfortunately, we don’t have particularly flexible data on the farm businesses, as we can’t speak to the dynamics (as we do with non-farm enterprises) since we only have data one year after the cessation of transfers. We also don’t have particularly granular data, as outcomes are measured at the household level, and typically reported by the main agricultural decision-maker, most often the foremost male in household. Since farming is the primary income-generating activity for many households in our study, changes here have a big impact on overall welfare. Importantly, our sample focuses on the ultra-poor, many of whom started with essentially no reported yields at all — so when we see improvements, they’re often going from nothing to something, which is meaningful in both relative and absolute terms.
The evidence points to a few plausible channels. We observe increases in spending on agricultural inputs, as well as greater deployment of family labor (mostly male) on plots, both of which are consistent with the uptick in yields. And while we can’t directly test this, it’s reasonable to speculate that the broader cash injection into the local economy — and the accompanying rise in microenterprise activity — may have improved the ability to invest in agriculture as there is more available cash to invest and more local demand for farm products (remember there is an increase in crop-processing, oil-processing, and cooking businesses in these communities). While most of the agriculture produced is consumed by the growing households themselves, some production is sold or traded on the local market. Even though agriculture doesn’t always get the spotlight in cash transfer evaluations, it’s a crucial part of the broader story of economic activation we’re seeing.
* * * * * *
The paper looks at an impressive range of impacts. You also find that there was a pivot to greater joint decision-making power within married couples. Do you have a hunch if this is the result of the cash going to women within the households itself, or a benefit from higher income levels or maybe greater income security for the CT households?
To measure women’s economic empowerment, we use a modified version of IFPRI’s pro-WEAI index, which gives us a structured way to capture shifts in agency, resources, and decision-making. For the more nuanced aspects of household bargaining, we incorporated a vignette-based measure to understand how decision-making dynamics change between spouses. What we see is that many women move from a position of having no say in household decisions to being part of joint decision-making processes with their husbands — which, in this setting, is a meaningful shift.
Interestingly, this pattern emerges not just among cash transfer recipients, but also among some non-beneficiaries. That leads us to think the mechanism has to do with the broader rise in women’s economic participation — as more women enter the labor force and generate income, they gain leverage and recognition within the household.
That said, we find only limited change in personal gender attitudes or perceived social norms about women working outside the home. Our interpretation is that enterprise activity may challenge some norms — particularly those around public visibility — but when work remains close to home, women can navigate these changes while still aligning with prevailing gender expectations. It’s a subtle but important distinction in how empowerment unfolds in this context.
* * * * * *
Lastly, I’d appreciate hearing your reflections on spillovers that can go in the other direction. Specifically, how do you think about this new Nigeria paper (with positive spillovers from CTs) in light of your earlier work (with negative spillovers from CTs).
Yeah this is a great question. Earlier work with Deon Filmer, Eeshani Kandpal, and Junko Onishi looked at a very different spillover induced by a pilot conditional CT program in the rural Philippines. In that setting the cash transfer, which led to many benefits among the beneficiaries, appeared to shift the relative price of certain goods important for the nutritional status of young children, namely animal-source protein rich foods, such as eggs, which are critical for physical growth in early life. This shift in prices was severe enough to affect the affordability among the non-beneficiary households, resulting in a rise in stunting rates among young children in households that didn’t receive the cash transfer.
It’s a striking result but a few caveats are in order. For one, the results only emerge among relatively remote villages – villages that are less integrated in regional markets and more reliant on the local production base for animal-source protein rich foods. In less remote villages, any rise in local demand for these food goods (as a result of cash receipt) can be presumably accommodated by imports from the broader region and so no relative price shift occurs. Hence no adverse spillover to non-beneficiaries. Further, many households in program villages were eligible for the cash transfer, so the aggregate cash transfer to the village economy was much larger than in our Nigeria setting discussed above. In some of the Philippine study villages, 80 or 90 percent of the households with age-eligible children received the transfer. We observe that as this level of program saturation increased, the greater the adverse spillover to non-beneficiaries.
On the other hand, Egger et al. 2022 studied prices in Kenya after a cash transfer and saw little change in the overall price level or in the relative prices of various goods. So, clearly, the context matters and the relevant dimensions of context include (a) the degree of integration of the study area with wider regional markets (b) the local market structure of the key commodities – in this case eggs, fresh fish, etc. and (c) the level of program saturation in the study population. Even when the transfer value is large, if only a small fraction of the population receives the transfer, there may not a large enough shift in local demand to affect local prices. In the Nigeria case, unfortunately we don’t have sufficient price data to look at this issue (although we can say that key asset prices such as for livestock are unchanged). However, only about 8% of the total village population in the Nigeria study was treated, so even with the spillovers from beneficiaries to other ultra-poor neighbors, there was likely not a sufficiently large increase in the local aggregate demand to shift relative prices. In general, collectively, we need to think much harder about how program design interacts with population characteristics, local market characteristics, and other site features to produce beneficial or adverse spillovers.
* * * * * *
Thanks Jed, I really appreciate that you took the time to write my blog for me 😉
It has really been a pleasure to speak with you. As you mentioned above, I’m a DI alum, with the effort from the beginning, so it’s especially great to return to discuss this new work.
Join the Conversation