Suntory and Toyota International Centres for Economics and Related Disciplines (STICERD) LSE RSS Contact Us YouTube Twitter

A large fraction of the world's poor earn their living through informal small-scale businesses. While their average productivity is low, many micro-enterprises have high marginal return to capital - raising the question of why so few are able to grow. One understudied hypothesis proposed in the literature is that pressure from family and friends to share income reduces production incentives. This project examines whether this distortion, "kinship taxation," is a constraint on micro-enterprise growth. Specifically, I ask how large this kinship tax is, who is most affected by it, and what effect it has on economic efficiency and income.

Funding from STICERD paid for a pilot study in early 2014, conducted mostly in slums near Nairobi. Participants in this pilot were eager to discuss the effect of kinship taxation on their lives. For example, a 40-year-old female entrepreneur said: "I sell second-hand clothes without anyone knowing, far from home. I hide from my friends because I believe not all friends will be happy with my success, and from family to create a picture that I have no money, for them to work hard for their own money. My previous business, a street-side restaurant, failed due to my in-laws using me for money, yet I wanted to expand it." To study this phenomenon, I combine two types of analysis. First, I describe results from a lab-in-the-field experiment to elicit kinship tax rates. These kinship tax rates tell us the magnitude of the distortion faced by each participant. This allows me to answer the first two questions: how large the kinship tax is, and who is most affected by it. Second, I collect firm-level data from entrepreneurs who participated in the lab experiment to study the effect of kinship taxation on micro-enterprises.

I measure marginal kinship tax rates for 1805 participants in 17 villages across Garissa County in Eastern Kenya. The lab experiment used to do this elicits a person's willingness to pay to hide income. Participants were asked to choose whether they prefer to receive $5 "in public", announced to their co-participants, or to receive a smaller amount of money "in secret". By asking a series of questions with different secret amounts, I can determine the maximum fraction of income each participant is willing to pay to forgo others knowing about it. I show in a simple model that willingness to pay to hide income is a summary measure of the distortion from kinship taxation a person faces. This holds in the presence of altruism as well as any other benefit from sharing income, such as increased status.

I find that 23 per cent of participants face costly kinship tax distortions. Among those who do, the mean marginal kinship tax rate is 50 per cent, suggesting potential distortions are large. Entrepreneurs face higher tax rates than non-entrepreneurs. Tax rates also are higher for men, and increase with education and with number of siblings. Surprisingly, marital status is not significantly correlated with kinship tax rates for either women or men. In this setting, it seems to be relationships between households, rather than within, that result in kinship taxation.

figure 1


Before using these estimates to measure inefficiency, I confirm that they reflect true preferences. First, when asked why they prefer a smaller amount "in secret", answers such as these are typical: "It can be I need the money and if I take it in public, friends and families can come and ask for some of it, which I cannot refuse"; "There are so many people who are poor, they will ask you to give them something"; "People will ask me and I will give out, remaining with none." Second, participants hide from friends and family, but not from strangers. That is, their relationship to co-participants matters in deciding whether to hide money from them. Finally, educated participants are more likely to hide income, which makes it unlikely that these choices are mistakes.

To measure the effect of kinship taxation, I ask entrepreneurs about their businesses. Capital stock, labour use, and profits are combined with a simple model to impute distortions affecting each entrepreneur. To estimate the effect of kinship taxation on productivity, I combine these results with those from the lab experiment described above. Specifically, I perform a counterfactual exercise where capital and labour inputs are reallocated across firms after removing the disincentive effects from kinship taxation. In my sample of entrepreneurs, removing distortions from kinship taxation would increase productivity by 26 per cent. This is the first estimate, to my knowledge, of the cost of kinship taxation on output or productivity.

To put this number into context, I exploit an additional feature of the data: entrepreneurs in my sample were also participants in a randomized cash transfer study. In this separate study, treatment households received 18 months of transfers equal to half of mean household income. Entrepreneurs unable to borrow were able to use these transfers to accumulate capital. Indeed, I find that capital stocks of entrepreneurs are significantly larger for the treatment group. Exploiting the structure of my model, I can calculate the gain in productivity from removing all distortions that affect capital and labour inputs. These include credit constraints but also capital and labour market imperfections - the data do not allow us to estimate these separately. Removing all such distortions simultaneously increases output of entrepreneurs who received cash transfers by 4 per cent, and those in the control group by 20 per cent. This supports the idea that input distortions are coming mostly from credit constraints, which are alleviated by the cash transfer. Given these results, kinship taxation seems to be at least as costly as credit constraints. Further, if before doing the above procedure I first remove distortions from kinship taxation, the increase in output for both treatment and control more than double. That is, removing one type of distortion increases the gains from removing the other.

These results speak to the multiple constraints facing entrepreneurs in poor economies, and the fundamental challenge in spurring microenterprise growth. This may help explain why the literature finds modest gains for microenterprises coming from microfinance, as well as the uneven effect of cash grants. If entrepreneurs have insecure property rights over their output, relaxing credit constraints or fixing other markets may lead to disappointing results. It is possible that offering long-term, reliable welfare benefits (such as unconditional cash transfers) reduces the reliance of entrepreneurs on their extended family, and allows them to more efficiently grow their businesses.