Bouncing along a dusty road in Ghana, I had an eye-opening conversation with a colleague who was supervising a survey we were doing. It turns out he had been offered a more prestigious job, with a significantly higher salary, elsewhere in Ghana. But he was turning it down.
Why? I asked. He went on to explain that this other job was much closer to his extended family. And this meant that there wasn’t a real raise in salary in it for him. In fact, he expected he would end up with less each month. Thus, while I was in the midst of my PhD, I discovered the other PHD – the pull him down syndrome.
This pressure to redistribute or, in some cases, the urge to see successful people brought down to earth, is something that is oft discussed in the context of Africa – a quick search shows this featuring in the national discussions in Ghana, South Africa, and elsewhere. And now a recent paper by Pamela Jakiela and Owen Ozier provides some nice empirical insights into this.
Jakiela and Ozier jet off to Western Kenya (the cradle of the new development economics), where they run an experiment across 26 rural communities. The set up is nifty. Subjects are randomly given an initial endowment of either 80 or 180 Kenya shillings ($1.05 to 2.35 – a significant figure relative to daily wages) in private. The subject then chooses to allocate this to a zero-risk savings account (for the duration of the short experiment) or a risky investment which pays off 5 times the investment with probability 0.5 and a complete loss of the investment with probability 0.5 (the payoff was decided by a coin toss – by the subject).
On top of this, participants were assigned to one of three other treatments. In the public treatment, they had to make an announcement to the whole group of how much they had put in the risky investment and what had happened in the coin toss. In the private treatment, everything was kept secret. Finally, in the price treatment, individuals would have to announce their investment outcomes unless they paid a price from 10-60 shillings (again, randomly assigned) to keep things quiet. In this treatment, the price of keeping information private was revealed before the investment decision, but the subjects could wait to make the decision on privacy until after the outcome of the investment was revealed. In the end, there are two ways to conceal information here – first, subjects could choose to only invest 80 shillings if they got the larger endowment – thereby tricking those watching into thinking they didn’t get lucky at the start. Second, of course, if they were assigned to the price group, they could pay to keep information secret.
Now the experiment is stratified by gender, so this allows Jakiela and Ozier to examine outcomes separately for women and men. So what do they find? In terms of the initial investment, women who got the large, 180 shilling endowment, are more likely – 25 percent more likely -- to invest exactly 80 shillings in the risky investment when they are in the public information treatments than when they are in the private treatments. For men, there was no such difference.
Then it gets more interesting. Women who have close kin attending the experiment are much more likely to investment exactly 80 shillings when they get the high endowment, and they invest significantly less when the outcome will be public than when it will be private. These are the women driving the result – they do this more than women without kin present. Men, of course, don’t change their behavior.
So we might expect that the kin presence is a kind of proxy for the husband – after all this would line up with other results we’ve seen that the root cause is fear of expropriation from the husband. Jakiela and Ozier try to allay our concerns that this is the case: they interact public treatments with an indicator for the spouse being present. This is not significantly associated with investing 80 shillings or less. So, it appears that it is the family, not the spouse, which is driving this urge to hide the endowment.
Jakiela and Ozier then go on to look at how this behavior is correlated with village level indicators of wealth – assets, formal/skilled employment, wages, and fertilizer use. These results add a nuance to the pull her down syndrome as it turns out that income hiding is associated with lower levels of household wealth.
Finally, it is worth looking at what people would pay to avoid revealing their investment outcomes. Out of 690 subjects assigned to the price treatment, 627 could afford to pay to avoid the public revelation of the outcome (recall, they could lose their investment) and in the end 30 percent of these folks chose to do so – paying out an average of around 15 percent of their earnings. Although women were not more likely than men to pay to avoid the revelation, their wiliness to pay is significantly related to the amount of income they get and this isn’t so for men.
So this gives us something to chew on regarding the pressures for redistribution within communities – and how this might differ across gender lines. And makes us think again about how and when we target women for either transfers or programs to boost their earnings.