Standard economic theory would suggest that a one-time infusion of cash should have at most a temporary effect on business profitability – over time, individuals facing high returns should be able to re-invest business profits and bit-by-bit bootstrap themselves up to the steady-state size. Yet in an experiment I did with Suresh de Mel and Chris Woodruff in Sri Lanka, we find a one-time grant has sustained impacts five years later on male microenterprise owners. One possible explanation for this is behavioral effects – individuals lack self-control to continually set aside and reinvest the small absolute amounts of returns, but once the money gets into the firm, there is a flypaper effect whereby it stays there.
Can a short-term boost to savings get the right mental account going?
I was therefore very interested to come across a new paper by Simone Schaner which looks at how a short-term boost to savings can have long-term impacts on enterprise ownership and performance.
In her earlier paper (one of my favorite development job-market papers from the past few years) Simone worked with married couples in Busia, Kenya (aka RCT-central) to test the hypothesis that a reason for inefficient household decision-making is that husbands and wives have different time preferences in some households. 749 couples were given an offer to open up to three bank accounts – a joint bank account, a bank account just for the husband, and a bank account just for the wife – with the account opening fees paid for them. For each account they opened, they then received a 6-month interest rate subsidy, which was randomly chosen – varying between 0 and 20 percent annual rate for the individual accounts, and 4 and 20 percent for the joint account. The job market paper then drops a few polygamous households and then looks over this 6 month period at the extent to which couples with more or less aligned time preferences take advantage of the relative differences in interest rates across these accounts.
The new paper goes back to these households three years after the initial experiment to measure the longer-term impacts of these short-term savings subsidies on households, 2 and a half years after the subsidies have ended. She finds:
Temporary individual interest subsidies lead to long-term increases in income, assets, and working in non-farm business activities. Going from a 0 to a 20 percent subsidy, while only resulting in a trivial amount of interest (99th percentile of interest received was just $5, mean was $0.24), led to:
- 11-12 percentage point increase in the probability of having any business capital or income – appears to be from helping existing business owners stay in business and grow, rather than from creation of new businesses.
- A $38 increase in business capital, and a $15 increase in monthly income. However, it had no significant increase on jointly held assets.
In contrast, the subsidies to joint bank accounts had more limited impacts on income and no increase in non-farm business ownership, but instead increased investment in livestock and home renovations.
- The point estimates are much smaller for income and not significant (e.g. an insignificant $6 increase instead of a $15 increase in monthly income), but she typically can’t reject equality with the individual treatment. For non-farm business ownership, the point estimate is 0, and she can reject equality.
- The households receiving this joint interest subsidy are 9 percentage points more likely to have made a renovation, and appear to have made increases in livestock holdings.
Is this mental accounting?
Simone considers several possible explanations for the long-term impact of a short-term savings subsidy, ruling out non-convexities in production, learning about the value of bank accounts, and simple savings habit formation. Her preferred explanation is that of mental accounting – she suggests that joint subsidies could have encouraged couples to open mental accounts for shared investment goals, while individual accounts could have served to set up a mental account in the business owner’s mind for their private saving goal.
This seems plausible, and is certainly consistent with the speculation I started this post with. However, the evidence for it is basically a) finding a difference in impacts of the joint and individual subsidies; and b) a process of elimination, trying to rule out other explanations.
The challenge for this paper, and others in this literature, is trying to provide more direct evidence on mental accounting. This is difficult to do empirically (maybe the neuroscientists will find the right tools to view mental accounts), but some things I would like to see future work in this area explore are:
- Much more frequent measurement: this paper - like several other recent savings papers – finds increases in income, and tries to infer what returns to capital must be based on assumptions of reinvestment behavior. Ideally we would like to see how frequently business owners are reinvesting in their firms, to see if it is little by little bootstrapping of returns, versus saving up for lumpy purchases.
- Directly asking about mental accounts: we find it hard to ask these sorts of questions, but the idea pushed by Drexler et al in their business training work of teaching business owners to have separate drawers or places for storing business and household cash, suggests we could do more with just measuring how people manage their money for business versus household uses.
- Deriving some sharper testable theoretical predictions and identifiers: I don’t know the theory as well as I should here, but apart from a general view that money need not be completely fungible across different uses or accounts, I haven’t seen a lot of sharp predictions that would help in i) identifying ex ante from survey questions which individuals are more likely to use mental accounts; and ii) helping distinguish mental accounting from other types of transaction costs and explanations for some of the times one finds money getting used differently.
The finding that a short-term boost to savings can have long-term impacts on income and enterprise ownership is an important one, so I look forward to seeing more studies which explore further these mechanisms.