If you just took a break from home-schooling the kids to read this, then this post is for you. We know that female firms have lower profits than male firms (e.g. data on firms in Africa here), and we have some suggestions as to why (sector, for example). But we lack good clear insights into what precisely a lot of the mechanisms are, particularly when profit differentials persist within sector.
Into this space comes a nicely done forthcoming paper from Solène Delecourt and Anne Fitzpatrick. They look at firms in Uganda. Focusing on shops that sell drugs (anti-malarial drugs to be precise), they find no outlier here: female owned firms have profits that are, on average, 60 percent lower than their male counterparts.
Let’s start with a bit of context. First, Uganda has a very high fertility rate. And as Delecourt and Fitzpatrick point out, this means by age 30 the likelihood a woman has had at least one child is close to 100 percent. And these women breastfeed for awhile: data indicates the median duration is 19.8 months.
Delecourt and Fitzpatrick have a unique dataset. The data was collected as part of an audit study, so not only do they have data on firms, they also have data from actual and (researcher-sent) mystery shoppers. This gives them unannounced visit insights into what is happening in firms.
And that turns out to be the key variable (and insight). Women who have a child (under 2) in the shop with them, as observed by enumerators, have 48 percent lower profits than other women. Zero men have a child in the shop with them in case you were wondering.
Now this is correlational, not causal. And Delecourt and Fitzpatrick do a nice job of walking us through potential other explanations.
Is it that women with a young child work less hours than others?
Nope. Both sets of women (the yes-baby and the no baby storeowners) are working 6 days a week, 13 hours a day.
Is it that young children distract an owner cognitively?
Nope. They score the same on a cognitive ability index. And a host of other observables.
Is it that customers prefer not to see kids in the shop they are patronizing?
Nope – at least on the face of it. They have the same number of customers. And these customers are identical on observables and the reasons for why they chose that shop.
Is it that owners with smaller kids are more likely, for some reason, to be in less profitable markets?
Nope. Keep in mind that everyone is selling very similar (or identical) products here. And these are products where customers prefer convenience versus other store attributes. But these results hold with market level regressions.
So what’s the mechanism?
Well, it turns out the mystery shoppers that are sent in to buy antimalarial drugs complete a sale significantly (at 10 percent) less times in a yes-baby shop. Why? The yes-baby owners report being completely out of stock of malarial drugs during 13.9% of the visits, while the no-baby owners are out of stock 5.7% of the time. It’s interesting to note that men are out of stock 3.3% of the time – keep in mind that almost all of the no-baby store owners are still moms. Getting stock isn’t trivial for these businesses – they don’t just click on a website. It usually involves going to a wholesale outlet and buying things in person. And that’s harder with a baby (likely breastfeeding) in tow.
Delecort and Fitzpatrick go the extra mile to show us this is the mechanism. First, there isn’t a significant record keeping gap between the yes and no-baby owners. But the yes-baby owners are 15 percentage points more likely to report that they wait until stock is out before they restock (this isn’t statistically significant).
This is a neat paper. Maybe even more neat since it’s a secondary use of the data, which were gathered for another study. It takes these data and does some thoughtful inferential work to provide insight into the mechanisms by which the norms on childcare might be a big chunk of the story in male-female firm profit differentials. And now, back to helping with trigonometry homework.