This post is for my buddy Dan, I'll try not to make it too technical for others. See his full comments on Wednesday's post. In short he asks how can wages for adults go up enough to compensate for a ban on child labor. Basu generally seems to think child labor bans won't work, because if they are not monitored (tough to do) then some firms will hire child workers and be successful, while goods produced with child labor will be cheaper and the workers in the developing country will be worse off. It is hard to test if child labor bans work, since it is difficult to measure the impact of one ban or another, since we don't know what happens in the absence of the ban. But, here is a long explanation of how a ban might actually help in a rare case.
If a long theoretical explanation isn't for you come back later this week for discussion of churches going to the movies and a $250 memory test
So let's start with Basu and Dan's assumption that each person is paid their marginal product of labor. That is whatever the value of extra output each person produces that is their wage.
Now Basu assume that the marginal product of labor is diminishing. That is for each extra worker we have the next worker's extra output (marginal product) falls. So if we cut out children then adults wages go up.
Can adult wages go up enough to compensate for the loss of child wages? Here is when the good being made is important. In most cases where we talk about bans in child labor we are talking about exports from a developing to a developed country (say Bangladesh to the US).
So if adult wages are going to go up enough to offset the lost in child wages then the price of the export has to go up. However, as the price goes up people in the developed country buy less of the good.
In a more recent paper Basu and his co-author create a model to show that when the quantity demand changes a lot with little change in price (ie elastic goods) then child labor bans will make families worse off. However for goods that their quantity demand is not impacted much by price (inelastic goods) then bans might work.
The only problem is most goods made with child labor are very responsive to price (elastic) so a ban is unlikely to work.
So to sum up child labor bans might work if they can be monitored and the good being produced is inelastic (ie not often)