A while back, Dylan Matthews asked where the Laffer Curve bends. A few numbers were thrown out. I would tend to estimate something around 70%. However, I want to talk Laffer Curve Theory for a minute.

So leaving behind for a moment the important issues of avoidance and evasion it’s key to remember that there is no particular reason to think taxes in and of themselves have any impact on people’s incentives.

What matters is the net of tax return to the activity, working, saving, etc. So suppose that for whatever macro-economic reasons the return to being a hedge fund manager has increased 2 fold over the last 10 years. Now suppose that I raise that hedge fund manager’s total marginal tax rate from 44% (including state and local) to 72%, so that her take home ratio falls from 56% to 28%.

Even though I have radically increased her taxes, her return to work is essentially the same as it was a decade ago. If a decade ago it made sense to become a hedge fund manager and work 90 hours a week, it will still make sense today.

This fact makes the observation that there is little secular trend in hours worked per American important. There is strong trend in the return to working and it is going up. There is also fairly turbulent flow in the mix of labor, more women / less men, etc.

However, the total hours worked is per American is not changing that much and what change there is seems to be dominated by the unemployment rate. Now in the interest of intellectual honesty this goes both ways. Supposing I have a constant number of structurally unemployed workers, then adding more workers to the mix will lower the unemployment rate. However, I think with other data series we have ample evidence that the unemployment rate is dragged around primarily by monetary policy.

The graph below shows the average hours worked per working age American in blue and the inverse of the unemployment rate in red. They track each other pretty closely. I would suggest this implies that the amount of work done in America is determined by the number of Americans who can find work.


This leads to the astounding conclusion that the vast area of social, economic and policy changes that we have had over the last 40 years have had very little impact on the total amount of work being done in America.

To wax extremely nerdy for a moment, but this is exactly what you expect if you had labor supply decisions made at the extended family level and something functionally akin to logarithmic utility. In more human terms, that says that we don’t just think of ourselves when we decide to work but our children and our parents and to a lesser extent our siblings and cousins. It also supposes that the relative happiness generated by having more stuff and having more free time is more or less written into human DNA and is not altered that much by relative costs of each.

As an interesting side note, while something like logarithmic utility would imply that taxes and wages “don’t matter” they do not imply that “welfare” doesn’t matter. Under logarithmic utility welfare dramatically reduces the incentive of people to work. Explaining that dichotomy is beyond the scope of this post but it is easy to verify mathematically.

So suppose that utility is something like logarithmic. Then extend family behavior is invariant to real wage and things like entry of women into the workforce. The decision of women to work more, under this formulation, is implicitly the decision of men to work less.

This isn’t by any means a flawless representation of what’s going on but we can’t ignore the fact that it predicts the historical data. A massive lifting of labor force constraints on women seemed to have only a small and temporary effect on labor supply. Taxes, unions, trade, etc are barely noticeable.

Again even the trend that we do see could largely be explained by changes in the unemployment rate. Changes that I believe are mostly driven by monetary policy.

Some commentators – I am thinking Reihan Salam in particular – might wonder how this could ne possible when we can easily observe massive social change and differences in work patterns over the last 40 years. I would first say that the aggregate stats are what they are and we have to accept that whether or not they make sense.

The point of science is to come up with theories that explain and predict the data. Not, to adopt a theory and then attempt to impose it upon the data.

Intellectually though I would say these facts suggest that there are deep fundamental forces at work and that what we think of as our culture and social norms are actually products of those larger forces.

For example, one might suggest that it is precisely because women can now reasonably earn a living as single mothers that more families have children out of wedlock and the social culture allows the fathers to skip out on their children. If the economics did not allow these women to make it on their own then social culture would respond by berating and tracking down deadbeat dads.

Likewise it is precisely because the returns to skilled labor are so high that aggressive professionals will wind up working 80 hours a week just out of college. However, that same high return to skill is why so many of them feel fine taking 6 years to graduate college or going on to graduate school in a field unrelated to what they will actually do. If the return to labor was less they would get themselves out into the workforce sooner, work fewer hours when they go there, retire later and actually end up supplying roughly the same amount of labor over their lifetime.

In short, there are all of these trends going on but they are honed and balanced against one another by the basic economic returns. If the “fundamental” family labor supply function is nearly vertical then all sorts of complex social dynamics will evolve to ensure that total labor supply is constant in the face of changing incentives.