From the NYT

Nearly a quarter of South Korean men over 75 are still in the labor force, as are 14 percent of Japanese men. In the United States, a 10th of such men are working or seeking work, compared with half of 1 percent in France.

Put another way, a Korean man over 75 is more likely to be working than a Frenchman in his early 60s.

One point I hope to consistently make is even consumption functions which are almost completely insensitive to taxation – as I believe people really are – are not insensitive to transfer payment.

For example, the very simple, logarithmic utility function predicts zero net response of individual to changes in taxes or wage rates. Taxes decrease your direct incentive, but they also make you poorer which increases your indirect incentive. In a logarithmic function these two effect always exactly balance.

So, for example we shouldn’t be able to predict whether a woman making $7.50 an hour will work more or less than a woman making $750 an hour. The $750 woman faces a much stronger direct incentive, but couldn’t you easily imagine the $7.50 woman working double shifts to make ends meet. 

Now, note that from the point-of-view of the worker, the former is facing the equivalent of a 99% tax rate on the latter’s salary. That’s a 99% tax rate and we can’t tell off the top of our heads whether it will cause our hypothetical citizen to work more or to work less. If taxes have clear net effects the answer to this question would be laughably obvious – its not.

However, the logarithmic utility function does predict a massive drop off in labor supply from a program like Social Security or other defined benefit pensions. This is because in old age you are much wealthier than you would have been without the plan, but there is no increased incentive to work. The wealth effect is the only game in town and it causes you to quit working.

This will hold for almost any transfer program.

Let me give you another example of why this is important. Some people have suggested that we switch to something like the FairTax which levies a flat tax rate on consumption and then pre-bates revenue back to people.

If the argument I am presenting holds then the reduction in marginal tax rates from moving to a flat system will have little effect on incentives. Yet, the pre-bate will have potentially massive effects on incentives.

The pre-bate will cause marginal workers to simply drop out of the labor force and attempt to live only on the pre-bate. It will cause full-time moms to switch to part-time. It will cause people considering early retirement to go ahead and retire.

Even though you have dramatically lowered marginal tax rates I would expect to see labor supply go down. This is because taxes have offsetting incentive effects but transfers do not.