Quite a Leaky Bucket; In Fact My Feet are Soaking Wet

In one of our research projects, Doug and I created an environment in which individuals could endogenously choose not only personal decisions leading to voluntarily provided public goods, but they also had the power to vote to tax their group for the same purpose. In the enforced taxation option, we wanted to capture what Arthur Okun had famously called "the leaky bucket," namely, the idea that government transfer programs designed to help the poor suffer a loss of efficiency. We chose a parameter of 20 percent loss of efficiency largely for reasons specifically relating to the experiment, not because we thought that 20 percent (as opposed to 10 percent or 30 percent) was the actual level of the leaky bucket.

Nevertheless, we've had a fair number of people conjecture about this extrapolation: was 20 percent "too big," or "about right" when one is talking about actual governments. This article in the Washington Post is a great illustration of the problem of the leaky bucket, and why we were justified in wanting to incorporate it at some level in our experiments.

Thanks to Instapundit for the tip to the link.