Saturday, November 3, 2012

Guest post from Andrew Bossie on Prices in Disaster Areas

My own thoughts, before pasting in Andrew's post, is that certainly prices wouldn't be the only way to direct disaster relief. After all the places that need goods and services most are probably going to be the ones with sparse or missing markets! But for the areas that are doing not great but not terribly, I imagine prices would be a good signal to direct aid.


So, Dan’s post the other day about allowing the price mechanism to determine disaster relief really stuck in my craw. I’ve been chewing it over for the last 24 hours or so and I’ve come up with a highly simplified islands model to explain my objection to the idea. I'm curious how it holds up. In a nut shell, the price signal confuses ability with willingness to pay and that in order for the price signal to give you any meaningful information, you actually need more information (demographics) to come to any meaningful conclusions about what prices tells anything who is not a profit maximizing firm.

Imagine you have two islands. One called Park Slope and another called East New York. Both Islands are far enough apart so that people from Park Slope cannot get to East New York and viceversa. Gas is priced the same on each island because of perfect competition in gasoline markets and no difference in transportation costs to the two islands. The only difference between the two islands is that the people in East New York have half of the income of the people on the island of Park Slope.

Now, assume that on each island gas initially costs $5 a gallon when a big storm causes gasoline supplies to be disrupted. The ruler of the Confederate Islands of Brookland (CIB), Lord John Russell, has decided to allow the market to sort out the response to the storm. Since the marginal buyer of gasoline in East New York can pay $10 a gallon and the marginal buyer of gasoline in Park Slope can pay $20 a gallon the prices are set as such. Now, what does this tell you? Is gasoline more dear to those in Park Slope than in East New York?

Now assume that Russell is deposed (unlikely, though, because he has kept the influential isle of Park Slope happy) and he is replaced with Richard Nixon. Nixon imposes price controls at $5 a gallon for both island.

Now, assuming that gasoline still runs out because the supply is still inadequate. The quantity signal alone is unambiguous and tells us what is wrong perfectly adequately. While allowing only adjustments in quantity is inefficient, it is clear. The price signal actually introduces noise here. For instance, high prices mean that any limited private supply form outside the islands will flow to Park Slope. Potentially, Park Slope can have more gas at higher prices than East New York does.

Now, if the government has a tanker of gas and it operates as a profit maximizing firm, then we are done with this discussion. Clearly, the government should send the tanker to Park Slope. But if there is any other goal pursued by the government the price signal is pure noise.

Furthermore, suppressing the price signal forces private firms to act as surrogates for the government. If gasoline prices are kept at $5 suppliers will be indifferent between supplying both islands.

A stronger point to make is to say that it is probably reasonable to assume that--due to their more limited means--the marginal benefit of relief is higher for East New York than Park Slope as they are less in a position to respond and rebound from a disaster. Perhaps it makes sense to let gas prices to rise to $20 in Park Slope, allowing the market to take care of it and subsidize East New York by sending all the government gas there.

Anyway, I consider Daniel’s original point a natural extension of the idea that the government should be more responsive to market signals because it will make it more efficient. That efficiency of the price signal, however, is only meaningful to the government if you accept it as strictly a profit maximizing entity.

1 comment:

All anonymous comments will be deleted. Consistent pseudonyms are fine.