What sort of data should an AS/AD graph display?
I don’t think the current AS/AD paradigm is very useful at all. For one thing, the Y axis label is hopelessly confused. Should an interest rate be there, and if so, which one? There are literally millions to choose from, when you account for duration, risk and expected inflation. Should it be the price level? No, because there is no such thing as an aggregate price level. All prices are relative prices, and no economy can have an average relative price of anything but 1. It can be the change in the nominal price level, or inflation, but the 70s showed that to be unreliable. If people expect inflation, its’ effects will be altered. Workers start to demand wage increases equal to the rate of inflation and their wages become sticky downward at the rate of inflation, not at 0. Inflation becomes integrated into interest rates. So any Y axis must include some measure of expectations and change relative to those expectations. Rational Expectations models may be a bit extreme, but you really can’t fool everyone all the time.
As a quasi-monetarist, I think that the most important variable to gauge the demand side of the economy is the rate of spending. If expenditures are increasing more rapidly than the previous trend line, demand is high and if expenditures are below trend, demand is low. Unlike Keynesians, Quasi-Monetarists don’t care whether the spending comes from government or the private sector, and unlike Monetarists, they don’t assume you can infer much from monetary aggregates. So, I think that one axis should display the rate of spending realative to trend.
Nominal spending has two components: the quantity of goods exchanged and the price of those goods. We would like to have some way to say: given a change in spending, will it primarily change the quantities of goods or the prices of goods? If all scarce resources are fully employed, adding money will only cause people to bid up the resources’ price. There will be no quanitity adjustment because the productive capacity of the economy is already fully used. If, however, there are lots of unemployed resources, an increase in spending might cause people to start working again and employing those resources in productive ways. So the second piece of information we want to include is the level of slack in the economy. How much will an increase in spending cause prices to rise? I’m trying to figure out a way to graph these two data in a meaningful way that someone could use to describe the economy. I haven’t figured anything good out yet, but I’ll keep thinking about it.
Update: Tyler Cowen comments.