Thursday, September 25, 2008

Supply and demand?

When the word 'economics' is mentioned, many people first think of 'supply and demand'. This is considered by many to be the quintessential economic model - as close as one can get to a foundation for the discipline.

Astounding, then, that it is in large part a fairy tale.

I don't mean this as a criticism - fairy tales are useful. They teach life lessons, are memorable, and can even help to shape a culture, or provide insight on how to deal with life's difficulties.

Little Red Riding hood is a beloved cautionary tale, but few adults would think of taking it literally, or trying to attach real-world locations, names, numbers or even physical laws to the story.

To make my point clearer, consider a single farmer who owns an apple orchard. All she does is farm apples, year in and year out.

When customers at the nearby market are willing to pay more for apples, she farms more. The extra money makes the extra effort worthwhile. When these same customers are sick of apples, she farms less. It's no use farming apples that won't sell, or that will only sell at a very low price. This is supply.

Now consider the customers. When the farmer charges a high price for apples, they want to buy less of them. They can spend the money they save on oranges, or rabbit meat. When apples are cheap, they want to buy more of them. Maybe they use money that had been earmarked for a bushel of plums to buy three bushels of apples. This is demand.

Graphing supply and demand for the apple market is easy. We first draw two lines at ninety degrees to each other. We label one 'price of apples' and the other 'bushels of apples'. These are our axes. We then plot out all the combinations of price and quantity that the farmer is willing to supply at, and customers are willing to buy at. For example, maybe when apples are $1 a bushel, the farmer wants to sell 10 bushels, and when apples are $10 a bushel, she wants to sell 20 bushels. Connecting the dots will give us our 'supply curve'. Doing the same for demand will give us our 'demand curve'.

In general, the two curves will only meet once. At this point, the farmer is selling as many apples as she wants to sell, AND customers are buying as much as they want to buy. This is the 'equilibrium' point (after the Latin for 'equal weights', or 'balanced'). Economists like it, and solve for it, because the market will tend to move toward it. For example, if there's a shortage of apples - that is, the farmer sells less than people want - the price of apples will probably rise, encouraging the farmer to grow more of them. If there are too many apples, and a lot are rotting unsold or selling below cost, the farmer will lower production.

So far, so good. The model is useful, and it works.

The problem happens when we're talking not just about apples, but about apples AND plums, say. Or, on a larger scale, when an economist wants to look at the supply and demand for all goods produced by an entire country.

Suppose our farmer farms twenty different kinds of fruits, and we want to draw a supply-and-demand graph for fruit in general.

Let's start as we did before, with our axes. We have price on one axis, and quantity on the other.

Here's a problem: quantity of what? We could just label it 'bushels of fruit', but this wouldn't be very informative. We want to know how the farmer changes her willingness to supply when prices change. A bushel of apples is very different, in terms of value and effort required to produce it, than a bushel of raspberries.

The problem would be even worse if our farmer were a part-time shoemaker, and we had to add production of shoes to production of fruit.

Economists get around this by finding the VALUE of total production, and using THAT as quantity. In other words, 'how much does the fruit sell for at market? that's your output'.

If you're scratching your head at this point, you've probably understood everything correctly.

We want to know how the quantity of fruit supplied (and demanded) changes when price changes. However, our measure for 'quantity of fruit' INCLUDES the price level. This leads to a lot of problems, because it makes it very difficult to tell what's actually going on. Suppose the price of oranges rises. Total fruit production may rise as farmers are more willing to produce oranges. Even if production did not change at all, though, we'd STILL see a rise in MEASURED production, simply because the price of oranges rose.

It gets worse. What is our measure of 'price'? When we looked JUST at apples, it was the price of apples. When we look at twenty fruits... it's not as clear what price we should use. Economists often use a price index, such as the CPI, which tracks the price of a 'typical consumer's' spending. Suppose we use a 'fruit price index', which tracks the price of a particular combination of fruits, in a way similar to how a stock index tracks the value of various stocks.

Our supply and demand diagram is now a lot less useful than it was, because we're plotting not price against quantity, but the cost of a particular subset of production (whatever combination of fruits we choose) against the cost of total production.

It's not easty to untangle the various effects from each other.

This is why 'aggregate' or 'added-up' supply and demand is pretty much a fairytale. We can THINK about it - we can pretend that we can actually measure output. Doing so gives us a lot of interesting and useful results, such as allowing us to understand how an oil shock can cause a recession AND inflation, or why price controls seldom work. However, any attempt to take the story literally and apply real-world data to it is going to run into problems.

To be fair, professional economists often use advanced statistical techniques (such as regressions) to separate out the various effects, with a great deal of success.

This does not help much with a basic problem - that the textbook theory of aggregate supply and demand has been built on an awkward mathematical foundation. Geometry and simple algebra aren't the right tools for the task.

No comments: