What is Demand, Anyway?
Matthew Martin 9/15/2012 10:29:00 PM
But what do we mean by demand-side? In an undergrad economics course, we are taught to thing of "demand curves" and "supply curves," but these may actually obscure more than they reveal, so I will propose an alternative way of thinking about it:
Let us just define, for the purposes of the rest of this post, the term "demand" to mean the quantity of goods and services purchased in the economy. There are three things that cause demand to change:
- a change in relative price (at the macro level, the pertinent relative price is the amount of future goods and services we can buy with a unit of current goods and services),
- a change in the consumers' budget sets, or
- a change in the consumers' preferences.
The budget set tells all the different combinations of goods and services that the economy is capable of supplying, but not which one specific combination of goods and services that consumers will want. This is where consumers' preferences come in--consumers will choose the element of the budget set that they most prefer. (Strictly speaking, there is nothing in economics that says that the element consumers will choose need correspond to the one that they like the most; this is just a philosophical belief that most economist hold, but is in no way necessary for the validity of our models). In short, any thing that affects consumer preferences is a "demand-side" effect.
Now suppose that Krugman and the economists are right that this recession is "demand-side," meaning that it is due to a shock in consumer preferences. This means that demand--as defined specifically for this post--has fallen. Because we know that demand has fallen, we know that we have to now be on the interior of the budget set, since nothing else has changed yet, and as a result, expanding the budget set (enacting supply-side policies) won't affect demand. That leaves only option (1)--we have to lower the relative price to increase demand. The relative price here is the amount of future goods and services we can buy per unit of today's goods and services. One way to measure this is by calculating the interest rate. So we need the interest rate to fall.
Or do we? As a thought experiment, consider the case where the nominal interest rate is at zero, and we experience a negative shock to preferences. We need the interest rate to fall, but it is already at zero. Of course, what matters is not the interest rate per se, but the amount of future goods and services we can buy per unit of today's goods and services, so an alternative to decreasing the interest rate is to simply raise future prices, or else reduce current prices. Since future prices are constrained by the money supply (or alternatively think of this as future prices are constrained by expectations), this effectively means that without government intervention, current prices need to fall. In a sense, this is what "should" happen in a free market. If it does, then there is no need for any government intervention--the mechanism in (1) will offset any effects of mechanism (3).
My point in explaining this is to point out that (3) by itself says nothing about why we are in a depression, or why we need government policy to intervene, even given that there is a zero lower bound on interest rates. In fact, as far as flexible-price economic models are concerned, there is no need to even have any interest rate at all--we can achieve analytically the same relative prices by adjusting either the interest rate or the price level--they're redundant. But in reality, prices are sticky, meaning that they can't immediately fall in the face of a negative shock to demand. This means that in reality, price levels and interest rates are not redundant, and we need to have the interest rate able to fluctuate in the short run to take care of the fact that prices can't. And since we rely primarily on interest rates to respond to short run changes in preferences, it matters a lot that we are at the zero lower bound. To sum up, you need both sticky prices and a negative preference shock to cause a depression like the one we are in.
Some people have mistakenly argued that because the depression arises because of the combination of both negative preference shocks and sticky prices, we can end the depression by simply lowering current prices. Their confusion is understandable, because this does appear to be an implication of the model. The error arises, however, because the government doesn't actually set any of the prices in the economy. Instead these commentators are inevitably arguing for policies such as tax cuts that would actually put even more downward pressure on prices, which actually increases the amount by which prices would need to fall, thereby reducing demand even further. The same result holds for any policy aimed at lowering prices, such as union-busting or other so-called labor market reforms. Even if these policies are beneficial in the long run, they are inevitably contractionary at the zero-lower-bound.
So the real socialists are the ones who argue in favor of lowering prices rather than increasing government spending. They are socialists because for their idea to work, the government would literally have to adjudicate every single contract in the economy and set all the prices directly, financing all the adjustment costs with taxpayer funds. If firms are free to set prices themselves, then we inevitably have sticky prices and the potential for demand-side depressions.