This post is a sort of response to a recent post by Tyler Cowen, which was skeptical of whether the “nominal AD” shock model could provide an adequate explanation of our current crisis. To explore this issue I would like to start by asking what a pure nominal shock would look like.
Let’s suppose that the economy is in equilibrium at the natural rate of unemployment, and suddenly the Fed tightens monetary policy enough to reduce NGDP growth about 8% below trend. Also assume wages are sticky. What would happen? The answer depends on the structure of the economy. If all output is one all-purpose good, output of that good might fall modestly, and prices might decline as well. The total decline in real output and prices would be 8%.
Does this look like our current situation? Obviously not. Some industries have suffered severe declines while other have hardly been touched. Does this mean that we need a sectoral explanation of the current recession? No, rather we need need to account for the existence of capital goods, consumer durables, and services. And we need to recall that the permanent income hypothesis predicts that during a recession there will be a sharp decline in capital investment and consumer durables, whereas the output of food, clothing, and many services will decline by a relatively small amount. If you had a sharp but temporary decline in your income, would you cut back more sharply on consumption of food, haircuts, or purchases of new cars? The answer is obvious. And that is why a pure nominal shock will produce massive sectoral shifts, even if in the absence of a nominal shock the economy needed no sectoral adjustments at all.
Now let’s return to the original assumption that the nominal shock came out of the blue. How likely does that seem? In my view monetary policy shocks are not arbitrarily produced by central bankers, but rather reflect a flawed monetary policy process that puts too much emphasis on interest rates as an “instrument” of policy (actually short run target), and inflation stabilization as the goal. As you know, I favor using NGDP futures as the “instrument,” and expected NGDP as the goal variable.
The monetary policy process suffers from two flaws. If there is an autonomous decline in the Wicksellian equilibrium interest rate (the rate consistent with macro equilibrium) and the Fed keeps its target rate constant (as it did throughout much of 2008) then NGDP growth expectations may plummet. This can cause a recession.
A second problem occurs if there is a severe oil shock during a period when NGDP growth falls below trend. If the central banks focuses on headline inflation instead of NGDP, they may tighten policy, or refrain from cutting the target rate when needed. This also occurred in mid-2008.
When both of these errors occur at the same time you can have a severe recession. Just to be clear, here I am only trying to explain the severe recession that began in mid-2008. So let’s go back and look at the early stages of the recession, the first half of 2008. There were three problems:
1. A sectoral shift out of housing and related activities that had been occurring since mid-2006. This sort of sectoral shift did not and typically does not cause a recession.
2. A severe energy shock began to develop in 2007 and worsened in early 2008. This type of shock will hurt energy intensive industries, but will not generally cause a recession. Most economists were not predicting any sort of severe recession as of mid-2008, despite the fact that they already knew of the sub-prime fiasco and the $147 oil.
3. In the first half of 2008 there was a mild AD shock, as NGDP growth slowed from its usual 5% down to below 3%. Even the combination of all three shocks was not able to produce a major rise in unemployment, nor did most economists expect a major rise in the future, despite the fact that they knew about all three problems. Let’s just stand back and marvel for a moment at the astounding resilience of the US economy. It takes three hammer blows and it is still standing.
But then in just a few months it all started to unravel. The culprit? Something as seemingly innocuous as a decline in NGDP caused by the sort of flaws in our monetary system described earlier. In other words these massive real shocks did relatively little damage to our economy (or our financial system for that matter) but as soon as M*V started falling everything fell apart. What can explain this? Any principles of macro text can explain what happens when NGDP falls. I prefer texts like Cowen and Tabarrok, which define AD growth as a given increase in NGDP. Suppose NGDP growth falls 8% below trend. Because wages are sticky, relatively few workers will get hourly wage increases 8% below trend. Instead, three things will happen: Profits will fall sharply, bonuses will fall sharply, and hours worked will fall sharply.
What makes macro so endlessly fascinating is that the none of this will appear to have been caused by the factors I describe. Even in an economy with only one good, no one can “see” a fall in M*V, so it will appear that some other mysterious factor will have caused the recession. But in an economy with heterogenoeous goods and sticky wages my explanation will seem even more counterintuitive. Some sectors (especially capital goods) will decline sharply. Many prices will be relatively sticky, but asset prices will crash. It will look like the wealth destroyed in the asset price crash and the sectoral shocks are the two prime causes of the recession. But according to mainstream macro theory both are actually symptoms of falling nominal spending/income.
How could my hypothesis be disproved? If the Fed pegs the price of NGDP futures and we continue to see events similar to what we observed in late 2008, then I’ll throw in the towel.
Where does policy go from here? There is one additional problem. Not only do real shocks such as financial turmoil and/or energy shocks often trigger bad monetary policy, but the resulting recession often leads to bad supply-side responses. By far the worst example was the NIRA in July 1933. But even in this recession we have a dramatic lengthening of the duration of unemployment benefits, from 26 weeks to as much as 73 weeks. (I also recall reading that Obama planned to reverse Clinton’s welfare reform, but I haven’t followed the issue closely.) In a normal recession workers who have exhausted unemployment benefits are more willing to accept any job at any wage rate. This increase in labor supply puts downward pressure on wage rates. In the current recession wages have been stickier than in 1921, and thus the unemployment will be more prolonged. FYI, 1921 was probably the purest nominal shock in American history. It is worth studying intensively if you want to see what the AS/AD model purports to describe.
[As an aside, many people get confused when some workers accept big wage cuts and still get laid off. The problem is that if you are in a highly cyclical industry, it is not enough that you accept wage cuts. Instead, to prevent high unemployment in cyclical industries it is necessary for everyone in the economy to accept big wage cuts, even government workers. I know people working for high tech firms who haven't yet had any wage cuts in this recession. We are far from the sort of wage flexibility required to make nominal shocks neutral, despite all the reports of wage cuts you might have read about. And wage cuts mean cuts in hourly wage rates, not annual income.]
BTW, I hate the term “aggregate demand” as I have no idea what it means, and when I read other economists I immediately realize they mean something much different than I do. I prefer to define AD as NGDP, an approach adopted in the new macro text by Cowen and Tabarrok. But Cowen also recognizes that economists often mean something very different by ‘aggregate demand’ and thus discusses ”real AD” shocks in his recent post:
If someone wants to insist that “this is really an AD shock, not a sectoral shift,” I’m not so keen on fighting to keep one term over the other. I would insist, however, on an issue of substance, namely that not all AD shocks are alike. If we are going to switch terminology, it could be said that this is a real AD shock and not just a nominal AD shock. (Though there have been nominal AD shocks too.) A nominal AD shock can be offset more easily by goosing up some mix of M and V and restoring the previous level of nominal demand.
Thus Tyler Cowen correctly observes that real AD shocks are essentially sectoral shocks. Thus a fall in wealth may make people shift their consumption away from some goods and towards others. It doesn’t make people want to work less, but I think both Tyler and I agree that during a transition period it causes frictional or structural unemployment that cannot easily be papered over by printing money. Where we may disagree a bit is that I think the fall in wealth due to the sub-prime fiasco was not severe enough to cause even a small a recession, and the much bigger fall in wealth after mid-2008 was caused by sharply falling NGDP expectations—i.e. by a nominal shock.
Would reversing this nominal shock be able to reverse all the negative effects I described? I can’t say for sure. We have structural problems like extended unemployment benefits that we didn’t have in mid-2008. But I still think it would do a lot of good. More NGDP could help because asset prices are highly volatile and forward-looking, while nominal wage rates are very sticky. A strong increase in 2011 NGDP expectations would sharply raise asset prices, but hardly budge current nominal wages (given 10% unemployment.) This would boost employment and output.
Every day that goes by my preferred policy response becomes slightly less effective in reducing the problems we currently face. But oddly a decision to “be irresponsible,” i.e. to have “excessive” NGDP growth in order to catch up to the 5% trend-line, would make it far less likely that NGDP growth expectations would collapse in the next crisis. This time they collapsed because the markets guessed the Fed would not try to correct its error and return NGDP to the old trend line. And as each day brings new stories of a Fed itching to tighten policy even as the fiscal authorities are contemplating new stimulus (a policy mix showing our government is approaching “banana republic” level of incompetence) it becomes more and more evident that the bearish speculators of late 2008 were right—NGDP is going to stay on a new and permanently lower growth track. There will be no nominal recovery. And with the lengthened unemployment benefits the real recovery will take longer than normal.
PS. This was a frustrating decade at times, and ended with me missing my flight home on the last day. But at least humanity had its best decade ever, at least if you believe economic development and peace makes people happier. So I guess I shouldn’t grumble.
My new year’s resolution? How about full RSS feeds? I have only a vague idea of what they are, but I’m told I should want one. And that means I want one.
Happy New Year!