Show Me The Model

Three opinion leaders in the blogosphere have laid out how they think about the macroeconomy. They talk about “models” but unfortunately don’t deliver. Instead, they provide lists of beliefs or facts to be explained. Economics is a science precisely because it has progressed beyond such lists. Economists build models—consistent, well-structured and clearly specified (and thus, mathematically formulated) stories.

But here are the lists: Scott Sumner’s “Musical Chairs model” (blog):

In the short run, employment fluctuations are driven by variations in the NGDP/Wage ratio.

Monetary policy drives NGDP, by influencing the supply and demand for base money.

Nominal wages are sticky in the short run, and hence NGDP shocks cause variations in employment in the same direction.

In the long run, wages are flexible and adjust to changes in NGDP. Unemployment returns to the natural rate (currently about 5% in the US.)

Tyler Cowen’s “model” (blog):

In world history, 99% of all business cycles are real business cycles.  No criticism of RBC can change this fact.  Furthermore the propagation mechanism for a “Keynesian business cycle” (arguably a misleading phrase) also relies on RBC theory.

In the more recent segment of world history, a lot of cycles have been caused by negative nominal shocks.  I consider the Christina and David Romer “shock identification” paper (pdf, and note the name order) to be one of the very best pieces of research in all of macroeconomics.  Sometimes central banks tighten when they shouldn’t, and this leads to a recession, due mainly to nominal wage stickiness.

Workers are laid off because employers are often (not always) afraid to cut their nominal wages, for fear of busting workplace morale, or in Europe often for legal and union-related reasons.

Overall I favor a nominal gdp rule for monetary policy.  But most of its gains would come in a few key historical episodes, such as 1929-1932, or 2008-2009.  In most periods I don’t think we know what the correct monetary policy should be, nor do we know that it matters.  Still, that uncertainty does not militate against an ngdp rule.

Once workers are unemployed, nominal wage stickiness is no longer the main reason why they stay unemployed.  In fact nominal wage stickiness is largely taken out of the equation because there is no preexisting nominal wage contract for these workers.  There may, however, be some residual stickiness due to irrational reservation wages, also known as voluntary unemployment due to stupidity.  (You will find a different perspective in Scott’s musical chairs model, which I may cover more soon.)

Monetary stimulus to be effective needs to be applied very early in the job destruction process of a recession.  It is much harder to put the pieces back together again, so urgency is of the essence.

The successful reemployment of workers depends upon a matching problem, a’la Pissarides, Mortensen, and others.  Yet this matching problem is poorly understood, and it can involve a mix of nominal and real imperfections.  Sometimes it is solved more quickly than expected, such as in the recent UK experience, and other times more slowly than expected, as in current Spain.  Most of the claims you will read about this reemployment of workers are wrong, enslaved to ideology or dogmatism, or at the very least unjustified.  Hardly anyone wants to admit this.

Really bad recessions involve deficient aggregate demand, negative shocks to intermediation, some chronic supply-side problems, negative wealth effects, and increases in the risk premium, all together.  It is hard to find a quick fix.  Furthermore models where AS and AD curves are independent and separable are often misleading, despite their analytic convenience.

Given that weak AD is only one of the problems in a bad downturn, and that confidence, risk, and supply side problems matter too, the best question to ask about fiscal policy is how well the money is being spent.  The “jack up AD no matter” approach is, in the final political equilibrium, not doing good fiscal policy any favors.

You should neither rule out nor overstate the relevance of Hayek and Minsky.  Their views have much in common, despite the difference in ideological mood affiliation and who — government or the market — gets blamed for the downturn.  For really bad recessions, usually both institutions are complicit to say the least.

All propositions about real interest rates are wrong.

The Economist’s Free Exchange response to Cowen’s model (blog):

Supply-side policy is hard. Why is America the richest large economy in the world? Well, because output per person has grown at about 2% per year, on average, for a very long time. How did it manage that? I have a long list of policy choices and characteristics and historical accidents that I believe contributed, but I would find it very difficult to say which of those factors were most important. If someone gave me free reign over the German economy and asked me to raise its output per person to American levels, I know the sorts of things I would do, but I have a low level of confidence that I could succeed, or even close much of the gap, within a generation.

That doesn’t mean that supply-side policy should be ignored. Supply-side reforms (of the sort this newspaper tends to favour) are politically difficult to achieve, but many of them are probably at least somewhat useful and should be undertaken whenever the political environment is amenable (though with very modest expectations regarding detectable effects on growth).

With supply-side policy, the precision of a policy action is not the problem; accuracy is. With demand-side policy, it is the opposite: it is pretty easy to meet broad policy goals, so long as you’re not too concerned about hitting them square on the nose.

We know what an economy with way too much demand looks like. It has high and accelerating inflation.

We know what an economy with way too little demand looks like. It has high unemployment and deflation.

Within those two extremes, it can be tricky to identify exactly where an economy stands: how close or far away from potential output it is.

Both too much and too little demand are economically costly, but history suggests that too little demand is far more economically costly and politically risky than too much demand. So policy should err on the side of too much demand rather than too little.

The determined use of monetary policy is almost always going to be sufficient to generate the right sort of “too much demand”. But an independent central bank might not always be able to muster the appropriate determination. In some cases a central bank may flounder until a clear political consensus emerges supporting the determined use of monetary policy.

It is generally unwise for countries to sacrifice monetary-policy autonomy, either by adopting a constraining exchange-rate regime or by introducing an excessive level of capital-account openness.

In countries with autonomous monetary policy, which are stuck at the zero lower bound on interest rates, fiscal policy is almost by definition too tight, and it is probably quite difficult to conduct fiscal stimulus in a way that generates long-run economic costs. That is because the long-run supply-side and fiscal benefits of getting off the ZLB are probably pretty large.

Fiscal policy is subject to political constraints, and it may be easier to introduce a large stimulus in emergency situations if the pre-emergency public-debt burden is low. That suggests that prudence in normal times is a good idea (though do remember point number 10).

Don’t subsidise debt.

The level of financial- and banking-sector liberalisation at which it can be demonstrated persuasively that further liberalisation will generate net benefits is probably not that high.