Sunday, May 24, 2009

Never mind the gap

In case you were looking forward to my judicious insights on how to best get off the subway, you're about to be disappointed: It’s the output gap I’m gonna talk about—the difference between aggregate demand and an economy’s potential supply.

Some inflation hawks out there have taken the task of trashing the usefulness of the output gap as a guide for monetary policy decisions, in order to extrapolate scenaria of high inflation as a result of the Fed’s current money-printing enterprise.

But the argument is misplaced… Not because the possibility of high inflation is far too remote (I’ll come back to that in a minute); nor because the criticism re. the output gap is unfair (it isn’t). It’s just that the inadequacies of the output gap framework fail to provide a reason for casting Cassandra-isms on the inflation outlook.

But let me go a step back.. on the output gap... The concept has been used for years now by economists—including the Fed—as an important component of their modeling toolkit to forecast inflation and guide their monetary policy decisions.

The idea goes like this: When aggregate demand is higher than potential supply, inflation should be expected to go up—therefore the Fed should raise interest rates to prevent an undue acceleration of prices. The opposite holds when aggregate demand is lower than potential supply, which is believed to be the situation today.

The problem with this framework is that nobody has ever seen an output gap—the “potential output” of an economy is unobservable and has to be modeled for. As such, it is subject to a lot of uncertainty, given different model specifications and/or the frequent data revisions that occur as our information about the economy improves over time.

That may not be as big a deal for someone who wants to conduct a historical analysis. But it’s a serious issue when it comes to policy-making in real time, as has been frequently pointed out by renowned monetary economist Athanasios Orphanides (formerly at the Fed). Orphanides has found that the often inevitable mismeasurements of the output gap in real time could lead to inferior policy responses, even compared to policy responses that are guided by a simpler framework, e.g. one that uses past output growth instead.

That’s all solid and good. But is there any connection with the inflation omens of our times?

One supposed connection is the idea that we may be mismeasuring the output gap right now, by a significant amount. In other words, the Fed may be under the illusion that aggregate demand has fallen way short of potential output and, as a result, keep monetary policy too loose for too long.

The reason for the illusion could be that the economy’s potential supply has fallen beyond imagination: For example, the "equilibrium" unemployment may have risen as laid off workers need time to retrain in order to find new jobs; meanwhile, firms may shy away from putting new capital to work as they seek to repair their balance sheets a retool for a tougher financing environment.

While plausible, the drop in potential supply would have to be pretty severe to justify a high inflationary scenario based on the output-gap framework. It's like saying that the "natural" unemployment rate has gone up to 9-10% (ie near or above current levels). Not impossible, only that since nobody has perfect information about the current output gap, the conjecture is as inadequate for guiding the Fed's policy decisions as the argument of the opposing camp.

Then you have to think of the alternatives. I mean, fine, academic research has shown that the Fed’s output gap framework is not the holy grail of real-time inflation forecasting: Alternative models of future inflation could produce superior policy responses, e.g. models of inflation based on past inflation alone, or a combination of past inflation with past output growth.

But what would these models tell us today? With both inflation and output growth having gone downhill recently, even these “superior” models would still fall short of predicting a scary inflationary scenario.

So hawks need additional ammunition. And they find it in the Fed’s expansion of base money—the result of its credit/quantitative easing operations. But now things become flakier.

As I’ve argued in past posts, the usual monetarist argument that a large-scale expansion of money supply leads to inflation appears naïve at present, as it assumes a stable demand for money. But this has clearly not been the case, as people, banks, everyone rushed away from risky assets towards liquid alternatives (read “money”!).

So should we sleep in peace then?

Well no... I didn't say there has been no policy blunder. In my view there has been, though not in the choice of an inflation model, but in the effort to manage inflation expectations.

First by communication: The Fed has yet to be crystal clear about the hierarchy of its objectives in the case where inflation, growth or credit conditions start drifting in conflicting directions (e.g. inflation heading up while growth remains subpar).

Importantly (and I can’t say this enough), the Fed screwed up in my view by deciding to buy Treasuries. Apart from the very shaky empirical evidence on the effectiveness of Treasury-buying in bringing down long-term rates, the policy has spread confusion about the Fed’s intentions, fuelling suspicions about debt monetization at a time when the incentives to do so are pretty strong. (See here for more on this issue).

Indeed, one could conceive a spiraling scenario whereby large Treasury purchases end up undermining confidence in the Fed’s credibility and the US dollar, prompting a sell-off in the currency, leading to inflation, undermining the Fed’s credibility further and so on. We kind of saw tiny hints of this last week with the dollar sell-off gaining momentum.

I don’t want to sound like a Cassandra myself. Especially because economists have a very poor understanding of how agents form their inflation expectations and, importantly, how expectations react to different sets of policies.

So I could go on speculating that the Fed’s Treasury buying will fuel inflation via a collapse in the dollar, while you argue for the opposite by pointing to the various indicators of inflation expectations (TIPS, surveys to economic forecasters, wages, commodity prices, etc), which (still) remain fairly well-contained.

Perhaps if there is any conclusion to be drawn out of this is an admission that we are all pretty much clueless.

3 comments:

Paul W said...

What do you think about the Fed buying mortgages? At these prices, that is a lot of extension risk being added to the balance sheet. Combine this with a housing market that will be weak for a long time, I don't see how the fed could sell these securities off to reduce money supply. Am I wrong? Is it even an issue?

Chevelle said...

I actually objected to the MBS purchases (here http://modelsagents.blogspot.com/2009/01/messing-around-with-credit.html) but the argument was not one based on effectiveness but one based on credit allocation.

My point is that the Fed should not be in the business of allocating credit to a specific sector --like housing. And what a sector, by the way! One where inventories are still near record highs and will require further price reductions to clear, for good or for worse.
Now, if the government/society decides housing is a sector that needs support (more than it currently receives!), it's the government who should take up the bill--e.g. through transfers to homeowners--, and not the Fed.

With regard to the Fed's ability to reverse the monetary impact of these purchases:
Actually, the Fed is developing tools to allow it to withdraw liquidity without having to sell these securities: e.g. by paying interest on bank reserves or by reverse-repoing these MBS. It will be a matter of willingness to do so, not ability.

Paul W said...

Thanks for the clarification.

Your blog is a great antidote to the MSM.