Scott Sumner  

What went wrong in 2008?

A Proposed Bet for Bob Murphy... Why Should We Worry About Stoc...

Over at TheMoneyIllusion I discuss a new paper by James Bullard, et al, on the case for NGDP targeting. The Wall Street Journal reported some additional comments by Bullard:

He said his model is consistent with the Fed maintaining its 2% inflation target, but would simply mean the central bank would work more aggressively to counter any undershooting or overshooting of that objective. For instance, a nominal GDP target would have likely prescribed even more aggressive Fed monetary easing during the crisis-ridden 2007-2009 period, Mr. Bullard said.

"If you took this paper seriously we should have done even more in 2007-2009," he said. "Something like nominal GDP targeting, if it's appropriately formulated, does look like optimal policy."

In their model, the central bank would aim to raise the price level when there is a negative economic shock, such as we saw in 2007-09. Here I'd like to consider why the Fed did not do this optimal policy. I'll argue they failed because they were focusing on stabilizing inflation, not NGDP growth.

Paul Krugman has suggested that pre-1985 recessions were usually caused by the Fed raising interest rates to fight inflation, but more recent downturns were "balance sheet recessions", and were not preceded by a sharp upward spike in interest rates. There's some truth to the claim that recent recessions differ from those occurring before 1985, but I'll try to show that fighting inflation played a larger role in policy mistakes during 2008 than you might assume.

I could focus on NGDP and inflation, which behaved very differently in 2008. NGDP slowed throughout 2008, whereas inflation rose sharply until midyear, and then fall sharply. But the NGDP data is reported with long lags, and has been revised downward substantially since 2008. So policymakers in real time may not have perceived a big NGDP problem. Instead I'll focus on the unemployment rate, which is reported in a timely fashion, and generally isn't revised very significantly. Here's Michael Darda on the unemployment rate and recessions:

. . .anytime the U.S. unemployment rate has risen 0.5 percentage points or more from year-earlier levels (after a period of annual declines), the U.S. has either been in recession or on the cusp of one with no exceptions to this rule in post-war history.
I've made a similar point in previous posts, but I like Darda's formulation better. Keep in mind that a 0.5% increase is really small, and that it's hard to find non-trivial indicators that correlate 100% with post-war recessions. (Note that this is different from predicting recessions before they occur. For that goal economists tend to rely on yield spreads, which are less accurate but more timely. Here I'm more interested in identifying recessions that have already begun, or are about to do so.)

Below I've provided the 12-month change in the unemployment rate, and also the 12-month change in the price level (i.e. PCE inflation) for key dates in 2007 and 2008:

Date ** Change in U ** PCE Inflation

Oct. 2007 ** +0.3% **** 3.1%
Nov. 2007 ** +0.2% **** 3.5%
Dec. 2007 ** +0.6% **** 3.4%
Jan. 2008 ** +0.4% **** 3.1%
Feb. 2008 ** +0.4% **** 3.2%
Mar. 2008 ** +0.7% **** 3.2%
Apr. 2008 ** +0.5% **** 3.2%
May 2008 ** +1.0% **** 3.4%
Jun. 2008 ** +1.0% **** 3.9%
Jul. 2008 ** +1.1% **** 4.2%
Aug. 2008 ** +1.5%

Here I have included all of the data available to FOMC members as of the September 16, 2008 meeting. It's striking that the unemployment rate for December 2007 was the first to provide a recession signal. That's the date the recession actually began, but even as late as mid-2008 the Fed did not expect a recession.

Of course just because a signal has always been correct in the past, doesn't mean it will work in the future. And the 12-month change in unemployment did moderate to 0.4% in January and February of 2008. Then another even stronger signal in March (plus 0.7%), followed by a fallback in April (to plus 0.5%). We seemed to be skirting on the edge of recession. After May, however, the economy went into a deep tailspin, with the unemployment rate showing a plus 1.0% in May, rising to a huge plus 1.5% in August.
How did the Fed react to these very scary unemployment numbers? By doing nothing, or by tightening, depending on your perspective.

The Fed held its fed funds target stable (at 2.0%) from late April to October 2008. This despite extremely serious deterioration in the labor market. But why? In the statement issued after the September 16, 2008 meeting, which occurred two days after Lehman failed, the Fed gave us their reasoning. They indicated that they saw equal risk of recession and high inflation. Given that they perceived these risks as balanced, they decided not to change policy.

In my view, policy was being continually tightened throughout this period. If we had had a NGDP futures market, it would have indicated falling NGDP expectations. TIPS spreads were declining, and reached 1.23% for the 5-year spread by the time of their September meeting.

Today we know that the real problem was deflation, which had set in by early 2009. But the Fed was worried about high inflation, and this led them to unintentionally tighten monetary policy. Even though the policy rate was stable at 2.0%, the Wicksellian natural rate was falling fast between April and October. Hence the effective stance of monetary policy was getting progressively tighter.

The new paper by James Bullard (who is president of the St Louis Fed) and three other researchers, tells us what went wrong in 2008. The Fed needed to allow a higher price level, to offset the negative shock from the housing/banking crisis. That would have required them to ease monetary policy. But the inflation numbers were very scary. Not only were they higher than the Fed's 2% target, they were increasing up until July 2008, the last data point available by the September meeting. Little did they know that inflation had already fallen to 4.0% in August, and would reach less than 0.4% in December. By March of 2009 the 12-month inflation rate was negative---we were in deflation.

The markets figured out this problem before the Fed. If the Fed had adopted NGDP targeting, they would have ignored the inflation numbers and focused on NGDP growth. Even so, because of data lags they might have been a bit behind the curve. But the markets would have expected a catch-up from any near-term NGDP undershoot, and thus NGDP growth expectations would have stayed strong enough to prevent the economy from hitting the zero bound.

The problem was not so much that the Fed didn't cut rates quickly enough in 2008 (although that was a problem) but rather that they didn't have a monetary regime in place that would have prevented the Wicksellian natural interest rate from falling to zero in the first place.

The Bullard, et al, paper represents a huge step in the right direction, toward a better understanding of how monetary policy failed in 2008.

Comments and Sharing


COMMENTS (16 to date)
marcus nunes writes:

Scott, my only concern is that the idea was put forth by Bullard, who arguably is the most fickle of FOMC members:

John Hawkins writes:

Scott, How would a NGDPLT perform during a supply shock, as opposed to a demand shock? How would it perform relative to 1) an inflation target, and 2) a productivity norm?

M.R. writes:

Did other central banks (ECB, BOE, etc.) make essentially the same mistake as the Fed in 2008 (tightening)?

fralupo writes:

An interesting case for NGDP targeting. What did NGDP change look like for those same periods?

Michael Byrnes writes:


If you had gotten your Hypermind market in place in, say, mid-2007, what do you think it would have done over the next 2 years?

Does Bullard say anything about level targeting in the paper?

Scott Sumner writes:

Marcus, Fickle? Or open-minded?

John, It is during supply-shocks that NGDP targeting really outperforms inflation targeting.

MR. The ECB did even worse, raising interest rates in July 2008. The BOJ's policy was also too tight, but in that case it had been for 15 years.

Fralupo, NGDP growth slowed sharply in the first half of 2008, and then NGDP fell in the second half.

Michael, I don't recall him saying anything about level targeting. I believe the Hypermind market would have shown the folly of Fed policy by September, at the latest.

fralupo writes:

It seems like the principal problem with an NGDP target, then, is that it gets updated so infrequently. Does anyone publish monthly NGDP numbers on Earth?

Brendan Riske writes:

Would it be correct to say that in 2008 the US had an number of structural factors which caused NGDP growth to decline, and that Fed policy has been unable to allow the economy to heal. Even with the policy you are suggesting if it was the economic structure that lead to a NGDP decline, then Fed policy could not do much to fix it in the long term anyway. Can I say that is a fair interpretation of part of your argument?

What is the explanation of the mechanism by which "easier money" increases real GDP?

It seems to me (ISTM) that tight money would show as a lack of bank reserves to make loans against reasonable collateral, preventing creditable companies from expanding their operations. Equivalently, it would show as an increase in interest rates demanded by banks as they rationed available capital.

Why is it hard to determine if money is tight?

If banks are not lending at very low interest rates, and they have reserves granted by the Fed, then who is lending? At worst, why don't banks raise their price, the interest rate, at which they will make loans to companies?

Alternately, maybe companies won't borrow and expand even at very low interest on loans. Or, rising stock prices are funding equity investments, leaving out the banks.

If companies don't want to borrow, even at low interest, then ISTM that money tightness is not the problem. The problem would be regulatory friction or high risk perceived from other causes not related to monetary policy. Nothing monetary is going to fix that by much.

How does increasing the level of prices help? Is that supposed to force companies into borrowing, to hedge their cash balances and not lose cash in an expected inflation?

If companies and a free market are the optimal way to allocate capital and discover useful projects, then why should forcing, I mean encouraging, them to borrow be an efficient way to run an economy?

What are the costs to cash savers on the sidelines, especially retirees?

Don Geddis writes:

@Brendan Riske & Andrew_M_Garland: You both misunderstand. Yes, the US had some real / structural factors leading to low real growth. But those were relatively minor, and would only have caused a minor recession (at most). The major economic damage came from the unnecessary negative shock to aggregate demand. Which was mostly in the form of plunging velocity (= increasing demand for money).

It's not about banks, or borrowing money. It's about total spending in the economy. If spending drops, then each company's revenue drops. When costs are sticky (wages, debts), but revenue drops suddenly, then the only option is to reduce the workforce and reduce output.

The goal of monetary policy is not to encourage more investment. It's to maintain spending along trend expectations, so that we don't have unforced idle workers [higher unemployment], and thus lower total national wealth created.

It's important to realize that most investments are not "good" or "bad" in some absolute, objective sense. Whether an opportunity is a "good investment" typically depends on monetary policy, because you're trading off a nominal cost now, in exchange for a nominal future return. And monetary policy determines the real value of that future nominal return.

marcus nunes writes:

In the US Macroeconomic Advisers publishes monthly NGDP & RGDP:

Brendan Riske writes:

@Don gettis

So you believe that the end of the housing bubble would have caused a "minor recession" at most? I don't think thats accurate. If we had et the banks blow up from their risky bets (which would have been good) the effects would be systemic and potent. The entire US economy needs to make major changes to grow. Way too much regulation, way to much crony capitalism.

I love discussions of monetary policy. Its all a sham. Say what you want about maintaining aggregate demand, or keeping the price level stable, or keeping inflation constant. But in practice, they are the lender of last resort and bailout kings. They can't create financial stability, definitely not with the methods they have been using. You can not control a complex system like an economy with a few policy levers. The more they try to control, the more the forces of creative destruction build up and the greater the chaos from their inevitable release. No central planners have a chance at managing a national and certainly not a global economy. Its too complex. The Fed has thought itself very smart for a number of years, I find their beliefs misguided.

Scott Sumner writes:

Fralupo, Macroeconomic Advisers published monthly estimates. I believe Canada does as well. The US government could do so if it wished to, but of course they'd be revised as new data came in.

Brendan, Absolutely not. My argument is that a tight money policy at the Fed caused NGDP to decline in 2008. They caused the recession.

Andrew, I don't believe monetary policy should have any relationship to bank lending. Money and credit are completely different things.

Monetary policy affects NGDP via the hot potato effect, and NGDP affects RGDP in the short run due to sticky wages and prices. It has NOTHING to do with lending.

Brendan, Banks got into some difficulty in 2007 and early 2008 due the the housing crash, but nowhere near enough to cause a recession---the huge losses to banking in late 2008 were caused by the drop in NGDP, i.e. tight money. Having said that, you are right that we need to reform our banking system; pity that Dodd-Frank did not do so.

Scott Sumner writes:

Brendan, You said:

"The Fed has thought itself very smart for a number of years, I find their beliefs misguided."

We certainly agree on that point! The Fed should not try to control the economy, they should merely keep the monetary system stable.

Brendan Riske writes:

Well I am glad we agree on that! I do enjoy reading your theories on NGDP and the possibility of targeting it. Certainly much better than watching inflation. I do think there is something there as far as explanation. But I do get cautious using the numbers to explain everything when there are so many structural issues that need to be address. Tax policy and regulation needs work. The labor market is going through some stress as the economy transitions towards more flexible employment. And technology and globalization alot pressure the labor market. Demographics are changing and populations are aging. I also see plenty of fraud and market rigging in finance and politics. Big issues with pension and retirement funding. The US healthcare system is still to expensive relative to the world. The US higher ed system has too much debt. Crony capitalism is a big deal in making the developed economies inefficient. and on the whole there is a switch to a more sustainable growth with less environmental impacts. In the face of structural issues on almost every front which don't seem to be getting resolve, I worry about the economy.

Mr. Sumner, I found your blog post explaining the Hot Potato Effect:
09/01/2013: The “hot potato effect” explained

So, here is rough summary of what I understand from it.

If some entity provides a large amount of new money, then people will want to use their cash (sell it) ahead of seeing it lose value. But, thay can't do this in aggregate, because there has to be a buyer for each seller. So, all cash will lose real value (prices will go up) until the new amount of money falls into a new equilibrium with the items it can buy.

=== ===
[Sumner, edited] The extra gold is a “hot potato” that people try to get rid of by selling it. This works at the individual level, but it doesn’t work in aggregate. Now someone else has the extra gold. (That’s why attempts to understand money at the level of the representative consumer fail.) The only way for society as a whole to get rid of the extra gold is by driving down the price of gold until people want to hold the new and larger quantity.
=== ===

This extra money can be real, or possibly just the credible threat that the central bank will supply that extra money soon. That is, a credible threat that money will lose a good part of its real value as prices go up in anticipation or as a response to the increased supply.

=== ===
[Sumner, edited] If prices are sticky then other things will change to bring about a short run equilibrium in the gold market before the price level has had time to fully adjust. One of those other things might be a change in interest rates, in asset prices, and/or in real output.
=== ===

In extreme summary, this says: Impose inflation in prices by supplying much more money or threatening to supply it. The resulting inflation will require a new long-run equilibrium, and a short run adjustment which MAY affect interest rates, asset prices, and/or real output.

I hope that my summary is roughly correct, but I would like an integrated explanation of the mechanism which starts with inflation and leads to lasting increases in real output, not just short-term effects. Have these details been measured and confirmed by experience?

I confess I also don't know the following things:

() How does the central bank cause inflation, if not through lending too cheaply into the investment market, say at low interest rates? Would Keynes call for a helicopter drop of cash?

() Market participants see the CB doing this. What other undesireable effects may happen? Could prices go up without much increased real output? I have read that inflation is bad for real output.

() Is anyone injured by this policy, say cash savers? How much of increased output would be funded by their losses? Say there is an overall benefit, that Increased Output minus Loss to Savers is positive; does that justify harming one group of people to help another group even more?

() In what way are prices sticky? Does sticky merely mean that some participants remain unaware of the new price pressure and so cannot adjust in time to avoid larger losses?

() When the stock market adjusted to new information in 2000 (tech bubble) and 2008 (housing valuation bubble) it adjusted in a week. That doesn't seem sticky to me.

Comments for this entry have been closed
Return to top