The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.

Authors for macroblog are Dave Altig, John Robertson, and other Atlanta Fed economists and researchers.

« Prospects for a small business-fueled employment recovery | Main | A look at another job market number »

October 13, 2009

Reviewing the recession: Was monetary policy to blame?

Reviewing the Recession: Was Monetary Policy to Blame?
In a recent speech given at the University of South Alabama, Federal Reserve Bank of Atlanta President Dennis Lockhart added his voice to what is now the general consensus: "I agree with all who are declaring that a technical recovery is under way."

There is still much work to be done, of course, not least the continuing examination of just what led to the recession and how a repeat performance can be avoided. One theme of this examination appeared in last week's Financial Times:

"It is certainly true that the most recent bubble, its bursting and the Fed's actions in the aftermath have inspired existing critics and recruited new ones. The first charge is that interest rates under Alan Greenspan, [current Fed Chairman Ben] Bernanke's predecessor, were kept too low for too long, contributing to a bubble of easy credit."

Here is an exercise that that I find intriguing. Suppose we try to estimate, as closely as we can, the actual interest rate decisions of the Federal Open Market Committee (FOMC) over the period spanning the beginning of Greenspan's tenure to the present. It turns out that by using an approach based not only on measures of inflation and actual output relative to potential but also on the lagged fed funds rate, you can actually get pretty darn close to statistically describing what the FOMC did:


I'm going to resist the temptation to call this approach a "Taylor rule." The estimating "rule" used here does in fact include realizations of inflation and a measure of the output gap (that is, a measure of how different gross domestic product is from its potential). These are the essential ingredients of the Taylor rule, but not the source of the close fit evident in the chart above. Over the period covered by the chart, you could have done a pretty good job mimicking the actual federal funds rate outcomes in any given month using knowledge of the previous month's rate. (If you are interested in the details of the estimating rule, you can find them here.)

The interpretation of the tendency for today's federal funds rate to generally follow yesterday's rate—sometimes referred to as interest rate smoothing—is controversial. Glenn Rudebusch (from the Federal Reserve Bank of San Francisco) explains:

"Many interpret estimated monetary policy rules as suggesting that central banks conduct very sluggish partial adjustment of short-term policy interest rates. In contrast, others argue that this appearance of policy inertia is an illusion and simply reflects the spurious omission of important persistent influences on the actual setting of policy."

Rudebusch is decidedly in the second camp, but for our purposes here the exact interpretation may not be that important. Though I am glossing over some not insignificant caveats—such as the difference between final data and the information the FOMC had to react to in real time—the chart above suggests that whatever the underlying structure of policy decisions, after the fact the FOMC appears to have behaved in an extraordinarily consistent way over the period extending from the late 1980s. This observation, in turn, suggests to me that there was nothing all that unusual about monetary policy in 2003 once you account for the state of the economy.

Which leads me to my main point on the chart above: If you are of the opinion that interest rate policy was good through the late 1980s and 1990s, then there seems to be a good case the FOMC was just sticking with "proven" success as it set interest rates through the dawning of the new millennium.

There is, of course, the possibility that the pattern of the funds rate depicted in the chart above was incomplete all along in the sense that whatever variables are explicit and implicit in the estimated rule, they did not include information to which the FOMC should have responded. In calmer times, the story would go, not including potentially pertinent information was not much of a problem. Eventually the missing-data chickens could come home to roost. The prime omitted variable suspect would, of course, be some sort of asset prices.

Scratch any gathering of macroeconomists these days and out will bleed a steady stream directed at incorporating credit and financial market activity into thinking about the aggregate economy. The necessity of proceeding with that work was emphasized by no less an authority than Don Kohn, vice chairman of the Federal Reserve Board of Governors, speaking at just such a gathering of macroeconomists last week:

"It is fair to say, however, that the core macroeconomic modeling framework used at the Federal Reserve and other central banks around the world has included, at best, only a limited role for the balance sheets of households and firms, credit provision, and financial intermediation. The features suggested by the literature on the role of credit in the transmission of policy have not yet become prominent ingredients in models used at central banks or in much academic research."

I will admit that economists were not exactly ahead of the curve with this agenda, but prior to 2007 it was not at all clear that detailed descriptions of how funds moved from lenders to borrowers or how short-term interest rates are transmitted to longer-term interest rates and capital accumulation decisions were crucial to getting monetary policy right. Models without such detail tended to deliver policy decisions not far from the sort depicted above, and, as I noted, they seemed to be working quite well in terms of macroeconomic outcomes.

Thus far, one of the lessons from models in which financial intermediation is taken seriously is that interest rate spreads or stock prices or other asset prices do become part of the policy rate recipe. Your response might well be something along the lines of "duh." You are entitled to that opinion, and I won't push back too hard. But it is yet far from clear that the financial crisis can be explained by a misstep in the setting of the federal funds rate caused by the failure to make whatever adjustments might have been indicated by the inclusion of pertinent financial variables in implicit rate-setting rules of thumb. Furthermore, early versions of research I have seen that combines capital regulation policy and interest rate policy suggest that the macroeconomic consequences of getting the former wrong may be much greater than the consequences of getting the latter wrong. To me, that conclusion has the ring of truth.

By David Altig, senior vice president and research director at the Atlanta Fed

October 13, 2009 in Federal Reserve and Monetary Policy , Monetary Policy | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Reviewing the recession: Was monetary policy to blame? :


On the former G.6 release, all debits cleared thru demand deposits, with the exception of Mutual Savings BAnks (an obvious error).
I.e., real estate, financial transactions, etc. all cleared thru DDs. Thus if you take the 2 rates-of-change in bank debits, you can measure all economic activity (including bubbles). Contrary to all other economic theory, the methodology has been infallable.

Posted by: flow5 | October 13, 2009 at 02:21 PM


The graph is very intriguing. One question I have is what inflation measures are being used. If one had substituted the Case-Shiller index of home prices for the Owner's equivalent rent part of the CPI, would that significantly change the resultant interest rates?

If the inflation metrics used to set policy aren't adequately representing true inflation, then the prescribed policy response as set by the so-called rule may also be wrong.

Posted by: uber_snotling | October 13, 2009 at 02:33 PM

At what point does coming up with a myriad of form fitting Taylor rules just look like Butter in Bangladesh (datamining)? Why does the fact that there exists some mechanical rule amidst many possible ex ante formulations which renders all Fed policy consistent provide information whether that is a meaningful consistency? I can't help but wonder if we could come up with a seemingly relevant Taylor-type formula that would fit almost any pattern of random Fed Funds rates. Of course, I also don't think deviating from one of these supposedly tight-fit rules like the actual Taylor is likewise great evidence that the Fed made a mistake in the first place, though many seem to disagree.

Posted by: dlr | October 13, 2009 at 02:55 PM

The recession had as its origin the sub-prime lending crisis. Further, home prices in many US cities were being driven very high, year after year, in comparison to the CPI. However, this price growth of this one asset, residential housing, could not have occurred without the strong availability of mortgage credit. It follows that what is required is a very intense study of the sources of that mortgage credit. In the end, it is money supply, but from what origin? Interest rates alone cannot account for the availability of mortgage credit. The mortgage credit availability drove a very powerful "inflation" in one asset. The consequence of the eventual "deflation" was the dire impact on the balance sheets of the banks, as their collateral collapsed and the sure-fire real estate growth ceased and then reversed. It seems quite doubtful that the Fed by itself could have caused all of the excess mortgage credit availability. This is what must be studied, what kind of institutions made the mortgage loans and from whence came the money supply to fuel the loans. The answer is likely quite complex.

Posted by: mme | October 13, 2009 at 09:37 PM

David - I enjoy your posts, but I wanted to add one observation about this particular argument. I agree that a monetary policy misstep had nothing to do with the proximate cause of the current events, but some of those who criticize the Fed...of whom I am one; see my book "Maestro, My Ass"...actually are critical of the stability and forecast-ability of the rate. By leaching uncertainty from the market, the Fed enabled much greater leverage than if they had been spastic, or at least more stochastic. Your post actually demonstrates that the rate path was HIGHLY predictable (which we knew, but it's a great illustration of just how predictable), and that is in fact the main error the Fed made...not the level, but the variance.

Keep the posts coming! They are very thoughtful.

Posted by: Mike Ashton | October 14, 2009 at 08:01 AM

The technical appendix to your estimating rule states: "The primary difference from the original Taylor rule is the inclusion of the lagged Fed funds rate". Taylor likes to remind people that according to his understanding of his rule, the Fed funds rate became too low around the beginning of '02 and didn't catch up with his rule until sometime around '06 (see, e.g., "The Financial Crisis and the Policy Responses: An Empirical Analysis of What Went Wrong"). Your estimating rule works well as an approximation of actual interest rate decisions, so is the inclusion of the lagged Fed funds rate really making up essentially all of the difference between how well your rule works and how well Taylor's rule works?

Posted by: Rich Fitzgerald | October 14, 2009 at 08:32 AM

There are some other aspects to the "rates to low" argument that are essentially implicit in your "rule". For one thing we were after all entering a 2nd war while still being involved in a first in '03. Then job creation was extremely weak and didn't pick up until '03 and peaked on a YoY basis in '04, indicating a very weak recovery. Finally there is the infamous conundrum - long rates stayed low even after s.t. were raised. What else was the Fed to do - quantitative tightening? We now know (ahem) that the conundrum was based on "excess" savings in the trade surplus countries being re-cycled into the US to finance over-consumption on leveraged debt. Which by the by gets to the heart of the real criticism of Uncle Allen's last years - the failure to enforce existing regulatory regimes. In other words you are so right about which factors turn out to be important. Sadly none of this has entered the common wisdom or general discussions.

Posted by: dblwyo | October 14, 2009 at 09:33 AM

In my experience, the rates-were-too-low crowd can be shut up by asking two questions:

1. How much higher should rates have been? (If this doesn't stop them in their tracks completely, they will nominate something fairly moderate, say 2 percentage points. Scarily few of the people in this camp have actually thought about the answer to this question.)

2. Do you REALLY think that rates X percentage points higher would have stopped the speculative behaviour in housing and credit markets?

The Fed could not have stopped the bubble with interest rates alone. They would have had to jack rates up so far it would have killed the economy.

Posted by: michelle color me shocked | October 14, 2009 at 07:08 PM

I think you mean:

"The Fed COULD have stopped the bubble with interest rates, BUT they would have had to jack rates up so far it would have killed the economy."

Of course this was already baked in to the picture. Wealth cannot be created by simply increasing the money supply, the misallocation of capital that is caused by steady increases in unsound money supply always results in a in a future crisis...cause by the misallocation of capital. This is good for the banks that have implicit ability to tax the masses to socialize losses whenever they want(Goldman Sachs, JP Morgan), but it is bad for all of those who cannot(regular hard working people raising families).

Posted by: Gabe | October 15, 2009 at 10:13 AM

Given that the Great Recession had as one of its primary causes the government-sponsored easing of residential mortgage lending criteria over the past decade plus, I am intrigued by the statement that "...early versions of research...that combines capital regulation policy and interest rate policy suggest that the macroeconomic consequences of getting the former wrong may be much greater than the consequences of getting the latter wrong." Such research is sorely needed.

Posted by: sts | October 15, 2009 at 04:19 PM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

Google Search

Recent Posts



Powered by TypePad