Sunday, March 9, 2014

Monetary Policy and the Financial Crisis

The FOMC transcripts for 2008 are indeed interesting reading. If you want to dip into these, I would recommend skipping some of the talk at the regular scheduled meetings, some of which is just formal statements by the participants, and pay close attention to the inter-meeting conference calls.

It's clear from the transcripts that the FOMC was dealing with situations for which it was unprepared. Even for people with a knowledge of historical financial panics - the National Banking era panics (1863-1913) and the Great Depression - what happened in 2008 was unprecedented. The fact that the dysfunction was centered in very large financial institutions, and in the financial markets in which they were active, rather than in retail commercial banking, made all the difference. I'm going to focus on two main (and related) points: (i) Decisions by the FOMC about the fed funds rate target during most of 2008 ultimately mattered little for the effects of monetary policy, or for the unfolding crisis, during 2008; (ii) Effective power in the Federal Reserve System during the crisis was concentrated at the New York Fed and the Board of Governors. The Presidents of the regional Feds had little influence on the big decisions that drove the response to the crisis.

Monetary Policy and the Fed Funds Market

During the September 16 FOMC meeting, which occurred the day after the Lehman failure, Bernanke said the following:
We have been debating around the table for quite a while what the right indicator of monetary policy is. Is it the federal funds rate? Is it some measure of financial stress? Or what is it? I think the only answer is that the right measure is contingent on a model. As President Lacker and President Plosser pointed out, you have to have a model and a forecasting mechanism to think about where the interest rate is that best achieves your objectives. It was a very useful exercise to find out, at least to some extent, how the decline in the funds rate that we have put into place is motivated. In particular, the financial conditions do appear to be important both directly and indirectly—directly via the spreads and other observables that were put into the model and indirectly in terms of negative residuals in spending equations and the like. The recession dynamics were also a big part of the story. I hope that what this memo does for us—again, I think it’s extraordinarily helpful—is to focus our debate better. As President Plosser pointed out, we really shouldn’t argue about the level of the funds rate or the level of the spreads. We should think about the forecast and whether our policy path is consistent with achieving our objectives over the forecast period. I am sympathetic to the general view taken by the staff, which argues that those recession dynamics and financial restraints are important, that we are looking at slow growth going forward, and that inflation is likely to moderate. Based on those assumptions, I think that our policy is looking actually pretty good. To my mind, our quick move early this year, which was obviously very controversial and uncertain, was appropriate.
The "quick move" Bernanke was referring to here was the 3/4% drop in the fed funds rate target (to 2.25%) on March 18, following the assisted merger of Bear Stearns with JP Morgan Chase on March 16. At the September 16 FOMC meeting, in spite of the Lehman collapse, Bernanke seemed confident that a 2% fed funds rate target could be maintained through the crisis, as he stated later in the FOMC conference call on October 7:
On the economic growth side, what is particularly worrisome to me is that, before this latest upsurge in financial stress, we had already seen deceleration in growth, including the declines, for example, in consumer spending. Everyone I know who has looked at it—outside forecasters and the Greenbook producers here at the Board—believes that the financial stress we are seeing now is going to have a significant additional effect on growth. Larry gave some estimates of unemployment above 7 percent for a couple of years. So even putting aside the extraordinary conditions in financial markets, I think the macro outlook has shifted decisively toward output risks and away from inflation risks, and on that basis, I think that a policy move is justified. I should say that this comes as a surprise to me. I very much expected that we could stay at 2 percent for a long time, and then when the economy began to recover, we could begin to normalize interest rates. But clearly things have gone off in a direction that is quite worrisome.
Of course we know that things were about to go off in an even more worrisome direction, and a forecast in October 2008 of unemployment slightly above 7% for a couple of years was way off the mark.

So, was the decision to hold the fed funds rate target at 2% at the September 16 meeting a big mistake? On September 16, the most recent information the FOMC had was the July year-over-year PCE inflation rate, which was 4.2%, and the August unemployment rate number, which was 6.1%. Suppose the FOMC had been following Janet Yellen's "balanced" approach, i.e. a Taylor rule with equal weights on squared unemployment deviations and squared inflation rate deviations from their respective targets. Also assume, following Yellen, a target unemployment rate of 5.5% and a target inflation rate of 2%. Then, if the FOMC were thinking in terms of a long-run real interest rate of 2%, Yellen's rule would have dictated a fed funds rate target in excess of 4% in mid-September. Adjust downward for financial market stress (as Yellen would be inclined to do), and a 2% fed funds target does not look unreasonable. On September 16, the FOMC participants seemed to be thinking that financial market disruption would be on the order of what occurred in March 2008 when Bear Stearns was taken over. It's hard to argue why they should have thought otherwise, given what they knew at the time.

Thus, in order to make the case that the FOMC goofed on September 16, we would have to argue that Janet Yellen's view of the world is/was wrong. I don't think the people castigating the Fed for too-tight monetary policy in the face of the crisis want to make that argument.

But I think we can argue that the FOMC did not quite have a grip on how to read developments in the overnight financial markets, and the meaning of the fed funds rate at the time. To see this, first consider the time series for the effective fed funds rate and the 1-month T-bill rate (the shortest-maturity interest rate on U.S. government debt).
You can see that, before late 2007, the 1-month T-bill rate tracks the fed funds rate reasonably closely (with fluctuations in the difference that I don't quite understand). But, note that the margin between the two rates becomes more volatile from late 2007 through 2008. If we look at the difference between the two rates, we can see that more clearly:
Clearly there is a large margin, on average, and a lot of volatility, from late 2007 to mid-2008, and again around the Lehman failure. If we look only at 2007-08, the two interest rates look like this:
The fed funds market is unusual, relative to overnight markets that exist in other countries, and it has features that make it an awkward forum for central bank intervention, particularly in times of crisis. Trading in the fed funds market is over-the-counter, and some of it occurs through brokers, with the brokered transactions being the only ones that are accounted for in the calculation of the effective fed funds rate. Fed funds trades are unsecured. That's important, as in times of crisis, the interest rate on an individual fed funds market transaction will reflect counterparty risk associated with the borrower.

What do we know about the quantity of trading on the fed funds market? Perhaps not so much. This paper by Afonso, Kovner, and Schoar, at the New York Fed, measures fed funds quantities and prices using the "Furfine algorithm," which looks for patterns in payments through Fedwire that look like fed funds market transactions. The measurement is far from perfect, but the authors give us two important findings: (i) There was a substantial increase in the dispersion in interest rates across federal funds transactions following the Lehman collapse; (ii) there was only a small reduction in the quantity of fed funds market trading, post-Lehman. Thus, activity in the fed funds market reflected a substantial increase in counterparty risk after mid-September 2008, but the fed funds market did not "freeze."

The key measure of the effect of conventional monetary policy is the riskless overnight interest rate. That was certainly not the fed funds rate, especially during the financial crisis. Afonso, Kovner, and Schoar find a high degree of dispersion in fed funds transactions prices in fall 2008, reflecting substantial counterparty risk. But there seems to be something else going on, stretching back into 2007, which shows up in the charts in the time series for the 1-month T-bill rate and the fed funds rate. Unless there is something peculiar about 1-month T-bills, I think we can infer that the risk free overnight rate had fallen substantially below the fed funds rate as far back as mid-2007. Here's what the two interest rates look like from August-December 2008:
So, note that the 1-month T-bill rate is in low territory post-Lehman. It's at most 1%, and is typically under 50 basis points. Further, we can see in the next chart that payment of interest on reserves (initially different rates on required and excess reserves) beginning in October 2008 failed to put any kind of floor on the 1-month T-bill:
Therefore, given the actions of the Fed, including lending through the discount window, lending through the TAF (term auction facility), lending to AIG (beginning September 17, 2008), etc., the effective policy rate appears essentially to have gone to zero when Lehman failed - and stayed there. Some people on the FOMC appear to have been thinking the same thing, as is evident in the October 7 conference call:
MR. LACKER. ...I think it would be worthwhile, as we go on with financial markets in such turmoil, to reflect on whether ... our lending is the equivalent of pushing on a string, to use another metaphor from the Great Depression.
CHAIRMAN BERNANKE. Quoting Keynes, I see, Jeff.
Once the riskless overnight rate goes to zero, there's a liquidity trap. Broad-based lending programs or overnight repo transactions by the Fed will have no effect, though if trading has indeed ceased in segments of the financial market, or financial insitutions are indeed facing liquidity crises, then targeted lending or targeted asset purchases by the Fed can matter. You'll note in the quote above that central bankers can have a little fun while dealing with financial crises, apparently. Lacker is known for being on the laissez faire end of the central-bank-intervention spectrum, and Bernanke is kidding him for sounding like a Keynesian (though there is actually nothing particularly Keynesian about a liquidity trap).

So, what to make of this? There are hints that, when the Fed achieves "liftoff" (an increase in the Fed's policy rate above 0.25%), the New York Fed's standard day-to-day intervention mechanism will change. In particular, it seems the Fed plans on having a substantial quantity of reverse repos on its balance sheet on a regular basis, and may target an overnight repo rate rather than the fed funds rate. This will require that the Fed have timely information on overnight repo transactions on some homogeneous set of repos - e.g. those collateralized with a standard class of Treasury securities. I'm not sure if such information is readily available, but it needs to be for this to work. This will eliminate the effects of counterparty risk on the targeted overnight rate, and other features that were creating the margin between the fed funds rate and Treasury securities in 2007-8. This can only make monetary policy more effective when the next financial crisis happens. As well, repo market intervention by the Fed would tighten up arbitrage in the overnight market of interest for the Fed, for reasons discussed here.

Financial Crisis Intervention and the Distribution of Power in the Fed System
As I've mentioned before, the key policy decisions made by the Fed during the financial crisis related to lending programs (discount window and TAF), institution-specific interventions (e.g. related to the Bear-Stearns merger and lending to AIG), and unusual asset purchases (e.g. of commercial paper). Some of the regional Fed Presidents played some role in these decisions. For example, the Richmond Fed President would have played some role in the takeover of Wachovia (headquartered in Charlotte, NC) Wells Fargo, and the Boston Fed President would have been consulted in relation to intervention with respect to money market mutual funds (headquartered in Boston). But, much of the intervention involved only decisions by the Board of Governors and the New York Fed, or day-to-day judgement calls by the New York Fed. For example, the lending program to AIG, announced September 16, resulted from a decision by the Board, in consultation with the Treasury.

Some FOMC participants were worried about the ad hoc nature of the interventions, particularly with regard to large financial institutions. This statement by Bernanke is particularly revealing:
I have been grappling with the question I raised for President Lacker, and I would be very interested in hearing other views either now or some other time. The ideal way to deal with moral hazard is to have in place before the crisis begins a well-developed structure that gives clear indications in what circumstances and on what terms the government will intervene with respect to a systemically important institution. We have found ourselves, though, in this episode in a situation in which events are happening quickly, and we don’t have those things in place.
We don’t have a set of criteria, we don’t have fiscal backstops, and we don’t have clear congressional intent. So in each event, in each instance, even though there is this sort of unavoidable ad hoc character to it, we are trying to make a judgment about the costs—from a fiscal perspective, from a moral hazard perspective, and so on—of taking action versus the real possibility in some cases that you might have very severe consequences for the financial system and, therefore, for the economy of not taking action. Frankly, I am decidedly confused and very muddled about this. I think it is very difficult to make strong, bright lines given that we don’t have a structure that has been well communicated and well established for how to deal with these conditions. I do think there is some chance—it is not yet large, but still some chance—that we will in fact see a much bigger intervention at the fiscal level. One is tempted to argue that by doing more earlier you can avoid even more later, but of course that is all contingent and uncertain. So we will collectively do our best to deal with these very stressful financial conditions, which I don’t think will be calm for some time.
The idea that a consistent policy framework for crisis intervention did not exist at the Fed, and that the absence of such a framework may have contributed to the severity of the crisis, is an idea that has been explored, post crisis, by Jeff Lacker.

Here's something I think is interesting. The next chart shows repurchase agreements and reverse repurchase agreements on the Fed's balance sheet.
For the Fed, a repo is an asset - a secured loan to a private sector financial institution. Similarly, a reverse repo is a liability for the Fed - a loan by a private sector financial institution to the Fed. Repos can be overnight assets, or of longer maturity. As is evident in the chart, before the financial crisis the Fed was typically on both sides of the repo market on a given day, but there was much more variability in repos than in reverse repos. Thus, day-to-day intervention by the New York Fed consisted primarily of variation in the quantity of overnight repos to hit the fed funds target rate. Beginning in late 2007, and in a more pronounced way in 2008, the Fed began taking a large net position in the repo market, to the point where it was lending about $80 billion on net in the repo market in mid-2008. Presumably this was intervention intended to inject liquidity into the "shadow banking" system.

Next, look in more detail at only 2008:
Here, note that the New York Fed engineered a large increase in reverse repos after the Lehman collapse, to the point where it essentially became a large repo financial intermediary, borrowing from one set of repo market participants and lending to another set. Further, from early October 2008 until the end of the year, Fed repos were constant, and all the variation in Fed repo market activity was in reverse repos. In the second-to-last chart, you can see that repo holdings went to nil early in 2009, and reverse repos have increased, most recently to an average of about $200 billion, as the Fed experiments with what is likely to form the basis for a new operating strategy.

The New York Fed has a substantial amount of latitude in how it manages the System Open Market Account (SOMA). For example, the decisions about the repo market interventions that we can see in the last two charts appear to have emanated from the New York Fed. Over time, power in the Fed has become increasingly concentrated in Washington and New York. During the financial crisis, there was even more relative power wielded by the New York Fed in particular, given its proximity to Wall Street, and the nature of the policy interventions. Further, I think the crisis produced a permanent increase in the power of the Board of Governors and the New York Fed relative to the rest of the Fed system. I haven't looked at these numbers, but my guess is that we would see a substantial shift in resources - wages and other costs - in terms of what is spent in Washington and New York, vis-a-vis the regional Feds. That should certainly cause some concern in the regional Feds, and I think it's also a public concern. Part of the strength of U.S. central banking is (was?) decentralization.


  1. "That should certainly cause some concern in the regional Feds, and I think it's also a public concern. Part of the strength of U.S. central banking is (was?) decentralization."

    The large majority of people care about a central bank that cares, besides inflation, also about employment stabilization and financial stability. The big advantage of the Fed (over e.g. the ECB) is that it is not dominated by hawks (aka folks who care about rentiers and not the working man or financial stability).
    How federal a central banking system is is unrelated to this core issue.

    1. You think perhaps that the regional Fed Presidents don't care enough about the working man and financial stability?

    2. You know you've got a crackpot on hand when he uses the word "rentier".

    3. @Stephen: No. I just fail to see why a more federal central banking structure should be per se superior or more democratic than a more centralized one.
      I think somebody who is into political economy could talk more in depth about the issue of federalization.

      @Anon: You are in dire need of actually studying Econ 101, rentier is a standard term in our profession.

    4. I've seen this at work. In a centralized central banking arrangement, the tendency of underlings in the organization is to support the ideas of your superiors - you're punished for being contrary. The decentralized system in the U.S. permits independent ideas to develop in the regional Feds, and those ideas compete with each other. In the U.S., regional Fed Presidents (especially recently) have been speaking out and criticizing the status quo policies. That doesn't happen, typically, in other countries, and I think it's a good thing.

    5. I disagree. Bernanke and Yellen are respectively have been fine heads of the Fed.

      Of course there are always some folks who think that stopping QE or increasing rates in the worst recession are a smart thing to do. Doesn't mean that they are not totally wrong and that the implentations of their "ideas" would not make millions of people worst off.

      I don't buy the notion that a more federal system automatically leads to more competition and to less butt-licking.
      Think about the Civil Rights movement and how the federal government had to step in to prevent the mistreatment of black folks, think about how people can be discriminated against in a small community (the ideal picture of people who favour federalism).

      Perhaps I am living to close to France but democracy and centralism go hand in hand. This might sound too Jacobin for some folks but it simply takes a strong central power to crush reactionary forces (also applies for monetary policy).

    6. The Other AnonymousMarch 14, 2014 at 10:59 AM

      Would you also disagree if, say, Ed Prescott became head of the Fed?

    7. I fail to see how more federalism can prevent the emergence of braindead chairmen of the Fed.

      It's more a matter of dominance of ideas and interests. For example on this side of the big pond there is a slightly federal central banking system but it has not prevented the dominance of Germany's hard money stance and the ensuing catastrophe (until Draghi pushed through OTM).

      One does not need to have watched The Wire to understand that stupid ideas that create immense human suffering can emerge in any institutional setting.

    8. "I disagree. Bernanke and Yellen are respectively have been fine heads of the Fed."

      1. Yellen's just started on the job. Too early to say whether she's any good at this.
      2. One of Bernanke's strengths is that he is collegial. The regional Fed Presidents have had much more influence on policy under Bernanke than Greenspan.

  2. Steve, I've got a question for you. You probably won't be able to answer it fully, I think its an open question in the literature, but I just wanted to get your thoughts.

    With current monetary policy following an approximate Taylor rule (zero lower bound permitting), how close does this policy rule get to a policy chosen to maximize a social welfare function. It is my understanding that the derivation of the Taylor rule come from the Feds desire to uphold its duel mandate and thus minimized the sum of squared deviations of output and inflation from their steady state levels. But how well does a Taylor rule perform if the objective function were instead a social welfare function that incorporates a societies desire to increase financial stability (because of risk aversion), societies likely preference of having the output gap smaller than the inflation gap (why do people care about the inflation gap? its pretty reduced form and unclear in the models I have seen, even LW to some extent), and societies preferences for redistribution (which of course the specific nature of these preferences is unknown but newer papers are trying to answer this question)?

    I am not suggesting imposing these preferences in the SWF, but that these general ideas would come out of the standard preferences macroeconomists typical impose (risk-aversion gives a desire for financial stability, concave utility gives a desire for redistribution, preferences over consumption and not money reduce the importance of the inflation gap relative to the output gap).

    My intuition tells me the Taylor rule would not perform well, mostly because of the third channel (relative importance of the output gap to inflation gap). But I feel the other two channels would have more ambiguous impacts on the Taylor rule. What do you think?

    1. When Taylor first wrote down his rule, he didn't make any claims that there was any theory which would justify it as some welfare-maximizing policy rule. It seemed to capture the idea that the Fed should care about inflation, and that there exist some non-neutralities of money which the Fed could exploit in influencing real economic activity. He then claimed that it worked pretty well (in terms of an ad hoc welfare criterion) in some quantitative models. Woodford used the Taylor rule to obtain determinacy in NK models, and even argued that it was optimal under some special circumstances. But NK models are very special. Typically they ignore financial factors that I think we can agree are important, and certainly don't address issues to do with what you're calling "financial stability."

      So what's optimal? I don't think we have a good grip on this. I don't think we understand fully the nature of the central bank's influence on real economic activity, and we don't fully understand the costs of inflation. 2% inflation is optimal? Why?

      You might hope that we could muddle along with a simple Taylor rule driving central bank behavior. It's simple, and maybe close enough to optimal, given what we know. But it's well-known that it can have bad long run properties. For example, the Fed can think that 2% inflation is optimal, but converge to a long-run equilibrium in which inflation is too low. Or the Fed can have incorrect beliefs about the long-run real interest rate, or the output gap, which will imply that the policy is wrong.

    2. Many people feel that having a rule (so long as it is not obviously bad) is more important than making sure that the rue is optimal. The strength of the Taylor rule seems to be that it represents a reasonable interpretation of the dual mandate.

    3. "Many people feel that having a rule (so long as it is not obviously bad) is more important than making sure that the rue is optimal."

      I'm beginning to think that this view is wrong. The real rate of interest is endogenous, and affected by various factors other than monetary policy. We can't measure output gaps directly, and no one agrees on how to measure them. So any Taylor rule could go astray in a bad way. Look at what happened with the Fed's forward guidance. That was a kind of rule. Now the Fed is having serious second thoughts about the 6.5% unemployment rate threshold. If they could go back in time and choose another rule, they would do it.

    4. I gotta agree. Rules are superior to discretionary policies because of the usual time inconsistency problem and because it enhances transparency and reduces uncertainty ... but when there is a regime change the CB clearly has to change its rule.
      And when there is a not-so-standard, once-in-a-lifetime occurring recessions like right now you don't wanna inflexible folks at the helm.