Revisiting the Causes of the Great Recession

Ramesh Ponnuru and I have an Op-Ed in today's New York Times:

IT has become part of the accepted history of our time: The bursting of the housing bubble was the primary cause of a financial crisis, a sharp recession and prolonged slow growth. The story makes intuitive sense, since the economic crisis included a collapse in the prices of housing and related securities. The movie “The Big Short,” which is based on a book by Michael Lewis, takes this cause-and-effect relationship as a given. 

But there is an alternative story. In recent months, Senator Ted Cruz has become the most prominent politician to give voice to the theory that the Federal Reserve caused the crisis by tightening monetary policy in 2008. While Mr. Cruz (who is an old friend of one of the authors of this article) has been criticized for making this claim, he shouldn’t back down. He’s right, and our understanding of the great recession needs to be revised.

The crux of our story is that what would have been an ordinary recession got turned into the Great Recession because the Fed failed to do its job. Readers of this blog will be familiar with this argument. For those who are not here is a brief recap of the evidence supporting our claims. 

First, the Fed contained the fallout from housing crisis for almost two years. This can be seen in the three figures below. The first two figures show that even though housing peaked in early 2006, employment and personal income outside of housing-related sectors actually grew at a stable rate up until about early-to-mid 2008.

.

This third figure shows that nominal spending overall continued to grow fairly stable during the initial run on the shadow banking system, as seen by the Ted spread. That run occurred from occurred from August 2007 to about May 2008. It also shows that the biggest spike in the Ted spread occurs only after the Fed allows nominal spending to start dropping.

Second, the Fed tightened policy in 2008 and did so in two phases. First, beginning around April 2008 the Fed began signalling it was planning to raise interest rates as seen in the figure below. It shows the market expectation of the federal funds rate one year in advance. It rises all the way through the summer of 2008 and remains higher than the actual federal funds rate through October 2008. This is the first phase of tightening in 2008. It was an explicit tightening, albeit of the future path of monetary policy.

 

The second stage, as we note, in the Op-Ed occurs in the second half of 2008. Here the natural interest rate is falling fast and the Fed fails to lower its target interest rate until October 2008. This is a passive tightening of monetary policy and is reflected in the decline in expected inflation and nominal spending that starts in mid-2008.

To summarize, our argument is that the Fed was doing a decent job responding to the housing bust up until 2008. After that point it tightened monetary policy and catalyzed the reaction that lead to the Great Recession. By the time the Fed changed course in late 2008 it was too late. Interest rates had already cross the zero lower bound (ZLB). Once that happens monetary policy as it is currently practiced cannot do much. (For more on the crossing of the ZLB see my review of Atif Mian and Amir Sufi's book on the crisis. Update: also see this twitter conversation with Amir Sufi on the ZLB.) The central bank of Australia, however, acted sooner and never faced the ZLB problem, despite having a housing and debt bubble too.

For interested readers, I would direct to the work of Richmond Fed economist Robert Hetzel who has written an article and book that makes the same argument.  Also, see Scott Sumner's early critique of Fed policy during this time as well. 

P.S. Just to demonstrate how worried the Fed was about inflation rather than growth late into the crisis, here is an excerpt from the minutes of the August 2008 FOMC meeting. Note they they were expecting to raise rates at the next meeting. 

The Latest Central Bank Fad: Asymmetric Inflation Targeting

I have noted many times here how the Fed's treats its 2% inflation target as more of a ceiling than a symmetric target. Apparently, the ECB is even more brazen in its asymmetric interpretation of its inflation target (my bold):

The ECB has got itself into an extraordinarily difficult position. It has missed its policy target — a headline rate of inflation at “close to but below” two per cent — for four years. The target has lost credibility. Once people have lost confidence in an inflation target, it becomes very hard for the central bank to persuade them to trust the target again. 

It was touching to hear Mr Draghi last Thursday talk about failing to reach a goal, then to try again and to fail again. I do not doubt his determination but the minutes of the December 3 meeting of the governing council tell us that not everybody supports the target in the same way... One [governor] said that he would not accept a further increase in QE unless the eurozone was once again in deflation. The implicit message of that statement is that this particular governor’s policy target must be zero per cent, not two per cent. He will only act once prices actually fall. 

[...] 

[A former governing council member] confirms something I had suspected for a long while but was never able to confirm: he cares if inflation is above the target but less so when it is low. The target becomes asymmetric... Germany’s economic establishment has its unofficial inflation target, which I would put at a range of 0-2 per cent. If that were the target, no policy action would be needed now. 

[...] 

To me this all shows that, as an institution, the ECB is only partially committed to its stated goal. This is one of the reasons why it keeps missing its policy target.

This can mean only one thing: it is time to update my inflation targeting chart from this earlier post. Asymmetric inflation targeting seems to be the new fad at central banks, at least the big ones. 

The Balance Sheet Recession That Never Happened: Australia

Probably the most common explanation for the Great Recession is the "balance-sheet" recession view. It says households took on took on too much debt during the boom years and were forced to deleverage once home prices began to tank. The resulting drop in aggregate spending from this deleveraging ushered in the Great Recession. The sharp contraction was therefore inevitable.

But is this right? Readers of this blog know that I am skeptical of this view. I think it is incomplete and misses a deeper, more important story. Before getting into it, let's visit a place that according to the balance sheet view of recessions should have had a recession in 2008 but did not.

That place is Australia. It too had a housing boom and debt "bubble". It too had a housing correction in 2008 that affected household balance sheets. This can be seen in the figures below:

 

 

 

Despite the balance sheet pains of 2008, Australia never had a Great Recession. In fact, it sailed through this period as one the few countries to experience solid growth. And, as Scott Sumner notes, it was also buffeted by a collapse in commodity exports during this time. If any country should have experienced a sharp recession in 2008 it should have been Australia.

So why did Australia's balance sheet recession never happen? The answer is that the Reserve  Bank of Australia (RBA), unlike the Fed, got out in front of the 2008 crisis. It cut rates early and signaled an expansionary future path for monetary policy. It also helped that the policy rate in Australia was at 7.25 percent when it began to cut interest rates. This meant the central bank could do a lot of interest rate cutting before hitting the zero lower bound (ZLB). So between being more aggressive than the Fed and having more room to work,  the RBA staved off the Great Recession.

This experience in Australia speaks to why the balance sheet recession view miss the deeper, more important problem behind depressions: the ZLB. Unlike the RBA, the Fed was slow to act in 2008 and that allowed the market-clearing or "natural" interest rate to fall below the ZLB. Had the Fed acted sooner or had it been able to keep up with the decline in the natural interest rate once it passed the ZLB, the Great Recession may not have been so great (See Peter Ireland's paper for more on this point).

Here is how I made this point in my review of Atif Mian and Amir Sufi's book, House of Debt, in the National Review.

Why should the decline in debtors' spending necessarily cause a recession?

Recall that for every debtor there is a creditor. That is, for every debtor who is cutting back on spending to pay down his debt, there is a creditor receiving more funds. The creditors could in principle provide an increase in spending to offset the decrease in debtors' spending. But in the recent crisis, they did not. Instead, households and non-financial firms that were creditors increased their holdings of safe, liquid assets. This increased the demand for money. This problem was exacerbated by the actions of banks and other financial firms. When a debtor paid down a loan owed to a bank, both loans and deposits fell. Since there were fewer new loans being made during this time, there was a net decline in deposits [and thus] in the money supply. This decline can be seen in broad money measures such as the Divisia M4 measure. These developments—increase in money demand and a decrease in money supply—imply that an excess money-demand problem was at work during the crisis.

The problem, then, is as much about the excess demand for money by creditors as it is about the deleveraging of debtors. Why did creditors increase their money holdings rather than provide more spending to offset the debtors? ...Mian and Sufi do briefly bring up a potential answer: the zero percent lower bound (ZLB) on nominal interest rates.

The ZLB is a floor beneath which interest rates cannot go. This is because creditors would rather hold money at zero percent than lend it out at a negative interest rate. This creates a big problem, because market clearing depends on interest rates' adjusting to reflect changes in the economy. In a depressed economy, firms sitting on cash would start investing their funds in tools, machines, and factories if interest rates fell low enough to make the expected return on such investments exceed the expected return to holding money. Even if the weak economy means the expected return to holding capital is low, falling interest rates at some point would still make it more profitable to invest in capital than to hold money. Similarly, households holding large amounts of money assets would start spending more if the return on holding money fell low enough to make household spending worthwhile. This is a natural market-healing process that occurs all the time. It breaks down when there is an increase in precautionary saving and a decrease in credit demand large enough to push interest rates to zero percent. If interest rates need to adjust below zero percent to spur creditors into providing the offsetting spending, this process will be thwarted by the ZLB.

 It is the ZLB problem, then, rather than the debt deleveraging, that is the deeper reason for the Great Recession.

Australia never hit the ZLB. That is why it avoided the Great Recession. If we want to avoid future Great Recessions we need to find better ways to avoid or work around the ZLB.

Has Macroeconomic Policy Been Overly Tight?

Former Fed Minneapolis Fed President Narayana Kocherlakota believes macroeconomic policy has been overly tight the past few years. Consequently, he thinks the Fed is getting ahead of the recovery with its current tightening cycle. Is he right, has macroeconomic policy really been overly tight? To answer this question, consider the four following pieces of evidence. 

1.  Inflation has consistently fallen below the Fed's two percent inflation target for the past seven years. Here is a visual representation of this this development in terms of a shooting target:

Yes, core PCE inflation has has averaged near 1.5% over the past seven years despite a 2% inflation target. It is as if someone has been doing target shooting and persistently hits the lower half of the target. Maybe the Fed is actually aiming for something other than 2%--possibly a 1%-2% inflation corridor target--or maybe the Fed in its current form is not as powerful as we thought. Either way, inflation is being systematically kept below 2%. This suggests relatively weak aggregate nominal spending growth. This is consistent with tight macroeconomic policy. 

2. The risk-free real interest rate appears to still be cyclically adjusting. That is, the 10-year treasury interest rate adjusted for the risk premium appears to be following the prolong closing of the output gap. Some have confused the low levels of real interest rates as evidence for secular stagnation. As I have argued elsewhere, this need not be the case. The long decline in real interest that most observers invoke as the smoking gun for secular stagnation is misleading because it does not correct for the rise in the risk premium during the 1970s. Once one corrects for that you get the following figure:

The black line is the 10-year risk-premium adjusted real treasury interest rate. Note that it averages just under 2%--roughly tracking the average growth rate of the economy--but deviates around that average. Since the 1980s, those deviations closely track the business cycle as seen below: 

What this implies is that slow return of risk-free real rate to a higher level is simply a reflection of the slow unwinding of this business cycle. That it has taken this long to adjust only part way suggests that macroeconomic policy has not been very supportive.  

3.  Household portfolios still inordinately weighted toward safe assets. If one looks at the holding of liquid assets by households--defined here as cash, checking, saving, time, money market mutual funds, treasuries, and agencies--as percent of total assets it shot up during the crisis and still has yet to return to pre-crisis levels.  This indicates household demand for safe assets remains slightly elevated. This can seen in the figure below. Note that this liquidity demand measure leads the broad unemployment rate.

This measure also tracks the CBO's output gap very closely as well:

Were macroeconomic policy highly accommodative we would have expected households to adjust their portfolios faster. But they have not and this suggest macroeconomic policy has been tight. That is why this measure tracks the above measures of slack so well.

4.  Total dollar spending is still below its full employment level. This last bit of evidence is model-based and is part of a paper I am finishing where I estimate various ways to gauge the appropriate level of aggregate nominal expenditures. The key point here is that the sharp drop in nominal spending during the crisis never has been fully corrected for even after adjusting for changes in potential real GDP. It is hard to reconcile this with loose macroeconomic policy.

In short, macroeconomic policy has not been very supportive of a robust recovery over the past few years. So I agree with Narayana Kocherlakota on this point. Where people can reasonably disagree, I think, is whether this was the Fed's fault. Stephen Williamson, for example, sees little that the Fed could have done. For him, it was fiscal policy that was deficient. Specifically, he thinks there should have been more U.S. treasuries to alleviate the safe asset shortage problem.

My own view is that the bigger problem is the body politic's rigid commitment to low inflation. There is no way the Fed or Treasury could have spurred significantly more nominal spending growth without there being a temporary increase in the inflation rate. And that is simply intolerable in the current environment. The safe asset shortage problem and inability of the Fed to get much traction is simply a symptom of this problem.

That is one reason why I am a big advocate of NGDP level targeting. It would allow for temporary deviations in the inflation rate while still providing a credible long-run nominal anchor. Until we get something like this, expect regular bouts of macroeconomic policy being overly tight.

A Small Step Toward Better Fed Policy

I have a new op-ed at Investors' Business Daily on the Fed's increasing use of the natural interest rate to guide monetary policy.

The Fed began a new chapter in its history in December by raising interest rates for the first time in almost a decade. A key reason for this historic liftoff of interest rates is the belief by Fed Chair Janet Yellen and other monetary officials that the "natural" interest rate has risen and the Fed must follow suit. 

          [...]

This is a much welcomed development but for one big problem: the Fed will not reveal to the public its estimates of the natural interest rate. This makes no sense. When the Fed started targeting the M1 money supply in the mid-1970s, it reported the actual M1 supply. Similarly, when the Fed switched to targeting non-borrowed reserves in 1979, it reported the actual non-borrowed reserves supply.

So why isn't the Fed reporting its natural interest rate estimates now that they are increasingly being used as a kind of target to guide interest-rate setting? 

It is true that the natural interest rate is not directly observable and that there is some uncertainty surrounding its true value. But the Fed has sophisticated estimates of the natural interest rate that account for this uncertainty, as Yellen revealed in a recent speech.

Here is the chart from Janet Yellen's speech that reveals the Fed's range of estimates for the short-run real natural interest rate. 

So why not update and publish this data on a regular basis? Imagine how much more informed Janet Yellen's congressional hearings would be if these estimates were available. They would provide all parties a common framework from which to discuss the Fed's actions. It would only help the Fed as I note in the op-ed:

Imagine if this information had been available over the past seven years. The Fed could have pointed to it every time someone claimed, "The Fed has artificially lowered interest rates!" and shown that it was simply tracking the natural interest rate, which had fallen below zero percent. It would have saved the Fed a lot of grief, but instead the Fed chose not to reveal this information.

One Fed economist, Vasco Curida of the San Franciso Fed, is already providing regularly updated estimates of the short-run natural interest rate. His estimates are in the chart below.

It would be great if the Fed would follow Curida's lead and regularly report its estimates too. This would be a small, but important step toward better monetary  policy.

P.S. As noted above, these estimates are for the short-run natural interest rate. They are different than the Laubach-Williams (LW)  series which is a medium-term estimate of the natural interest rate. This later measure is equivalent to the intercept term in the Taylor Rule.

Related Links
Savers' Real Problem
The Fed Did Not Enable the Large Budget Deficits--You Did!

The Fed Gets What It Wants: A 1%-2% Inflation Target Corridor

So it is finally time for lift off. The Fed is poised to raise short-term interest rates over the next few days after seven long years of ZIRP. Exciting as this development may be, it is important to keep in mind that the guiding principle behind the Fed's decisions during this time has not suddenly changed. This principle says that no matter what happens--whether it be ZIRP, QE, forward guidance, the changing winds of fiscal policy, or the normalization of monetary policy--the Fed must always act in a manner to keep core PCE inflation within a 1-2 percent inflation corridor. 

Yes, even though the Fed has an official 2 percent inflation target revealed preferences indicate the real force shaping Fed policy has been a 1-2 percent inflation target corridor over the past seven years. Once you understand this point all other Fed mysteries begin to clear up. For example, why did the Fed sterilize its lending to banks between December 2007 and October 2008? Or, why did it introduce IOR just as the markets were imploding or make the asset purchases under QE temporary? The answer is that it did not want rapid growth in nominal spending--even though it was sorely needed--for fear of pushing inflation too high.  The Fed, in other words, was willing to sacrifice the economy at the altar of the inflation target corridor. This framework is likely to continue going forward.

But don't take my word for it. Let's look to the data and let the FOMC's revealed preferences speak for themselves. 

Consider first the central tendency consensus forecasts of core PCE inflation by FOMC members.  The figures below show these forecasts for the current year, one-year ahead, and two-years ahead horizons. A clear pattern emerges from these figures as you expand the forecast horizon: 2 percent becomes a upper bound. FOMC officials, therefore, have been consistently looking at an upper bound of 2 percent for core PCE inflation. If we add to this fact that the FOMC has meaningful influence on inflation several years out, then these revealed preference are saying Fed officials actually want and expect to get an inflation upper bound of 2 percent. This inflation corridor is a choice.

 

 

Now consider the actual performance of core PCE inflation since the crisis started. This is where the 1 percent lower bound on the corridor becomes evident. The Fed seems ready to pull the monetary trigger if core inflation drifts too close to this lower bound. Currently, core inflation has stabilized around 1.3 percent but should it start falling again I would not be surprised to see the Fed getting trigger happy once again.

These revealed preferences of the Fed have begun to affect the public's long-term inflation expectations.The figure below shows the annual average inflation forecast over the next 10 years from the Survey of Professional Forecasters. I would not call these forecasts unanchored, but they are gradually drifting down. In the past we worried about expectations becoming unanchored as inflation expectations drifted upward. Now it seems they are drifting in the other direction, though presumably anchored by a lower bound.

The low inflation environment of the past few years seems entirely in line with the revealed preferences of FOMC officials. There is nothing mysterious about it. Fed officials are getting what they want. Unfortunately, aiming for a inflation corridor of 1-2 percent does guarantee macroeconomic stability. This inflation target range could be either be too tight or too easy depending on the state of the economy. It would be far better for the Fed to simply stabilize the growth path of aggregate spending. Until then, expect the Fed's decisions to be guided by the inflation target corridor. 

P.S. See Russ Robert's interview of George Selgin on EconTalk for an interesting discussion where some of the Fed's mysterious actions over the past seven years are discussed.

Update: Janet Yellen, made this point in the press conference following the December 2014 FOMC meeting (my bold):

But it’s important to point out that the Committee is not anticipating an overshoot of its 2 percent inflation objective (p.13).

Upgrading to Abenomics 2.0

What has the 'monetary arrow' of Abenomics accomplished? To answer this question, recall that this part of Abenomics called for the Bank of Japan (BoJ) to double the monetary base and raise inflation to 2%. On the former goal the BoJ has been successful. On the later goal it is still a work in progress, though inflation has been trending upward. 

Assessing the Abenomics inflation record is a bit tricky because the 2014 sales-tax hike in Japan put upward pressure on the inflation rate. Nonetheless, after accounting for the tax hike inflation is overall  moving up. This can be seen in the next two figures. The first one shows the core inflation rate since 2000. The second zooms in on the past few years and shows core inflation with and without the tax hike. An upward trend is apparent in this figure. Core inflation is now around 1%, the highest it has been since the late 1990s. So inflation is being pushed up, albeit at a slower paced than originally envisioned under Abenomics. 

More generally, Abenomics has been mildly successful at reflating the economy. The figure below shows nominal Japan's NGDP with the Abenomics period highlighted in red. It has risen relatively rapidly.

The figure also highlights NGDP during Japan's QE program of 2001-2006. It was comparatively flat and was supported by what turned out to be a temporary expansion of the monetary base. This suggests that at least some part of the monetary base expansion under Abenomics is expected to be permanent.  

So the monetary  arrow of Abenomics has been moderately successful at reflating the economy. This reflation, however, has not done much for the real economy. It could do a lot for Japan's debt burden if Abenomics were to continue raising NGDP. And that, in turn, could help reinvigorate the economy. Maybe that is why Prime Minister Shinzo Abe seems so intent on upgrading to Abenomics 2.0.

What is Abenomics 2.0 you ask? It is the Prime Minister's plan to explicitly raise the level of NGDP from its current value of approximately ¥500 trillion up to ¥600 trillion. No, the Prime Minister has not gone all Market Monetarist on us. He is not asking for a growth path target for NGDP, but a one-time 20% increase in the level of nominal spending. Just how ambitious is this goal? Below is a figure plotting out three hypothetical paths to ¥600 trillion NGDP. The five-year path appears to be most in line with Japan's Cabinet Office time frame, as noted by James Cox. By historical standards, this path would be very ambitious for Japan.

Prime  Minister Abe first called for this this goal in September and spoke to it again last week. Here is the Nikkei Asian Review on his latest speech:

TOKYO--Raising Japan's nominal gross domestic product to 600 trillion yen ($4.9 trillion) is a reachable goal, Prime Minister Shinzo Abe told a group of economists in Tokyo on Tuesday.

"We are aiming to achieve a virtuous cycle of growth and redistribution with our 'Society in Which All Citizens are Dynamically Engaged' plan. This is our proposal for a new socioeconomic system," the Prime Minister told the group.

"We will aim to build a diverse society where anyone can succeed. That will lead to new ideas, which will lead to further growth. Through growth and redistribution, I believe our target of achieving nominal GDP of 600 trillion yen is well within our reach." Japan's current nominal GDP is around at 500 trillion yen.

Abe told the gathering that the government is putting together a 3.5 trillion yen supplementary budget and that he hoped to use around 1 trillion yen of the extra money on his "Dynamic Engagement" project.

[...]

Bank of Japan Gov. Haruhiko Kuroda, speaking at the same forum, gave Abe a vote of confidence, saying the number of "people who say Japan is in deflation has decreased. The bank will continue with its aggressive monetary policy to support corporations' efforts, and, if necessary, will not hesitate in making adjustments."

Delivering multiple speeches on this ¥600 trillion NGDP level target suggests that Prime Minister Shinzo Abe is serious about upgrading Abenomics. It is reminiscent of the fireside chats in the 1930s where FDR was signaling his desire to raise the price level. In both cases leaders were trying to reflate their economies. FDR was partially successfully with reflating, but botched up on the supply side. Prime Minister Abe is trying to get both right with his monetary and structural reform arrows. Here is hoping he succeeds.

P.S. These developments do not bode well for Neo-Fisherism.

Update: Scott Sumner says Abenomics is working even better for the real economy and inflation than I portray in this post. On the latter, he suggests looking to the GDP deflator. It does paint a better picture.