Monday, April 21, 2014

Bernanke vs. Friedman: Financial Intermediation Shock or Medium of Exchange Shock?

In preparing for a talk, I reread Jeffrey Rodgers Hummel's article comparing Ben Bernanke to Milton Friedman. It was good to look over it again. Hummel makes the case that Bernanke saw the crisis as a financial intermediation crisis where Friedman would have viewed it as a medium of exchange crisis. These two different perspectives imply different policy prescriptions. The Bernanke Fed focused on saving the financial system at any expense, including creating distortions in the credit market. The first phase of the Bernanke Fed response, for example, was to manipulate (but not expand) the asset side of its balance as a way to support distressed financial firms. This phase began in the fall of 2007. The Fed eventually did began to expand its balance sheet with QE1 in late 2008, but even then the emphasis was on 'credit easing' not stabilizing total dollar spending. The Fed's later QE programs, though better, also lacked the full commitment and predictability needed to fully restore total dollar spending to its full employment level.

The Friedman approach would have been to avoid picking winners and losers in the financial system and instead commit to adding as much liquidity as needed to maintain a stable level of total dollar spending. This medium of exchange approach would have let some banks fail while preventing a wholesale collapse in aggregate demand. Recall that the widespread bank run on the shadow banking system was not inevitable. As Hummel reminds us, we had severe banking crisis in the 1920s and in the 1980s with the S&Ls and yet there were no sharp economic downturns. The same could have happened in 2008-2009. One can look to how Australia fared over the past five years to see how things could have been done differently. The key to this approach is credibly committing to maintaining the level of nominal GDP in a predictable, rule-based fashion. From this perspective, the worst part of the financial crisis in 2008-2009 was the consequence of the Fed failing to stabilize expected path of total dollar spending.

In short, Bernanke viewed the crisis as result of a large shock to financial intermediation while Friedman would have seen it as a large shock to the medium of exchange. This echoes Bernanke and Friedman's differing views on the Great Depression. The former saw the 1930's banking collapse as the key catalyst behind the Great Depression while the latter saw the Fed's failure to stabilize expected total dollar spending growth as the reason for the banking collapse and resulting fall out in the broader economy. 

Read the Hummel's paper for more on this distinction.

Friday, April 18, 2014

This One Figure Shows Why Fed Policy Failed

What do you see in this figure from a recent New York Fed study

Technically, it shows that the Fed's balance sheet is expected to shrink and return to the path it would have been on had there been no large scale asset purchases (LSAPs) over the past five years. This projection is a reflection of FOMC's plan to eventually normalize the size of the Fed's balance sheet. Bonds markets have understood this plan from the beginning as is evidenced in their inflation forecasts. The FOMC formally announced its plan to normalize and shrink its balance in its June, 2011 meeting. Subsequent speeches, press conferences, and congressional testimony by Ben Bernanke have reinforced this understanding. The point is the Fed never intended the LSAPs to be permanent.

So again I ask, what do you see in the above figure? What is the bigger message it is telling? In my view, the answer to these questions is unambiguously clear. It signals the Fed never intended to unload both barrels of the gun and fully offset the collapse in aggregate demand. In other words, the figure reveals why Fed policy failed to end the slump.

This is a strong claim I am making. You may be scratching your heard wondering how I got that message out of the figure. Here is my explanation, based on previous posts. First, in order for there to have been sufficient aggregate demand growth the Fed needed to commit to a permanent monetary injection:
[O]pen market operations (OMOs) and helicopter drops will only spur aggregate demand growth if they are expected to be permanent. This idea is not original to Market Monetarists and has been made by others including Paul Krugman, Michael Woodford, and  Alan Auerback and Maurice Obstfeld. Market Monetarists have been advocating a NGDP level target (NGDPLT) over the past five years for this very reason. It implies a commitment to permanently increase the monetary base, if needed.
Doing so would help in the following ways:
The key reason is that it would create an expectation that some portion of the monetary base growth from the asset purchases would be permanent (and non-sterilized by IOER). That, in turn, would mean a permanently higher price level and nominal income in the future. Such knowledge would cause current investors to rebalance their portfolios away from highly liquid, low-yielding assets towards less liquid, higher yielding assets. The portfolio rebalancing, in turn, would raise asset prices, lower risk premiums, increase financial intermediation, spur more investment spending, and ultimately catalyze a robust recovery in aggregate demand. (One could also tell a New Keynesian story where the higher future price level implies a temporary bout of higher-than-normal inflation that would lower real interest rates down to their market clearing level.)
The key to the above story is that some portion of the monetary base expansion is expected to be permanent. If the public believes the Fed's asset purchases are not going to be permanent and therefore the price level and nominal income will not be permanently higher, the rebalancing will not take place. I bring this up because this same point applies to helicopter drops or any other kind of fiscal policy stimulus.
Now to be clear, all that is needed is a commitment to permanently expand the monetary base if needed. But such a commitment, if credible, would most likely raise the velocity of the monetary base. In this case, the needed permanent monetary injection would be smaller. In other words, the figure above attests to the failure of the Fed policy not only because it shows a temporary increase in the monetary base, but also because it shows such a large increase in the monetary base. Had the Fed credibly committed from the start there never would have been the need for all the subsequent LSAPs. The Reserve Bank of Australia did just this and its economy was one of the few not hit by the prolonged economic slump. 

Another way to see how the above figure reveals monetary policy failure is to compare the Fed to the Bank of Japan. It too tried temporary monetary injections but then switched to permanent ones under Abenomics:
[T]he Bernanke Fed never tried Abenomics. That is, for all the Fed has done over the past five years it never tried to do the kind of monetary regime change now being done by the Bank of Japan. A year ago, Japanese monetary authorities shook things up by credibly committing to permanently raising the nominal size of the Japanese economy. The evidence so far shows this program to be a smashing success. From the start, the Bernanke Fed took the opposite approach. It credibly committed to a temporary expansion of its balance sheet. The U.S. monetary injections, therefore, were never intended to be permanent and this makes all difference in the efficacy of monetary policy.
So for all the praise the Bernanke Fed gets for preventing the second Great Depression, it should be equally noted that it allowed the long slump. By failing to do Abenomics for the U.S. economy, the Bernanke Fed effectively kept monetary policy tight for the past five years. There is no other way to say it.

Okay, maybe there is another way to say it. The Bernanke Fed failed to meaningfully address the endogenous fall in the money supply and the decrease in money velocity. The Bernanke Fed could have done an American version of Abenomics, like nominal GDP level targeting, that would have arrested these developments. Instead, it did not and passively allowed total dollar spending to remain depressed. This failure to act is no different than an explicit tightening of monetary policy in terms of damage done to the economy. The only difference is that the public is more aware of the explicit form.
So again I ask, what do you see in the above figure?

P.S. Yes, the Fed's ad-hoc decision making contributed to this failure.

Monday, April 7, 2014

Financial Stability Concerns Are Overblown

Michael Darda of MKM Partners makes the case that concerns over financial stability are overblown and miss the forest for the trees. Consequently, he worries that the Fed may be pulling back too soon. 
[I]t is not at all clear that 1) risk premiums are too low, 2) leverage is problematic and that 3) “easy money” leads to “bubbles”. However, unemployment, underemployment and long-term unemployment remain high while inflation, if anything, is too low (and well below the Fed’s target). Thus, there would seem to be inordinate risks to the Fed pulling back sooner than would otherwise be the case due to potentially faulty “financial stability” concerns.
One way of framing this discussion is that a necessary (but not sufficient) condition for monetary policy to contribute to the buildup of financial imbalances is for the expected path of the federal funds rate to be below its natural rate level. Most evidence indicates that this has not been the case over the past few years. Rather, it is more likely the natural rate has been negative and below the actual interest rate for some time. If so, the Fed has been contributing to financial instability by being too tight, not too loose.

Below are the slides where Michael Darda makes his case. Take a look (link):

Sunday, March 30, 2014

Should We Be Worried?

A number of Fed watchers see the last FOMC meeting signalling a tightening of monetary policy.  Tim Duy, for example, sees it in the FOMC's effective lowering of its inflation target via the dropping of the Evans Rule:
[A] hawkish interpretation... starts with the end of the Evans rule. Everyone seems focused on the unemployment part of the Evans rule, while my attention is on the inflation part. The Evans rule allowed for the Fed to reach their inflation target from above. It provided wiggle room on the target as long as unemployment was above 6.5%. With the end of the Evans rule, the Fed sends a signal that they no longer find it acceptable to reach the target from above. They intend to reach it from below. 2% is officially once again a ceiling.
Others see it in the moving forward on the calendar of the first federal funds rate hike as seen in the FOMC projections or in the shifting of the FOMC conversation to when the policy rate hike will occur. Here is Ylan Q. Mui:
The nation’s central bank said Wednesday it will look at a broad swath of indicators... as it determines when to raise rates for the first time since the recession hit. The deliberately vague wording is a retreat from the Fed’s concrete promise to leave rates untouched. Though they disagree on when to act... the statement signals the moment has finally come within striking distance.
Gavyn Davies sees these developments as part of a longer tightening cycle that has been going on for some time:
[I]n a wider sense there has been an unmistakable shift in the FOMC’s centre of gravity in the past few months. The key to this shift is that the mainstream doves who have dominated policy decisions in the past few years have now essentially stopped arguing against either the tapering of the balance sheet or the start of rate hikes within about a year from now.
I found this interesting because on one hand we would expect some pulling back of monetary policy as the economy recovers. In this case, we would not call it tightening, but adjusting the stance of monetary policy to its new neutral position. For example, as the the economy recovers the market-clearing or 'natural' interest rate will rise. The Fed could be simply indicating its plans to match the expected rise in the natural interest rate. Doing so would not be tightening. On the other hand, it is possible the Fed is getting ahead of itself and pulling back too fast. In this case, the Fed would be rising its target policy rate higher than the expected natural interest rate. If so, this would begin choking off the the recovery.

So which one is it? Over the past few months I have been leaning toward the former view. I believed the Fed was not tightening policy, but simply adjusting it to match the recovery. Lately, though, I am not as certain. What gave me pause were the two figures below. The first one shows the bond market forecast of average annual inflation over the next years:

The figure shows that since Bernanke's initial tapering talk the inflation forecasts have fallen from 2.3% to about 1.8%. That is a sustained 50 basis point drop. I do not see any evidence in the productivity numbers to indicate the expected inflation decline is from productivity gains, so it leaves me thinking that the bond market sees lower nominal spending ahead. If so, Gavyn Davies is right that the recent FOMC meeting is part of a larger tightening cycle.

The second figure shows the year-on-year growth rate of the Divisa M4 money supply. This is a measure of money that includes both retail and institutional money assets. It also accounts for the different degrees of liquidity of the assets in its construction. This series started declining before Bernanke's taper talk, but has continued to be on a downward trend ever since. If it continues we could be looking at a burgeoning excess money demand problem.

One interpretation could be that the monetary policy tightening implied by the breakeven graph above is leading to less inside money creation as the economic outlook gradually weakens. The March FOMC meeting would be a continuation of this tightening cycle. The problem with this view is that I have a hard time finding other indicators that point to economic weakening ahead. Yes, one can point to evidence that there is still slack in the economy, but it is harder to find evidence that the slack is increasing. Maybe bond markets are simply pricing the 2% upper ceiling that has been in the FOMC central tendency forecasts. Also, the M4 growth decline could be due to regulatory contraction in institutional money assets. These charts may be red herrings, but they do give me pause. These type of declines in the past signaled future economic weakness. Should we be worried now?

Friday, March 28, 2014

Market Monetarism and Endogenous Money?

Yesterday, I was interviewed by  Erin Ade for the show Boom Bust. She asked me, among things, whether I believed money was endogenously created. If so, was my belief consistent with Market Monetarism? My answer was that inside money creation--money created by banks and other financial firms--is endogenous, but the Fed shapes in an important way the macroeconomic environment in which money gets created. Consequently, the Fed influences the creation of inside money. So yes, I believe endogenous money is consistent with Market Monetarist views.

To further unpack this idea, I want bring up a point I have repeatedly made here: open market operations (OMOs) and helicopter drops will only spur aggregate demand growth if they are expected to be permanent.1 This idea is not original to Market Monetarists and has been made by others including Paul Krugman, Michael Woodford, and  Alan Auerback and Maurice Obstfeld. Market Monetarists have been advocating a NGDP level target (NGDPLT) over the past five years for this very reason. It implies a commitment to permanently increase the monetary base, if needed.2

So why exactly are permanent injections so important and how does this relate to endogenous money creation? From a monetarist perspective, the permanent expansion of the monetary base will lead to permanently higher nominal incomes in the future. Given there is a negative output gap, the expectation of higher future nominal income from such an injection should also create expectations of higher real economic growth. This belief should lead households and firms to increase their spending today. In the process, asset prices rise, risk premiums fall, and financial intermediation increases. From a New Keynesian perspective, the higher future price level implied by the permanent injection would result a temporary bout of higher-than-normal inflation that would lower real interest rates down to their market clearing level. Once that happened the increased spending, lower risk premiums, and increased financial intermediation would occur.

Now let's expand on that last point. The permanent monetary base increase will lead to increased financial intermedation. For example, banks will start providing more loans as the improved economic outlook makes households and firms appear as better credit risks. Likewise households and firms will start demanding more credit. All of this leads to the creation of financial firm liabilities that function as money. In short, a permanent increase in the monetary base will lead to more inside money creation.

Below is a figure that tries to reflect this story. It shows the central bank (CB) doing a permanent monetary base injection that affects the expected path of monetary policy. These injections are exogenously determined by the central bank as it decides where it wants the economy to go in terms of inflation, the output gap, or in my ideal world NGDP. These exogenous changes in the path of monetary policy alter economic expectations and therefore shape how inside money is endogenously created.

Now some may object that during normal times the monetary base is endogenously determined for interest rate-targeting central banks. After all, for a given interest rate target central banks will accommodate changes in the demand for reserves. This is true in the short run, but not far beyond that. As noted above, central banks ultimately care about inflation and output gaps and consequently will adjust interest rates over time to hit their inflation and output gap objectives. Such changes in interest rates mean changes in the supply of bank reserves. Stated differently, it implies a different degree of policy accommodation to changes in the demand for bank reserves. These changes, then, mean the central bank is exogenously changing the path of the monetary base.

Let me illustrates this point with two extreme cases. Consider the implicit Taylor rule for the United States during the Great Inflation of the 1970s and the Taylor Rule for the ECB over the past few years. The parameters on the inflation term (which measure to degree to which policymakers respond to changes in inflation) were very different. For the United States the parameter was very low and at the ECB it has been very high. These different inflation parameters were exogenous policy choices and determined the very different paths of the monetary base for these two economies. So even with an interest rate-targeting regime, the path of the monetary base is ultimately determined by the central bank.

Now lets return to the original question posed by Erin Ade. Market Monetarism, in my view, is consistent with inside money creation being endogenous. Because of this understanding, we believe the Fed should create a macroeconomic environment that is conducive to financial firms creating the optimal amount of money. We believe a NGDPLT does just that since it is stabilizing, by definition, the product of the money supply and money velocity.

Update: This statement is not quite right: "These changes, then, mean the central bank is exogenously changing the path of the monetary base." The central bank does independently change the path of the monetary base as it responds to changes in the economy, but this path change is itself endogenous to its central bank's ultimate target. All variables other than the target variable are ultimately endogenously determined. The exogenously chosen nominal target variable, however, does constrain the endogeneity of all other nominal variables. Or, as Francis Coppola nicely put it, "Endogeneity itself is exogenously constrained."

P.S. Erin Ade interviewed Scott Sumner too on the program. Here is the link for the full program.

1Also, it is assumed the monetary base injection will not be sterilized by further increases in IOR.
2It is likely that a NGDPLT, through its influence on expectations, will raises the velocity of the monetary base. In this case, the permanent increase in the monetary base will be small. Nonetheless, it is the threat to permanently change the monetary base as much as needed that is the key catalyst here.

Thursday, March 27, 2014

Ad Hoc Monetary Policy

One of the defining features of U.S. monetary policy over the past five years has been its incredibly ad hoc nature. Over this time, the FOMC has conducted monetary policy with a spate of make-it-up-as-we-go-along programs (QE1, QE2, Operation Twist, QE3, and the Evans Rule) that it hoped would spur a robust recovery. These programs did get progressively better as they became more state dependent, but they were often implemented and ended in a haphazard fashion. This stop-go approach to monetary policy was politically costly and prevented the Fed from fully utilizing its ability to manage expectations of future nominal growth. 

A great example of the Fed's ad hoc management of monetary policy is the tapering of QE3. Former Fed chair Bernanke announced it in the Spring of 2013, but the Fed kept the markets guessing for almost nine months as to when it would actually begin. And once it got started, some FOMC members were still uncertain about how much to do. Consider also the Evans Rule. It was implemented to firmly shape expectations about the future path of the federal funds rate. It was, in other words, created to increase certainty. However, when the unemployment threshold in the Evans Rule was no longer was consistent with FOMC preferences, the FOMC simply dropped it. So much for certainty. With decision making like this, FOMC officials themselves are uncertain as to what they will be deciding at the next meeting.

Ramesh Ponnuru observes this same erratic behavior and concludes that is the reason for Fed's communication problems:
the [Fed's] muddled communications aren't a gaffe. They reflect a muddled policy. The markets are obsessed with every syllable Yellen utters because they're so unsure about what the Fed is going to do. Its behavior is difficult to predict. It has acted in an ad hoc way for the past several years and has never bound itself to any rule.


What the Fed instead said last week is that its policies "will take into account a wide range of information, including measures of labor market conditions, indicators of inflation pressures and inflation expectations, and readings on financial developments." There is nothing here that constrains the Fed's decisions or offers much guidance to those seeking to predict them. The statement might as well have added "the price of gold" or "the gut feelings of the Fed's open market committee" to its list of indicators.
Unfortunately, the muddled communication did not end with the Bernanke Fed. In Janet Yellen's first meeting as Fed Chair this month, she was dismissive of the interest rate projections shown in the FOMC's dot-plot figures. These figures showed that the FOMC's consensus view on the first interest rate hikes have moved up on the calendar. The FOMC, therefore, was predicting it would tighten monetary policy sooner than previously expected. In response to a question about this tightening, Yellen told reporters not to take the dot-plots too seriously. Tim Duy did not like this response because of the further confusion it creates:
I think it is absolutely ludicrous that the Fed is trying to claim the dots have no value.  Seriously, can they work any harder to raise the act of bungling their communications strategy to an art form?  If the dots have no value, then why force feed this information to market participants in the first place?
Tim Duy is right. The dot-plots were suppose to improve clarity about the future path of monetary policy. To claim now the dot-plots really do not show the future path only adds more noise to an already weak signal about where monetary policy is heading. Tim Duy notes elsewhere that breakevens--the expected inflation implied from the spread between nominal and real treasury yields--have become more volatile since the crisis. This can be seen in his figure below. Arguably, this too is a consequence of the increased uncertainty surrounding Fed policymaking.

Now to be clear, the point of this post is not that the Fed failed to correctly forecast the future economy. Rather, it is that the Fed failed to clearly spell out how it would systematically respond to differing states of the future economy. For the Fed to truly manage expectations (and therefore fully exploit its influence over economic activity) the public should know with some certainty how the Fed will respond to different economic developments. Over the past five years this certainty has been largely lacking. (The only thing that does seem certain over this period is that Fed wanted core inflation to fall somewhere between 1 and 2 percent. And even this understanding only became apparent in hindsight.)

Now imagine the Fed's monetary stimulus programs during this time had be done in a more systematic and predictable manner.  For example,  assume the Fed had announced a NGDP level target from the start and said asset purchases will continue until a certain level target was hit. There would have been no need to announce up front the large dollar sizes of the asset purchases that attracts so much political criticism. There would also have been no need to announce successive rounds of QE that make it appear the previous rounds did not work.  More importantly, this approach would have more firmly shaped nominal spending and income expectations in a manner conducive to economic recovery. In other words, there was a much easier and more efficient way for the Fed to respond the crisis. FOMC meetings would have been more predictable and consequently less important. We would not be hanging on the every word of our Fed chairs. Fed watchers and bloggers would be far fewer.

It is true that implementing something like a NGDP level target would have used up a lot of the Fed's political capital. Some will conclude, then, that this reality could never have happened. My reply is that it may have politically cheaper for the Fed to do a NGDP level target than it was to do all the impromptu programs it adopted over the past five years. But this is just Monday-morning quarterbacking and not the main point of my post. The key takeaway is that the right-kind of systematic monetary policy would have been far better in the crisis and would be far better moving forward. Instead, it seems that Fed policy is actually going the other direction. It continues to be ad hoc and unsure of itself. This does not bode well for future economic crisis.

Update: Philadelphia Fed President Charles Plosser makes a similar argument in a recent speech. He notes that forward guidance would be more effective if monetary policy was more systematic and followed some kind of rule. My implicit point above is that a NGDP level target would be just such a rule, as I have argued elsewhere.

Friday, March 21, 2014

Abenomics at the Brookings Institution

Joshua K. Hausman and Johannes F. Wieland presented a paper today on Abenomics at the Brookings Panel on Economic Activity. They focus specifically on the monetary policy portion of Abenomics: the Bank of Japan's new commitment to 2% inflation, open-ended asset purchases, and a doubling of the monetary base. Abenomics is frequently covered on this blog and these authors reach similar conclusions:
We show that Abenomics ended deflation in 2013 and raised long-run inflation expectations. Our estimates suggest that Abenomics also raised 2013 output growth by 0.9 to 1.7 percentage points. Monetary policy alone accounted for up to a percentage point of growth, largely through positive effects on consumption.
They see Abenomics' success coming from its commitment to a new monetary regime, one that credibly moves Japan away from its deflationary past. Like others, they compare Abenomics to FDR's regime change in 1933, but acknowledge it has not proportionally packed as much of a punch. The authors attribute this difference to the relatively small size of Abenomics as well as the Bank of Japan lacking full credibility on its new inflation target. They see the lack of full credibility as a consequence of its monetary history and demographics (i.e. pensioners hate inflation).

This  paper largely fits my priors. However, I do want to comment on a key difference in our views. The authors take a New Keynesian view on how Abenomics works. They see it working by raising expected inflation and lowering the real interest rate. My view is that Abenomics works because it has committed to a permanent expansion of the monetary base which, in turn, has implications for the future price level and future nominal incomes. Here is how I recently made this point in the context of a NGDP level target operating in a slump:
[NGDP level targeting] would create an expectation that some portion of the monetary base growth from the asset purchases would be permanent (and non-sterilized by IOER). That, in turn, would mean a permanently higher price level and nominal income in the future. Such knowledge would cause current investors to rebalance their portfolios away from highly liquid, low-yielding assets towards less liquid, higher yielding assets. The portfolio rebalancing, in turn, would raise asset prices, lower risk premiums, increase financial intermediation, spur more investment spending, and ultimately catalyze a robust recovery in aggregate demand.
While our views may be complimentary, I do think something is lost by focusing too narrowly on the New Keynesian channel. The portfolio rebalancing process I outline above is effectively the same thing as a reduction in the excess demand for money. And excess money demand--broadly defined to include both retail and institutional money assets--in my view is the deeper reason for the slump of the past five years. A permanent, unsterilized injection of the monetary base would have gone a long way in solving this problem. 
The Fed has always said its asset purchases are temporary and the public has bought into this view. Consequently, the Fed's QE programs have not been as effective as they otherwise could be. Here is a figure from an earlier post that makes this point:

Although Hausman and Wieland's credibility discussion alludes to this issue, I wish they had directly discussed the issue of a permanent versus temporary increase in the monetary base . Michael Woodford has stressed this point too. So has Alan Auerback and Maurice Obstfeldt. They formally demonstrate the importance of a permanent monetary base increase in great AER paper that has unfortunately receive far too little attention in recent debates. Here is an excerpt (my bold):
Prevalent thinking about liquidity traps, however, suggests that the perfect substitutability of money and bonds at a zero short-term nominal interest rate renders open-market operations in- effective as a stabilization tool...Yet, our analysis shows... that credibly permanent open- market operations will be beneficial as a stabilization tool as well, even when the economy is expected to remain mired in a liquidity trap for some time. That is, under the same conditions on interest rates that make open- market operations attractive for fiscal purposes, a monetary expansion that markets perceive to be permanent will affect prices and, in the absence of fully flexible prices, output as well...Our analysis suggests that Japanese policymakers should underscore the permanence of past operations, perhaps through an announced inflation target range including positive rates, and may need to increase the monetary base even more.
That last sentence is striking. It is call for Abenomics almost a decade before it was adopted. It would have been nice to seen a discussion of it in the Hausman and Wieland paper. Still, the paper overall is a good read and worth your time. Give it a look.

Addendum: Milton Friedman would probably be happy with Abenomics.