The Hole in Jackson Hole

Fed Chairman Ben S. Bernanke gave his annual address at the Federal Reserve Bank of Kansas City Economic Symposium, Jackson Hole, Wyoming, August 31. We review the speech, Monetary Policy since the Onset of the Crisis, and suggest the Chairman has it wrong. As a consequence, the further policy intervention that he is signalling will likely not help the economy but will continue to distort markets making an inevitable market correction more severe. Call it building a Minsky moment.

He begins by describing the failure of using the Fed Funds Rate (EFFR) to stimulate the economy, leading to the current Zero Interest Rate Policy (ZIRP: see our discussion around Figure 6 in Living In a Hypothesis). He then goes on to discuss some of what we have learned, beginning with our experience conducting policy using the Federal Reserve’s balance sheet.

Key points (bolded for emphasis) on the use of the balance sheet include:

  1. The acquisition of Treasury and agency securities, which are the principal types of securities that the Federal Reserve is permitted to buy under the Federal Reserve Act.
  2. Then this insightful rationale for QE1: as investors rebalance their portfolios by replacing the MBS sold to the Federal Reserve with other assets [principally Treasuries], the prices of the assets they buy should rise and their yields decline as well. Declining yields and rising asset prices ease overall financial conditions and stimulate economic activity through channels similar to those for conventional monetary policy [the failed use of the EFFR].
  3. Large-scale asset purchases (LSAP), otherwise known as QE1 and QE2, were intended to put additional downward pressure on long-term interest rates.
  4. Operation Twist or the maturity extension program (MEP), was designed to operate by putting  additional downward pressure on longer-term interest rates and further eases overall financial conditions.
  5. Measures of the effectiveness of LSAP policies:

    • Generally, this research finds that the Federal Reserve’s large-scale purchases have significantly lowered long-term Treasury yields.
    • it is probably not a coincidence that the sustained recovery in U.S. equity prices began in March 2009, shortly after the FOMC’s decision to greatly expand securities purchases. This effect is potentially important because stock values affect both consumption and investment decisions [Zero Hedge has repeatedly documented the flow of funds out of equity mutual funds and into bond funds. In addition, volumes on the major exchanges this summer have been unusually light. Also see Technicals flash amber as ECB and Fed struggle to validate rhetoric). There is evidence that the consumer has left equities so this effect may be a tad overstated].
  6. And a key admission: how the economy would have performed in the absence of the Federal Reserve’s actions–cannot be directly observed. If we are willing to take as a working assumption that the effects of easier financial conditions on the economy are similar to those observed historically, then econometric models can be used to estimate the effects of LSAPs on the economy. This is the basis of the Fed’s econometric modeling. The fallacies in this assumption are the admission that with ZIRP, using tools they have no experience with, and failing to understand that this is a balance sheet as well as a business cycle recession means they have no historical referent to compare to! Their models cannot possibly work.
  7. Given the previous admission, he then states: a study using the Board’s FRB/US model of the economy found that, as of 2012, the first two rounds of LSAPs may have raised the level of output by almost 3 percent and increased private payroll employment by more than 2 million jobs, relative to what otherwise would have occurred.
  8. And his summary comment on LSAPs: Overall, however, a balanced reading of the evidence supports the conclusion that central bank securities purchases have provided meaningful support to the economic recovery while mitigating deflationary risks. At this point, keeping this statement in mind, we need to look at the data.

Mitigating Deflationary Risks: The Evidence

In the following sections, we use the FRED® data tool to create our graphs. To begin with we note that our observations are based on visual representations of data. As a result this may be good but not precise. Analytical tools available to researchers in the field would yield more precise results perhaps but we do not think the conclusions would be materially different.

Admittedly there are many definitions or ideas about inflation/deflation. The one that most commentators and people would think in terms of would be the Consumer Price Index (CPI). Since Bernanke was not more specific such as “deflation in credit markets”, we will assume he is talking about the general economy and hence CPI. We understand the Fed may prefer to look at Personal Consumption Expenditure (PCE) rather than CPI but the two are similar in performance as Figure 1 shows, and we wont debate their relative merits.

Figure 1. PCE versus CPI. (Click on image to open in a new window)


First, lets get an understanding of the behavior of CPI relative to the economy represented by recessions as in the following graph in Figure 2. Shown is the year over year (yoy) percentage change in CPI.

Figure 2. CPI showing yoy percentage change. (Click on image to open in a new window)

What we note is that 8 of the last 9 recessions were marked by a drop of at least 2.5% that started sometime during the recession. Although the drop during the recent recession of about 7.5% is deep, greater drops were seen in the late 1940s and early 1980s. The only exceptional point of the recent recession is the speed with which the CPI recovered half its loss. The recession of the late 1940s offers the only comparable recovery. We also note that since the early 1990s, the yoy change in CPI has been running at a mean estimated at about 2.5%.

In fact, the period from the early 1990s is similar to the postwar (WWII) period up to the mid 1960s, marked by higher volatility around the 2.5 grid-line on the graph. In short, we see nothing anomalous about recent CPI behavior based on the historical period covered.

Now lets zoom in on CPI from just before the start of the credit crisis in 2007. In Figure 3 we have added with some degree of approximation due to our graphics tool, vertical bars representing the period s when LSAPs were operational.

Figure 3. CPI and GDP from January 2007. Periods of QE have been indicated by orange bands. (Click on image to open in a new window)

Based on our historical observations we would expect a disinflationary drop from a peak inside the recession to a low followed by a recovery towards a mean after the recession. This in fact is what we see with the blue curve in Figure 3. Noting Bernanke’s comment that we cannot observe what the economy would have done without QE, we cannot on the other hand, readily distinguish what influence if any QE had on the economy. QE1 operated during the normal post-recession inflationary period. Given the size of the LSAP, we would expect some notable anomalous recovery but we do not see such. As noted, we cannot discern an observable effect in any case because the baseline that such would be measured against, an economy without QE operating, is rightly unobservable according to Bernanke.

Further, if LSAP is intended to be inflationary, then it might be argued that QE2 was in fact working since it was  in operation during a period of inflationary increase. But then look at the MEP (Operation Twist). It is currently operational in a disinflationary period.

As a final comment, we will later suggest a lag of 1-3 years between interest rate policy tools and employment. We would certainly expect a lag between LSAP and other aspects of the economy such as GDP and CPI. Certainly from Figure 3, we could deduce no lagging impact across four LSAP events.

Our conclusion is there is no observable effect of QE or LSAPs on inflation or of mitigating deflationary risks.

Support To the Economic Recovery: The Evidence

We will address this topic in two parts, first with regard to GDP and second with regard to employment/unemployment.

Figure 4. GDP yoy percentage change (blue line) and Effective Federal Funds Rate (EFFR, red line). (Click on image to open in a new window)

Since GDP is a factor used in determining recessionary periods, the correlation is not surprising. What is notable is that the last recession is the deepest postwar recession. Also, economic activity after recessions recovers to a downward trend-line quite nicely. There is no evidence that this trend will not continue implying continued slowing of economic growth. This would correlate with research that connects this downward trend with high levels of national debt (The Policy of Doom). The sharpness of the recovery and a recovery to a point higher than when the recession was entered is consistent with many of the post-war recessions. As a matter of fact, if one accepts the downward trend in GDP growth as a fixture of the US economic environment, then one could conclude that the recovery appears normal. At this scale we cannot conclude that LSAPs have any demonstrable effect.

We have overlaid the EFFR (red line) since this is the primary monetary tool that the Fed used until it broke  in 2009 (ZIRP) and was replaced by LSAPs.

To determine the possible effect of LSAPs, let us return to Figure 3 and look at the red curve in relation to the periods of LSAP. The recovery in GDP we would expect as the normal response of an economy after a recession. Since the period of QE1 corresponds to a period of GDP growth, if we want to argue that it was in part responsible for that growth then we should expect such growth to continue through QE2 and the MEP, particularly since interest rate effects tend to lead economic activity by some months. With or without a lag, if LSAP woks, then GDP should not level off.

We conclude then that there is no observable effect of QE or LSAPs on supporting GDP growth.

Employment Recovery

In Living In a Hypothesis, we looked at the unemployment rate and compared it to the size of the Fed’s balance sheet. We reproduce Figure 3 as Figure 5 in this essay.

Figure 5. Total unemployment number (blue line) and the total assets (size) on the Fed’s balance sheet (red l0ne). (Click on image to open in a new window)

Using the same techniques used in the arguments above we compared the recovery from the last recession to previous recoveries and could not discern an LSAP impact, especially given the size of the stimulus. Instead, let’s look at the annual percentage change in employment directly. Figure 6 gives the historical perspective.

Figure 6. Percentage change in total non-farm employment yoy. (Click on image to open in a new window)

Again we note that this recession registered the sharpest percentage drop in the number of employed since the 1940s. What is disturbing is that we can draw a downwards trend-line for recoveries. This reinforces the evidence that this has been a “jobless” recovery but suggests the improvement rate may have maxed out. From the early 1950s to the post-1980 recession period, the recovery rate for a recession reached a maximum of about 5%. In the 3 recessions since then, the recovery maximum has steadily declined. This parallels the decline in annual GDP growth. Using an argument similar to the one used earlier, the recovery shows no unexpected anomalies that might be attributable to LSAP operation.

Focusing in on the 2007 to date time period as in Figure 3, we get Figure 7.

Figure 7. Total non-farm employment. (Click on image to open in a new window)

As noted previously, we have no means of assessing a direct impact of LSAP programs on employment because we have no observable base line. The implementation of LSAP precludes establishing an exact one. To try and establish a comparative baseline one must look to past performance and extrapolate into the LSAP impacted environment in order to create any kind of a quantitative comparison through some academic model. Our use of past patterns of data from previous recessions to do a visual comparison and projection is as valid as any Fed modeling procedure. It simply forgoes quantitative estimation.

In our analysis above we have used visual comparison of the latest recovery to past recoveries and inferred that it exhibits a pattern closely comparable to the pattern of employment recovery associated with past recessions. From this we would infer that any effect from LSAP is not large if it exists at all. We realize that this last recession is different in that it is primarily balance sheet driven, a fact that renders it somewhat impervious to Fed tools. So comparing to past recessions may contain dangerous exceptions that we are not aware of but could mask LSAP activity. Still, conjecturing in this direction does not prove LSAP efficacy. There is simply not enough experience with these tools as Benanke notes, to allow either a determination of impact or a lagged effect.

The other comparison we have used is the behavior of the data during different LSAP applications. If one wishes to argue that the upturn in early 2010 is the result of QE1, then one might be tempted to infer that QE2 and the MEP are having a positive effect also. QE1 would suggest a lag of at least a year before employment turned around. From Figure 6 in Living In a Hypothesis we inferred a lag between changes in the EFFR (interest rate policy tool) and the unemployment numbers of about 3 years.

Finally we can look at employment recovery from the last recession and compare it to previous recessions. Figure 8 shows the last recession is the worst of all postwar recessions in terms of recovery begging the question, what has four rounds of LSAP produced?

Figure 8. Job recovery in post-WWII recessions compared from the recession onset. (Click on image to open in a new window)

(source: Political Calculations Worst. Recession. Jobs. Recovery. Ever. Updated.)

Of the Fed’s claim of two million jobs created we would note that key point #6 at the top gives a basic premise behind their models that is false or at least faulty. The actual numbers they put forward come from their models (key point #7) and not from observable data. We suggest that observations derived from observable data trump the output of Fed models.

We will say that from comparison with past recessions in Figures 6 and 8, this recovery is the weakest of the postwar era.  As we found for GDP and CPI, we don’t see the footprint of LSAP in the employment data. If in fact there is somehow a measurable economic effect from four LSAP programs to date, we don’t see it in any of the data we have reviewed. But if LSAP does have some small effect, then if the Fed wants to effect a large economic stimulus it will have to be a massive monetization program.

Reading the Tea Leaves: The Issue of Structural Change in Employment

Bernanke’s concern is:

The stagnation of the labor market in particular is a grave concern not only because of the enormous suffering and waste of human talent it entails, but also because persistently high levels of unemployment will wreak structural damage on our economy that could last for many years.

But here is where we read the tea leaves differently. In Bernanke’s words,

Some have taken the lack of progress as evidence that the financial crisis caused structural damage to the economy, rendering the current levels of unemployment impervious to additional monetary accommodation. The literature on this issue is extensive, and I cannot fully review it today. However, following every previous U.S. recession since World War II, the unemployment rate has returned close to its pre-recession level, and, although the recent recession was unusually deep, I see little evidence of substantial structural change in recent years.

We note that in Figure 6 we are looking at the rate of change in employment levels rather than the absolute levels. However, the peak rate of recovery after a recession has been trending downwards since the recessions of the early 1980s. That’s a 30 year span taking in four recessions. We suggest this has a structural implication.

Turning to Figure 8 we see a similar implication. The length of time to recover lost jobs for all postwar recessions has taken progressively longer for the last four, starting with the 1980s recessions. The disturbing feature of this chart is the current recovery appears to be leveling off (note in Bernanke’s admission of concern: The stagnation of the labor market) at about a 56% net loss of jobs. Even if we assume the overall rate of increase is to continue, it would take another 2-3 years to regain all jobs lost.

That we appear to have established a thirty year trend is to us suggestive of a structural change in the labour market and the economy. Since Mr. Bernanke does not share this view, he must then believe employment problems are cyclical and amenable to a good policy goose. And given his concern stated above, we expect he will act and act quickly if the numbers deteriorate. If the problems are structural and not cyclical as we believe, his policies will be ineffective. Time will tell.

Other Points from the Speech

Bernanke made a very strong statement: The FOMC has spent considerable effort planning and testing our exit strategy and will act decisively to execute it at the appropriate time. Whatever it is – and we have seen no inkling published – the Fed is ready to act deciseivly and confidently.

Could the Fed actually be trying to spot bubbles? Could they identify one if they tripped over it? What if in fact the very policies they believe in are creating a massive bubble? Would they know? Consider: We have seen little evidence thus far of unsafe buildups of risk or leverage, but we will continue both our careful oversight and the implementation of financial regulatory reforms aimed at reducing systemic risk.

And consider this point: Extensive analyses suggest that, from a purely fiscal perspective, the odds are strong that the Fed’s asset purchases will make money for the taxpayers, reducing the federal deficit and debt. As we have commented in private correspondence, any income from Treasury instruments is offset by the debt created to pay the interest. It’s a net wash with a net loss to the taxpayer on the operational administration of the transactions by the Fed. Admittedly, income from the private sector to pay interest on MBS securities is a net gain to the taxpayer.

The Fed is not overly concerned about gains or losses apparently: monetary policy can achieve the most for the country by focusing generally on improving economic performance rather than narrowly on possible gains or losses on the Federal Reserve’s balance sheet.


In his speech, Chairman Bernanke argues that the Fed sees positive economic effect as a result of LSAP as having raised the level of output by almost 3 percent and increased private payroll employment by more than 2 million jobs. We argue that there has been no observable effect on either disinflationary tendencies, GDP growth or job creation. We think the key aspect of our analysis is that GDP growth is on trend and the level of employment is appropriate to the current economy and should not be expected to improve much.

The fact that Bernanke thinks his tools work leads us us to believe he will use them again. That they will influence markets we have no doubt – this is an observable fact. That they will not impact the economy in any measurable or observable way we also have no doubt.

What is disturbing about Bernanke’s position is that the Fed has used massive and unprecedented tools with little and likely no effect on the economy. That fact that he believes he is right (not all Fed governors do) means he may be following the path of other authorities and administrations that at the time believed they were right. Weimar Germany and the Fed in the 1930s comes to mind.

So the Fed will enact further LSAP programs. They will have significant effect on markets and no effect on the economy. Place your bets accordingly.

Update 20120917

It is always gratifying to get confirmation of our thinking from people much smarter and informed than we are. Today we got this eletter from Bob Eisenbeis, Chief Monetary Economist of Cumberland Advisors: We’ll Know It When We See It! These people understand the Fed, Fed operation and policies, and debt and bond markets as well as anyone and better than most. Bob quoted Chairman Bernanke from the latest FOMC statement on the limits to the MBS purchases announced:

The key phrase here is in the Chairman’s line “… if we do not see substantial improvement in the outlook for the labor market, we will continue the MBS purchase program, undertake additional asset purchases, and employ our policy tools as appropriate until we do so.

We wrote David Kotok, Chairman & Chief Investment Officer of Cumberland Advisors:

Regarding Bernanke’s statement: “The key phrase here is in the Chairman’s line “… if we do not see substantial improvement in the outlook for the labor market, we will continue the MBS purchase program, undertake additional asset purchases, and employ our policy tools as appropriate until we do so.” Is this akin to “we will continue to flog the dead horse until it gets up and walks?” What if the change in employment is structural and not cyclical?

His gracious reply was:

It is structural. Will take a long time.

This confirms a lot of what we have been recently arguing.

Related Posts

Powered by WordPress | Designed by: photography charlottesville va | Thanks to ppc software, penny auction and larry goins