Folly of the Fed: Why Janet Yellen May Be a Dangerous Choice.

Andrew Smithers
PBS Newshour - The Business Desk, 15th November 2013

Janet Yellen, President Barack Obama's nominee to replace Ben Bernanke as the chair of the Fed, was on the hot seat Thursday in front of the Senate Banking Committee for her confirmation hearing. British economist Andrew Smithers questions the monetary policy she's supported as vice chair of the Fed and defended at her confirmation hearing. You can watch the hearing here.

Smithers, author of "The Road to Recovery: How and Why Economic Policy Must Change," has appeared on the Business Desk before, most recently in this interview with our resident econo-crooner Merle Hazard (otherwise known as money manager Jon Shayne) and in this earlier interview with Jon. (Don't miss Merle's latest music video about the threat of the "Great Unwind" of the Fed's asset-purchasing policy.)

Are we on a sugar high, as Sen. Mike Johanns, R-Neb., suggested at Thursday's confirmation hearing, critiquing the same asset-buying policy? Here, Smithers echoes those concerns about a bubble and he explains his objections to Yellen's defense of quantitative easing. What would Yellen have to admit for Smithers to retract his criticism?

The pending appointment of Janet Yellen as chair of the Federal Reserve Board of Governors gives me great cause for concern. By all reports, she is highly intelligent and, in the opinion of her peers, highly qualified for her new role.

However, we've seen errors in monetary easing set off a bubble in asset prices and the worst recession since the 1930s, proving that at least one other quality is essential for the job she is about to undertake: the intellectual integrity essential for sound judgement.

The three most marked breaks in the long-term growth of developed economies during the past century occurred in 1929, in 1990 in Japan and most recently in 2008. Each instance was marked by extremely high levels of private sector debt and was triggered by large falls in asset prices. In Japan in 1990 and in the U.S. in 2008, both housing prices and equities fell sharply, although in 1929, the collapse was limited to the stock market.

Because the fall in housing prices has received particular attention after the recent recession, it is very important to note that houses are not the only asset class where price falls can trigger damaging recessions.

Debt levels have barely fallen since 2008, notably in business, where massive buy-backs of shares, running at around 3 percent of GDP per year, have reduced corporate equity and thus kept up leverage, which shows that it is nonsense to claim, as many have, that the low level of U.S. business investment is driven by a desire by companies to deleverage.

I and many others are therefore concerned that the Fed's policy of quantitative easing (QE), in which they currently purchase $85 billion a month of Treasury securities and mortgage bonds, has driven up asset prices, in both the bond and equity markets, to dangerous levels.

Yellen could legitimately defend quantitative easing in a number of ways. She could, for example, argue that the risks of high asset prices are counterweighed by QE's intended benefits, including the prevention of deflation. But, instead, she's defended the program with meritless arguments to claim that equity prices are not at dangerous levels.

The problem is the metric she uses to value equity prices. The metric she uses is a common one: the ratio of stock prices of companies overall to their overall earnings -- the traditional price/earnings (PE) ratio. The PE ratio is based, reasonably, on the premise that when you buy a share of any firm, you're essentially buying a stake in its profits, which will either be given to you directly in the form of dividends or invested on your behalf by the company.

A question, though: what measure of profits to use? Yellen is on record claiming that the PE multiple based on assumptions about next year's earnings can be used to show that the market is not overpriced. Here's what she said in 2011:

"Overall ... indicators do not obviously point to significant excesses or imbalances in the United States. ... forward price-to-earnings ratios in the stock market fall within the ranges prevailing in recent decades, and are well below the early-2000 peak."

This is nonsense. On several occasions in the past, the market, based either on the past 12 months' or the next 12 months' earnings, has appeared to be reasonably priced, despite being demonstrably and dangerously expensive based on more reliable measures.

Those more reliable measures include q, which is the ratio of stock market value to real-world replacement value. (It is similar, but not identical, to James Tobin's q, as I explain here.) Another more reliable measure is the Cyclically-Adjusted Price-Earnings ratio or "CAPE," which compares the value of the stock market not to just one year's worth of earnings, but rather to the inflation-adjusted average of earnings for the past 10 years.

These inflated markets have been followed by unpleasant falls in the stock market and recessions.

And on the flip side, there have also been times when the market looked expensive based on the 12-month PE, but was demonstrably cheap using the much more reliable q or CAPE.

The point is that in order to know whether a market is expensive or cheap, we need many years of data. Cheap markets often fall in the short-term and expensive markets often rise. We cannot therefore judge whether a stock market was cheap or expensive simply by looking back at the following year's change in the share price. Similarly, we cannot even judge from looking at the return over the subsequent decade, say, because the market 10 years from the starting date may have been heavily depressed (as during the Great Depression) or massively over-priced (as in the Roarin' '20s or the '90s).

We can, however, largely avoid this problem by calculating the average return over many years. By calculating the average return for each subsequent year over 30 years, we can rank markets. Those which gave the lowest return can be deemed the most expensive and vice versa. By comparing these returns with the average long-term return we can show whether a given market was cheap or expensive and by how much.

By deducting from the average long-term real (inflation-adjusted) return, which from the end of 1900 to the end of 2012 was just over 6 percent, the average return for each year over the subsequent 1 to 30 years, we can show which years were expensive and which were cheap. (Because this takes 30 years of subsequent data, we can do this only up to 1982.)

An outstanding example of the misinformation about value given by PEs based on next year's earnings per share is 1916, which on basis of earnings, appears to have been the second cheapest market on record but was the third most expensive when judged by hindsight, and was followed by a severe recession.

The extent to which different criteria of value give accurate predictions about future returns can be judged by the correlation coefficient between hindsight value and the chosen criterion. I show these in Table 1 and it will be seen that next year's earnings-per-share is an even worse guide than this year's, and both are worse than the dividend yield let alone CAPE and q.

The fact that Janet Yellen chose the criterion most likely among this group to give a misleading signal raises the issue of intellectual integrity. Possible explanations are ignorance and the wish to defend the quantitative easing policy that has taken place on her watch. Ignorance about these other metrics is unlikely. Janet Yellen is married to Nobel Laureate George Akerlof who has co-authored economic papers with the newest Nobel Laureate, Bob Shiller, who has been the most prominent proponent of the use of CAPE to evaluate whether the current level of the stock market is out of line historically. Ignorance, even if true, does not strike me as a valid excuse.

Those who use their position as (then) deputy chair of the Fed surely have a duty to study the literature on a subject before pronouncing conclusions which have major policy implications.

Where we have the actual data, ratios based on prospective data have a worse correlation with hindsight value than PEs based on the earnings-per-share for the past 12 months and considerably worse than the dividend yield.

Using prospective earnings today is particularly suspect because no one knows what they will be; such forecasts are habitually overstated and they can be wildly wrong. Why were they used? Even the least cynical must suspect that they were used not because of their virtues but because of their defects.

Had the emphasis been placed on q, Yellen should have been worrying that the market on Wednesday, Nov. 13, (S&P 500 at 1782 ) was 68 percent overpriced according to q and over 80 percent according to CAPE. (It was also over 100 percent overpriced according to the dividend yield and while, admittedly, we know the dividend yield is a poor guide to value, it has at least better claims than the PE based on next year's assumed earnings per share.)

I hope, but do not expect, that Yellen apologizes for the error that she has committed. If she does this, my accusation of a lack of intellectual integrity will be promptly and happily withdrawn. It would provide a welcome change to other comments from central bankers who seem to think they can atone for their past mistakes by obstinately adhering to them.

This article was originally published on PBS Newshour - The Business Desk