1929 All Over Again?

Includes: DIA, QQQ, SPY
by: John Lounsbury

David Rosenberg is Chief Economist for Canada's Gluskin-Sheff, and former Chief North American Economist for Merrill Lynch. The graph below appeared in Rosenberg's investor advice newsletter Friday morning (July 31) here.

Others have also commented on the close parallel pattern, including both Doug Kass and Harry Schiller at Real Money (TheStreet.com) here.

I think many will agree with me that the economic resume for 2009 has major differences from 1929-30. For that reason alone there should be limited expectation of continuation of the remarked similarities to date in the above graph. I emphasized "limited expectation" because I do not want anyone to infer that I meant "rejection".

There are some significant parallels in the two time periods. Among these are the worldwide nature of the collapse, multigenerational deflationary pressures, massive insolvency issues in financial systems around the world, dramatic reduction in industrial production and drying up of international trade.

What Is Different Now?

There are significant differences, as well. There is massive sovereign interference with contraction now, orders of magnitude larger than anything conceivable in 1929-30. In the U.S., stimulus and government investment in the private sector is in the trillions of dollars when everything is added up. China is pouring its national reserves into stimulus and commodities on a scale relative to GDP that is several times what the U.S. and others are doing, Factors such as these will make the trajectory of this 80-year cyclical minimum much different from the last one. It is more likely to resemble the rolling recessions of the past 20 years in Japan than the 1930s of the U.S. Both had enduring pain, but the magnitude and depth of suffering is drastically different.

Rosenberg recognizes the possible correlation to the Japan experience. He points out that Japan has had a 20-year bear market with four large rallies with gains in excess of 30%.

My view is that the future will hold an experience that will probably have a mixture of flavors from many historic cuisines. I give low probabilities to a repeat of the Great Depression, as well as "V" shaped recovery to sustained 3%+ positive GDP growth. Something like the Japanese experience of the past 20 years has a higher probability.

But I think we will paint a new, original canvas. It may have recessions more frequently than experienced over the past 25 years. (We had four, counting the current one). It may have lower U.S. GDP growth potential, say closer to 2% rather than the 3%+ potential of the past two decades. It will probably see the G-12 move toward equality in economic power and many aspects of living standards with the G-8. (Note: I have divided the G-20 into two groups.)

What About Stock Valuations?

Rosenberg comments on the valuation of stocks with much skepticism. He says:

It is amazing that anyone would go long an equity market with a reported P/E multiple of 700x but that is indeed what we have on our hands. The end of the recession and the onset of a sustainable recovery, as we saw in 2002, are not the same thing. So this could still end badly but we will await confirmation signs that this is more than a very flashy bear market rally before shifting gears.

To understand how Rosenberg makes the statement about PE, you must recognize a key word - reported - as in "reported P/E". U.S. corporations keep two sets of books: One set is "reported earnings" upon which taxes are paid. These are the earnings that are stated according to GAAP (Generally Accepted Accounting Principles). Analysts like to use a second set of numbers reported by corporations, called "operating earnings". This allows corporations to exclude expenses that they designate as "non-recurring", such as acquisition costs, costs associated with payroll reductions (cost of firing people), etc.

Here is a table using S&P data (Standard & Poor's) with their latest results and estimates (as of 7/23) for "as reported" and "operating" earnings for the S&P 500 and the P/E ratios calculated by the author:

These numbers can be compared to recent actual results:

Before this crisis really got underway in the second half of 2007, there were usually not major differences between operating and reported earnings. Since then, they have often come from two different planets.

Analysts often like to use operating earnings, making the claim that they better represent ongoing business activities. However, when operating earnings differ substantially from reported earnings quarter after quarter, that argument must be questioned. Sooner or later, the two sets of books must come into close agreement or the accounting practices of corporations must receive closer examination. When do two sets of books with grossly different results become fraud?

This issue was discussed by David Pauly this week on Bloomberg.com, giving several examples of big impacts that analysts chose to omit when using the operating earnings for Intel (NASDAQ:INTC), Google (NASDAQ:GOOG) and others. Read the entire article here.

By the way, I don't show Rosenberg's stated 700x P/E ratio. His value may represent data from his own analysts (not the S&P) or may represent a different date than 7/23. For example, sometime in the first quarter the S&P estimates must have gone from negative P/E, through infinity and then on to a very large positive P/E.

In July, http://www.chartoftheday.com/ published the follow chart:

I do not know on what date the last data point was taken, or what the source of the data is, but I think this graph overstates the situation as I understand it today. This is definitely what the situation would have looked like in the first quarter of 2009, if the reported earnings at that time were projected for the entire year. Today I would expect the 2009 data point to be much higher, based on the current S&P estimates.

Shouldn't the Market Lead the End of the Recession?

There has been a lot of discussion regarding how the stock market responds to GDP. Rosenberg has another informative graph that relates to that question in his mid-day newsletter, also Friday, here:

Look at this graph in sections. In the first section there is a positive correlation between movement in stocks and GDP from 1965 to 1977, albeit with a delay, GDP lagging stock movement. For 1978 and 1979 there was a pronounced negative correlation, as was the case also from 1984 to 2000. From 1980 to 1983, a positive correlation existed. From 2001 to date we see a return of positive correlation.

In any of the time periods there were short periods of contrary relationships. For example, in 2001 to 2002, the stock market turned back down to a new low about nine months after GDP bottomed.

The Bottom Line

It is always interesting to compare current conditions to possibly similar past periods. I do it all the time and it often provides valuable insight. However, making many generalizations can lead one astray. I do not want to diminish the value of what David Rosenberg has to say. He is one of the best economic analysts in the world and I read his posts every day. I do encourage the reader to try to use his own brain as well as the thoughts of the writer and look for any value to be added.

Not only should we ask what could be repeated, we should also ask what things could be different this time.