The long term economic growth rate has been weakening for decades and has continued to plummet since the great recession. It is at 60 or 70 year lows with the growth rate still pointing down. There are at least two important questions. Does the trend continue? If so, when does it again send stocks down? On the first question there is empirical support suggesting the down trend continues with the climax of the crisis perhaps coming in 2014 or 2015. The timing of the next major bear market has eluded me, but should come soon. Perhaps the high was already put in August 2nd.
As I see it the Fed has the pedal to the metal and the engine is cranking out rapid monetary base growth. There is perhaps a transmission problem where the wheels representing money growth are turning at just a normal rate. The forward motion or economic growth is only about a third the normal pace. Worst of all, the car is fish tailing with S&P 500 earnings swinging wildly, perhaps about to spin out of control.
The money supply as measured by M2 (the dark blue line above) has been growing at slightly above its long term pace of about 5.5%, but is well within the normal range and growth has backed off in the last few months. The monetary base (light blue) has grown at 24.8% the last 6 years which is 4 or 5 times its normal pace.
Growth on the other hand has been slowing down. GDP has annualized 2.7% over the last 28 years. This is part of a multi-decade decline in the US growth rate.
The first point on the chart above, plotted at 1975 shows annualized growth of 3.7% for the 28 years from the first quarter of 1947 through the first quarter of 1975. So the decline in this long term growth rate should be independent of influence from W.W. II. The last point showing 2.7% growth is for the 28 year period ending in the 2nd quarter of 2013. Looking back further may also add some perspective.
From this perspective growth has been slowing for half a century. If we look at rolling 7 year periods the recent slowdown looks quite stark.
The notion that we are in a 2% to 3% growth environment may just be an artifact of nostalgia and looking at growth on a quarterly basis. Looking at growth over 7 year periods eliminates much of the fluctuation that makes such a misconception possible. The trend in the 7 year growth rate is down to 1.1%. There is some basis to believe growth is picking up since quarterly growth has sequentially increased the last two quarters. If this were true we would expect continued strength in the 3rd quarter. However, this quarter appears to be off to a very weak start. In July growth in industrial production, real personal income and real personal consumption expenditures all rounded down to 0% growth which would roughly be consistent with a 0.4% annual GDP growth rate. Durable goods were negative.
The decline in the 7 year growth rate started with 2001 growing 0.9% which was slower than the 4.0% growth in 1994. The next year, 2002 was slower than 1995 and this pattern has continued ever since. The 1.8% rate for the first half of 2013 is on track to be slower than the 2.7% growth in 2006. I expect the trend will continue with 2014 growing slower than the 1.8% in 2007. In fact I expect a recession with growth turning negative.
Looking further back the last 7 years have grown slower than any 7 year period since the one ending in 1952. The choice to look at 7 year periods is not completely arbitrary: while not statistically significant annual GDP growth has its strongest positive correlation with itself with about a 7 year lead time.
The conventional view is that the economy will strengthen in the second half of this year and that 2014 will grow more than 3%. Such forecasts apparently assume the influences that brought the weakness which included the great recession have passed. Some think it was just bad housing policy which is now fixed and since housing is recovering we should return to the normal post W.W. II growth rate of about 3.3%.
Since I started examining the influence of tax policy on growth in 2005 I have a different theory about the long term weakness. It appears to me there has been a multi-decade shift toward tax policy that favors consumption and financial investment at the expense of real investment and labor. The share of GDP going to labor and fixed private non residential investment has trended down in recent decades along with the economic growth rate.
The influence favoring consumption at the expense of growth shows up in the trade deficit. The broadest measure of the trade deficit, the current account, has a strong leading relationship with GDP growth when the trade deficit is worse than 2% of GDP. The chart below shows a scatter plot of the 7 year GDP growth rate with the current account as a percentage of GDP leading 8.5 years. So each point in the scatter plot shows what the growth rate was over the previous 7 years (on vertical scale) and what the trade balance was 8.5 years earlier (horizontal scale). For example, the lowest point on the chart shows that GDP grew at 1.1% in the 7 years ending in the second quarter of 2013 and that the current account deficit was 5.7% of GDP ($178 billion) in the 4th quarter of 2004.
The red curvilinear best fit line in the scatter plot suggests a positive relationship between trade and long term GDP growth 8.5 years later. When the balance of trade is worse than minus 2% of GDP the relationship with growth appears to be very significant. When the trade balance is better than -2% other factors appear to have much more influence on growth.
The data in the scatter plot is also graphed in a time series format on the right in the chart above. Using the curvilinear fit as a model, The trade deficit of 6.2% of GDP in 2005 Q4 and also in 2006 Q3 suggest the 7 year growth rate could drop to -0.2% in the 2014 Q2 or in 2015 Q1.
A trade deficit of 6.2% means Americans consumed about 6.2% more than they produced and either went into debt or sold off US assets to foreigners to finance the over consumption. The correlation above suggests that the influences that lead to profligate consumption in one period lead to weak growth 8.5 years later.
Since there is no historical US data to show the relationship (shape of the curve) between growth and trade when the trade balance is worse than 6% it might be useful to consider more than one possibility. Below is a chart with a linear best fit line for when the current account is weaker than -2% of GDP.
For the 7 year growth rate to reach the -0.2% in 2015 Q1, as suggested by the red curvilinear fit, GDP would need to shrink at an annual rate of 4.5% for the next 7 quarters. To hit the 0.7%, suggested by the green linear fit, GDP would need to annualize 0% for the 7 quarters. Either way implies a recession.
While the economy has slowed and may slow more earnings have swung more violently than ever. Looking at the standard deviation for annual real earnings growth over rolling 28 year periods reveals the last 28 years are more than three times as volatile as the 28 years that included W.W.I, The Great Depression and W.W.II. This calculation is done using real "as reported" earnings, which includes the expenses that are excluded from "operating" earnings. In the last 4 quarters as reported earnings of the S&P 500 (NYSEARCA:SPY) adjusted for inflation have grown just over 0%, but back in 2010 there was a hard swerve to the right with annual earnings growth hitting an all time record 783% (note scale only goes to 160%). In 2009 earnings fishtailed hard to the left with a record 89% decline. Preceding that decline was a fishtail to the right in 2003 with an 81 year record 73% annual growth. This was preceded by a 51% decline in 2001 which was the worst one year decline since The Great Depression. The next fishtail presumably to the downside may spin stock market out of control. The chart below shows the pattern of increasing volatility.
Historically, 3% GDP growth was the fulcrum around which real earnings grew or shrunk. If GDP grew faster than 3% earnings grew, if slower earnings shrunk. The change in earnings moved about 3 times as much as a change in GDP growth, such that if the GDP growth rate picked up 1% the earnings growth rate would increase around 3%.
In recent years the relationship between earnings and GDP growth has become less clear. They both drop together in times of recession, but the correlation during recovery and expansion is less obvious. While it could be otherwise, I believe the economy still underlies earnings, but that "managed" earnings mask the relationship. Assuming the relationship is intact, it appears the fulcrum point around whether earnings grow or shrink is now about 2% annual GDP growth. The multiple is much larger. Earnings growth may change 20% to 40% for every 1% change in the GDP growth rate. For example, in the 2001 recession annual GDP growth only got down to 0.7% while earnings growth rate got down to -51%.
Right now annual GDP growth has fallen to 2% and earnings growth has bobbled around 0% for the last year. If annual GDP dropped to 1%, earnings might contract 20% to 40%. I am expecting annual GDP to at least go down to 0% in the next few quarters and a much bigger drop for earnings.
Here is what annual real as reported earnings have done since 1870. The green line in the chart below shows a 1.9% growth trend since 1937. Earnings shown in black are about 27% above this trend. Earnings have only been this far above the trend about 7% of the time. I expect earnings to drop to and below the trend. In fact, I expect earnings to be so weak that the trend declines. A declining growth trend may itself be a trend. In 2000 the best fit trend from 1937 was a 2.0% growth rate.
So the long term growth trend of earnings may be in decline even as earnings volatility pushes all time highs. When earnings start declining with the next business cycle contraction you will probably not want to be in stocks.
One Extreme Follows Another
Along with earnings the long term rate of return in the stock market has been going from one increasingly large extreme to another. In this case I am looking at 17 year periods as the long term because valuation has its strongest correlation with stock market return over roughly 17 year periods. Extremity of the swings in market return has been expanding for 80 years. You can see this in the chart below with the widening gaps in the red trend lines.
In 1936 the strongest ever 17 year stock market return after adjusting for inflation/deflation with reinvested dividends was the period from September 1919 to September 1936. Of course, new record returns were set in 1965 and 1999. Since the weak return of the 17 year period ending in 1932 the lows have been getting lower. So far the only 17 year periods with negative real returns were ones ending in the 1981-82 recession. The worst was the period from March 1965 to March 1982.
While recent economic growth has been less volatile than during the world wars and The Great Depression stock market returns have become more volatile.
Historically, 17 year periods with strong return begin with a low valuation on the stock market, while 17 year periods with weak or negative return begin with a high valuation. The PEses which uses exponentially smoothed real earnings and the real price of the S&P500 has the best correlation with 17 year returns of any measure of value I have found.
In the chart above each point on the black line shows the return for a 17 year period and each corresponding point on the green line shows what the PEses was at the beginning of the 17 year period. The green scale for the PEses is inverted so high PEs are at the bottom to correspond with weak returns and low PEs are at the top to correspond with strong returns. The last point on the green line shows a PEses of 41.7 for August 2013. The correlation suggests the 17 year period from August 2013 to August 2030 will annualize loosing 2% after inflation and dividends. More ominously the PEses suggests the 17 year period ending December 2016 will annualize loosing 11% a year. For this forecast to come true would require a stock market decline of about 89% in the next 2 to 6 years. While I consider an 89% decline possible a 60% to 80% seems more plausible.
There are a couple of caveats with this analysis. First, valuation says nothing about timing. While there is a strong correlation with 17 year periods there is almost no correlation with returns over one year periods. Using valuation as a timing tool could easily lead to missing tops or bottoms by two or three years or even more.
Second, the latest data point is about 2 standard deviations above the estimated return. The annual return for the 17 years ending August 2013 is 5.0% while the forecast was only 1.0%. This gap might be a normal fluctuation which could go to the opposite extreme sometime in the future. On the other hand, conditions that created higher than normal returns might continue for some indefinite time. Conceivably, the relationship that has been in place for over a century was dependent on conditions that no longer exist.
The actual return also exceeded the estimated return by 2 standard deviations in September 1929 and in April 2000. Those points were of course followed by substantial bear markets. These points are in red in the chart below which shows the same data as the chart above, but in a scatter plot format.
You might notice there are more points below the negative 2 standard deviation dashed line than there are above the positive 2 standard deviation dashed line. Stock market returns typically do not quite fit the normal bell shape distribution:usually the negative tail is fatter. Or stated more simply the downside is bigger than people expect.
The downtrend in long term economic growth combined with the uptrend in volatility of stock earnings and returns along with high stock market valuations create a highly unfavorable risk reward ratio for the next 2 to 6 years. The downside for the broad US stock indexes may be 60 to 90% while the upside is probably less than 10%.
The downtrend in the long term economic growth rate shows no sign of abating. The bounce in quarterly growth rates the last 6 months may be over. Indicators that historically foretold changes in short term growth do not seem to be working now. The trade deficit and tax policy variables have continued to correctly indicate the decline in the long term growth rate and point further down. A climax of the crisis appears to be slated for the beginning of 2015. A recession appears likely to begin in the next two or three quarters. Much of the middle class is already in recession where median real income is in decline even though average real income is still rising, or at least was rising until July.
The stock market is stretched. The PEses implied return for the 17 year period ending in 2016 suggest the stock market is dancing on the edge of a bigger cliff than it has seen since 1929.
The driver of the market appears to be the psychological pull of a powerful 4 year uptrend. The driver may be over confident and unaware there is a transmission problem and how slick the road is. If major market indexes fail to make a new high by the end of October it is probably a sign the most destructive leg of the secular bear market in real stock prices that began in 2000 has begun. New nominal highs would only delay the eventual decline.
Disclosure: I am short SPY. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: There is no guarantee analysis of historical data their trends and correlations enable accurate forecasts. The data presented is from sources believed to be reliable, but its accuracy cannot be guaranteed. Past performance does not indicate future results. This is not a recommendation to buy or sell specific securities. This is not an offer to manage money.