Private Investment: Between a Rock and a Hard Place

investor1Note:  This is a working paper and will be updated over time with more discussion and data.

GEI News has reported that the McKinsey Global Survey has found most executives believe their companies are not making sufficient investment.  The survey found that the executives surveyed were very conflicted:  They regretted missed investment opportunities and yet they also had risk averse attitudes.  Go to the news article to read what McKinsey found out about the bias against investment. The news from McKinsey got me thinking about what experiences these executives had, besides being scared by the financial crisis, that could put them in the conflicted situations that McKinsey described. To look into this question I started with a look at historical data on private investment in the following graph from the St. Louis Fed.  What I saw was a chart pattern that has the last five years looking like the the deflation of a bubble that grew over the period 1991-2006.  This could put business executives in a frame of mind similar to a person considering possibly buying a home but having doubts about whether or not there is more air to come out of the housing bubble.


POP is the total population of the U.S.

The red curve is the gross private investment normalized for population growth.

There is a strong relationship between private investment and stock market returns, as seen in the following graph:


Hidden in the above graph is a dramatic story, which we will develop below.

Chapter One – Private Investment and Stock Returns

The following graph shows the ratio of the value of the S&P 500 to Gross Private Investment from 1957 to 2011.  The plot has been normalized to make the value of the ratio equal to 1 in the first quarter of 1957.


There are two distinctly different time periods in the graph:  1957-1982 and 1983-2011.  The latter time period might be subdivided at the year 2000.  The years 2000-2011 have been extremely volatile for the ratio of capitalization value to private investment.  That certainly is one factor that could heighten risk aversion of executives.

Chapter Two – Personal Income

Perception by executives about markets for their products can also influence investment decisions.  Shown in the following graph is the history of real disposable income growth.


When the return to stocks was lower for private investment (before 1980), the growth of personal income was 50% higher than for the more recent 31 years.  So in the years when corporate capitalization was growing more slowly, the expectation of future growth was supported by rapidly growing real income.  After the early 1980s, executives were conditioned to keep investing because stock prices were rising more rapidly, compensating for the lower disposable income growth rates.  Particularly in the years after the mid 1990s, consumer debt was escalating so that, even when the 2001-02 bear market crushed stock values, consumption remained strong in spite of the lowered real disposable income growth rate – that is until the credit bubble / housing bubble burst.

We can see a sharp break for growth of compensation paid to employees relative to investment about 1990. The following two decades showed employee compensation had no further growth relative to private investment.  That is the time period of the rise of the super sized credit card and home equity as a piggy bank, which kept the private investment party going.


COE is Compensation of Employees, paid.

After reading the initial draft of this article, Steven Hansen pointed out that private investment is correlated to PCE (personal consumption expenditure).  Steve provided the following graph, showing investment and PCE for durable goods :

Real gross private investment shows much wider swings in reaction to times of economic distress.  This is understandable because PCE contains expenditures for necessities of life and life does continue when the economy crashes.

The correlation seen qualitatively in the graph above has been calculated and plotted in the following graph.  The correlation definitions are defined here.

The results indicate the efficiency of the reaction of private investment to changes in PCE.  The best correlation exists for investment one quarter later than that for which PCE is reported and the second best correlation is coincident.  When one recognizes that PCE data is reported up to three months after the fact it can be argued that investment reaction is essentially coincident with the first announced PCE data.  That may reflect excellent planning research.

But the research is not perfect.  Looking at the minima in the graph provided by Steve (second above) the red bottoms and the blue bottoms are scattered both sides of being coincident.  And for the 2002 recession the investment minimum occurred well over a year before the PCE minimum.

Gross real private investment shows much poorer correlation with total PCE than with the durable goods portion of PCE.  The best correlation coefficient for total PCE is only 0.37 (coincident) and is o.35 for both one quarter leads and lags.

Chapter Three – Employment

Next we look at the systematic decline in employment growth per unit of investment over the entire 50-year period covered, even as real investment generally grew and then accelerated from 1980 to 2000.


PAYEMS is total non-farms payroll.

The effect seen in the previous graph is related to productivity improvement.  Private investment was trying to compound its gains through the wonderful dividend of productivity improvement.

After 1990 (and especially after 1995) there was an obvious “new nirvana” for productivity growth:


From 1948 to 1991 the average annual productivity gain was just under 1%. For the most recent 20 years the average is over 2% per year.  However, productivity gains have been experiencing a waterfall decline over the past two years and that should be making executives risk averse as well.

Note:  It would be a reasonable argument that the dramatic 12 years of high productivity growth from 1996 through 2007 was a computer/internet dividend which is now declining back to a long-term normal.  That might imply that the current productivity growth value is actually near what could be the loner-term average going forward.

Chapter Four – The Pendulum

One final graph shows how the pendulum has swung from labor to capital over the years.   In the following graph the ratio of compensation to employees to the value of stocks is seen to rise until about 1980 and has declined since.  Before about 1980 labor was making significant gains from private investment.  After 1980 those gains were continuously eroding away.


This graph plots the ratio of COE (compensation for employees, paid) to the value of the S&P 500 index. It points out that for much the 12 years 1957-1969 equity (capital) gained relative to labor, using the 1957 ratio as the reference. Then, from 1970-1982 labor gained relative to capital. From 1983-1996 the ratio returned back to the 1957 value. In the late 1990s gains by capital far outstripped labor (due to the stock bubble). If the two spikes related to the major bear markets of our era are skipped the ratio has remained near 1 in the 21st century.

An ideal economic system would not see such wild swings away from whatever ratio is selected as a reference. An ideal system would not see precipitous spikes from major bear markets, or, if they did occur, would see a return to a near normal ratio very quickly after the spike. Labor would share more of the pain of bear markets. The major “bulges” in favor of capital (1960s and 1990s) would also be muted in an ideal system. Labor would share in more of the gains in booms and bubbles. (An aside: If more of the gains from bubbles went to labor and less to capital, it is quite possible that capital would reduce short-term speculation and increase focus on long-term gains. That could reduce the extent of bubbles and instead create long-term investment and stable growth patterns.)

But, in spite of the apparently favorable ratio (from the labor point of view) over much of the past 15 years, median incomes have not been growing. If median family income is used instead of COE, a much different picture should emerge. That graph will be posted later. That brings focus back to income distribution which has reached historic extremes.

Taken all together the data shows a shift in the return on investment away from labor and to improved return on capital over the past three decades.  The findings of the McKinsey survey indicate executives may have doubts about whether the thirty-year improvement for return on capital may be over or not.  Certainly the fact that the current level of productivity improvement is below the pre-1991 average is not going to help change that.

The executive of 2011 has grown up through the bountiful years of the 1980s and 1990s.  Then he (she) has had his (her) teeth rattled and hair turned gray (or lost) by the economic turmoil of the 21st century.  Markets for their products are very uncertain and productivity growth is evaporating.  The potential for capitalization growth is also uncertain – there have been two vicious bear markets this century and they are preconditioned by recent experience to fear another.

One more sad part of all this is that not only is capital under stress, the lower level of productivity improvement does not appear to be producing a better situation for labor either. Right now there are no winners.

Related Article

McKinsey: Is there a Bias against Investment? (GEI News)