In a week ago's edition, Barron's featured a piece, "Cashing in on Capex", that linked capital expenditures and stock market returns. There were two themes embedded within the article: "Capex is important to stock investors because it signals that managers are spending money to finance growth, and because it bodes well for sellers of capital equipment." Although the latter may be true, the concept that an increased investment spending environment is good for the stock market has been contradicted by piles of academic research.
It's a logical conclusion, and an easy mistake to make. Even we noted in our blog post of January 30th that "Business Investment contribution to GDP was 1.19%. The rebound in business investment does point to some fundamental economic strength and optimism." Similarly, the Barron's article notes, "private investments in equipment and software jumped 12.4%, the biggest increase in five quarters. Conditions have rarely been better for a pickup in capex." Intuitively, that should be good. However, empirical results say that boosting capital expenditures is inversely correlated to stock returns.
A basis for much of the Barron's story comes from an Investment Strategy report by BMO Capital Markets' Chief Investment Strategist, Brian Belski, written on February 1st:
"We continue to believe that the U.S. is on the verge of an early capex recovery, one that is likely to have some considerable legs in our view given the fundamental and economic stability the U.S. offers its global partners… After declining for most of 2012, capex guidance has improved relative to consensus estimates in recent months…. Sooner or later, companies will need to invest in their business to provide themselves and investors with future growth opportunities. We analyzed performance trends for all capex cycles since 1990 and found that all areas of the market perform significantly better when the overall level of capex is improving."
According to the BMO report, the S&P 500 Index has a median gain of 10.8% in periods of rising capex, and a 7.7% decline as capex falls. There may be a relationship between capex cycles and equity returns as a class, but I would suggest it is a coincident, or even potentially, a trailing indicator.
Often companies make decisions to grow because their stocks are doing well, not the other way around. Conversely, in down markets, it is more difficult for companies to obtain financing that allows them to increase expenditures. A 2010 study of international firms in the Journal of Human Capital concluded that growth in capex was a direct result of growth in earnings, but the relationship between growth in capex leading to positive earnings was non-existent.
Basic finance tells us an investment is worth making if it creates value for its owner (in the case of public companies, its shareholders). The objective is for management to create value by identifying an investment worth more in the marketplace than it costs to acquire or build. On a more mathematic level, the appropriateness of an investment is based on its internal rate of return (IRR) and whether than IRR exceeds the required return -- often the weighted average cost of capital. However, estimating and discounting future cash flows of a project or acquisition is a fraught with hurdles, making the process more art than science -- even for an insider. Additionally, choosing when to make such long-term decisions in light of short-term market expectations only enhances the problems.
In the realm of public company investing, real-life analysis of capital budgeting decisions is virtually impossible. Managements are under pressure to grow and put the best spin on their uniqueness to exploit opportunities, yet public investors must rely on those same managers' estimates and biases since we are not privy to the cash flow calculations. Consequently, it is not uncommon for investors to be led into over-optimistic conclusions about the prospects of new expenditures.
Exasperating the issue, the markets' reaction and the level of new investment can often feed upon themselves. McConnell/Muscarella (1985) notes that announcements of increases in planned capital investment are generally associated with significant positive equity returns. Follow-up studies by Blose/Sheih (1997) and Vogt (1997) find a significant positive relationship between the magnitude of the reaction and the actual level of new investment made. In other words, the markets react to the announcement of capex plans, and managers ultimately spend more when the reaction is good.
Potentially for that reason, when using capex as a leading indicator, it points to underperformance, not excess returns. A 2004 study from the National Bureau of Economic Statistics produced by Titman, Wei and Xie finds that, "firms that substantially increase capital investments subsequently achieve negative benchmark-adjusted returns." Their evidence suggests that firms that increase their investment expenditures the most tend to underperform over the following five years. The authors note that a large portion of the underperformance occurs around earnings announcements, suggesting that over-optimistic prospects is the leading cause. Systematically-biased expectations are uncovered when new information is revealed normally through earnings announcements. Additionally, the authors find that the negative return relationship is independent of long-term return reversal theories (e.g., stocks that go up, must come down, as studied by Thaler) and/or secondary equity issuance theories (e.g., high-growth companies fund themselves by raising equity, which in turn dilutes existing holders).
The 2004 study created a hedged portfolio by going long the companies with the lowest capex relative to their last three year average, and short the companies with the highest abnormal capital intensity. With data back to 1973, the universe was divided into 40% long, 40% short and 20% not held. The portfolios were rebalanced each year. The spread between the long and the short was a statistically significant 0.168% a month, or 2.02% a year. The results are stronger in the second year (2.26%) and continue for five years (1.91%, 1.85% and 1.64% in the 3rd, 4th and 5th years, respectively).
Interestingly, there is one period of time where high capex companies outperformed low ones -- the hostile takeover boom of 1984-1990. The theoretical spread trade between capital intensities is positive for 15 of the 17 years, excluding the those hostile takeover years. The rationale for such a reversal during that period is that companies with unproductive capital investment were quickly subject to a hostile takeover, or were forced to implement shareholder-friendly measures to keep the barbarians at the gate, thereby skewing the results.
The University of Kansas' Christopher Anderson came to a similar conclusion in the Journal of Finance. His 2006 study observed that companies with accelerated investment spending experienced lower sequential stock returns, on average, than firms that slow investment spending.
The study showed that for the year after the portfolio's formation, average monthly returns are 1.18% for the companies with the highest two year growth rate in capex versus 1.75% for the lowest growth rates. Although intuitively, the exercise of growth opportunities will have a larger influence on the risk/return characteristics of small capitalization firms, Anderson notes that even among medium and large cap companies, high growth rates in capex predicted lower subsequent average returns.
At Bloodhound Investment Research, we have the ability create an investment strategy using specific fundamental criteria combined with pricing data and back-test the results using our 26-year point-in-time database without survivorship or look-ahead biases.
Similar to the Titman study, we created the following custom formula as a measurement of above/below average capital intensity using Meta Stock language.
We ranked all U.S.-domiciled stocks that were listed on the New York and NASDAQ exchanges in the Bloodhound System. The highest rank equated to the abnormally highest recent capital intensity compared to the company's previous three years of capex. The lowest rank is the least investment compared to the previous three years. We built two portfolios, Abnormally High (AH) and Abnormally Low (AL), with the top 100 and bottom 100 ranks, respectively, and reconstituted the portfolio annually. A score of zero would indicate the company's spending patterns have not changed. To avoid any small-cap effects, we limited the list to companies whose stock was above $5 per share, and whose market capitalization was greater than $500 million.
Over the last 20 years, the AH strategy had a risk-adjusted return of 0.59 versus the index of 1.03. Risk-adjusted return measures the expected return divided by the standard deviation, comparing the spread of values to the expected value. A value of less than 1 suggests that the risk of a negative return is quite high. The risk-adjusted return of the AL strategy is 0.93 over the same 20-year period.
The less capital intensive AL strategy generated an annual 12% return since 1987, whereas the AH strategy generated only an 8.7% return in the same period. In the 27 year periods (including YTD) since 1987, the AH portfolio underperformed the Value Line Arithmetic Index in 21 of them. In the years it underperformed, its average underperformance was 9.2%. The AH strategy's best year was (not surprisingly) 1999, when it generated a 40% return compared to the Index's 10.6%. Two-thirds of the gains that year were in Tech and Telecom companies. Echostar (DISH, now DISH Network (NASDAQ:DISH)) and Immunex (IMNX, acquired by Amgen (NASDAQ:AMGN)) were the only top 10 gainers, not from those sectors. As one would expect, the strategy significantly underperformed in the following three years.
The AL portfolio still underperformed the index in 16 of the 27 periods, but the average underperformance was less at 6.7%. A little surprisingly, the AL strategy performed almost as well as the AH strategy in 1999 (32.9%), but had significantly less pullback after the bubble burst. By the end of 2003, the AL strategy had an average 5-year return of 12.8%; whereas the AH strategy was only 2.3%. Overall, if one invested $1 million in the AH portfolio in 1987, it would be worth $5.2 million today. Conversely, if that $1 million were invested in the AL portfolio, it would be worth $13.4 million.
Building an AH portfolio of today's of 100 companies that have the highest change in capex spending would have a balanced cross section of industries. Healthcare and Consumer Cyclical would equally represent the two largest sector allocations, followed by Industrial and Technology. However, Healthcare dominates six of the 10 highest scores.
Interestingly, Healthcare dominates today's AL portfolio. Nine of the 110 lowest scores are healthcare companies. Like the AH portfolio above, the broader view of 100 names has a better cross section of sectors. Among the 100, Healthcare only represents 26% of the portfolio.
Overall, capex is not a leading indicator of future stock gains. Rather, companies that are increasing their capital spending levels are more likely building off past performance. There are a number of possible explanations for this effect. Unwittingly, the Barron's article may have addressed one of the more important factors when the author noted, "Corporations hold a record amount of cash. Managers have already boosted dividends and share repurchases, and the latter becomes less attractive as stock prices rise." Consequently, they need to do something, and many live by the motto -- grow or die. And growth isn't always the answer.