The Halo Effect by Phil Rosenzweig isn’t about discarding the people who claim to have discovered the secrets of business success, but to make a more subtle point, or rather to ask a more subtle question: why is it so hard to define success?
There are many delusions that shape the way we think about business.
In the physical sciences, knowledge seems to move forward. Science doesn’t deal with beauty of truth or ethics, it’s practical. It simply asks: if I do something here, what will happen there?
What leads to sustained growth? is a scientific questions. If a company does one thing or another, what will happen to profits or share price? The only way to investigate the answers to a scientific question is through trial and error. Conduct experiments, put together information in systematic ways to figure out the rules that govern the phenomena, which ultimately leads to accurate predictions.
But sciences like physics and chemistry can depend on experiments that are run in laboratories. And we can control the settings in these labs. That means that we can manipulate the variables, collect data, and try again. With so much information, it becomes easier to decipher the rules. But is that the case in the business world?
For some questions, “no”, since the business world is complex and messy. But in some cases, it is possible to conduct experiments. If you wanted to figure out the best location to place an item in a supermarket, how much a change in price will affect the demand for your product, or the effect of a special promotion, then you can run trials in different stores and compare the answers.
Wal-Mart’s success can be owed to its implementation of scientific rigor to merchandizing, studying the patterns of consumption, and understanding the behavioral traits of customers, and applying their insights. Amazon and eBay also benefit from the scientific technique. So do Casinos.
So, in a dynamic business world where consumer trends are constantly changing and old customers are disappearing and new customers are appearing out of nowhere, and people change companies, and new technologies are introduced, conducting scientific experiments to figure out “best practices” is impossible.
But it is possible to witness a high correlation with the business world’s top performers. And yet, correlation is not causation. Business books like In Search of Excellence and Good to Great don’t just offer descriptive advice, but prescriptive advice. The idea is that if you follow the advice, then success can be yours!
So, what happens, when after a while, those excellent companies start to fail? You get two contradictory headlines. Some analysts would say that a company failed because it refused to innovate, others would say that the company innovated too much, much like the Lego example.
The S&P 500 almost doubled between 1980-1984. Only twelve excellent companies grew faster than the overall market. The other twenty-three failed to keep up.
Some companies did very well (like Wal-Mart, which grew by a whopping 800 percent in those five years), but many well-known names like Caterpillar, Digital Equipment, Dupont, Johnson & Johnson, and Walt Disney didn’t even match the market average. You would have been better off investing in a market index than putting your money on those Excellent companies. If we go out ten years, the record is about the same: Only thirteen companies outperformed the market, which was up 403 percent, while eighteen didn’t match the market. Most of the companies weren’t even average, never mind Excellent.
But maybe there’s a simple explanation for the sharp decline since stock market performance is determined by the expectations of investors. If a company’s stock has been bid up to high levels with price/earnings ratios of 40 or more (Cisco or Google), and the company continues to perform well and deliver expected earnings, its stock price will stay steady but will not continue to rise. These earnings are already reflected in the stock price.
A company, through no fault of its own, may underperform the market over the next years.
Microsoft is a good example.
While its revenues and profits rose by more than 50 percent from 2001 through 2005, the price of its stock has been essentially unchanged, having been bid so high in the 1990s. Maybe that’s what’s going on here: The Excellent companies had their stock prices bid up to very high levels in the late 1970s, and a slide in stock price over the next years was due more to investors’ earlier optimism than any real decline in performance.
To remove the effect of expectations, I used a measure of performance that captured profitability: operating income as a percent of total assets. Again using data from Compustat, I found that for the five years after the study ended, fully thirty of the thirty-five Excellent companies showed a decline in profitability, some by a small amount and others by a lot. Only five improved their performance.
These results make plain that Peters and Waterman’s Excellent companies didn’t decline just because they failed to meet market expectations. These companies, chosen precisely for their outstanding performance, actually became less profitable in the years after the study ended.
The explanation for this decline is that no company is expected to do well forever. But Excellence ought to have at least lasted for a few years. Did the companies waver from their principles? Probably not. But those things were not enough to ensure enduring success. What if there was selection bias?
Guess how many companies on the S&P 500 in 1957 were still on the S&P 500 in 1997, forty years later? Only 74. The other 426 were gone — nudged aside by other companies, or acquired, or bankrupt. And of the 74 survivors, guess how many outperformed the S&P 500 over that time period? Only 12 out of 74. The other 62 survived, yes, but they didn’t thrive.
The only way to find companies that were truly visionary and maintained their success was to do so retrospectively, but never to look at the factors that are associated with success as predictive.
Does this mean that all company performance is just a matter of luck? Is it roughly equivalent to someone who flips a coin and gets heads ten times in a row, but stands no greater chance to flip heads on the eleventh try than anyone else? Not at all. Success is not random — but it is fleeting. Why? Because as described by the great Austrian economist Joseph Schumpeter, the basic force at work in capitalism is that of competition through innovation — whether of new products, or new services, or new ways of doing business. Where most economists of his day assumed that companies competed by offering lower prices for similar goods and services, Schumpeter’s 1942 book, Capitalism, Socialism and Democracy, described the forces of competition in terms of innovation. He wrote:
The only constant pattern we see is creative destruction, not stability or endurance.
The fundamental impulse that sets and keeps the capitalist engine in motion comes from the new consumers, goods, the new methods of production or transportation, the new markets, the new forms of industrial organization that capitalist enterprise creates.
Several researchers have studied the rate at which company performance changes over time. Pankaj Ghemawat at Harvard Business School examined the return on investment (ROI) of a sample of 692 American companies over a ten-year period, from 1971 to 1980. He put together one group of top performers, with an average ROI of 39 percent, and one group of low performers, with an average ROI of just 3 percent. Then he tracked the two groups over time. What would happen to their ROIs: Would the gap persist, would it grow, or would it diminish? After nine years, both groups converged toward the middle, the top performers falling from 39 percent to 21 percent and the low performers rising from 3 percent to 18 percent. The original gap of 36 percent had shrunk to just 3 percent, a decrease of nine tenths. Now, as Ghemawat pointed out, a persistent difference of 3 percent isn’t zero — it’s nothing to sneeze at. But the main point is that high performance is difficult to maintain, and the reason is simple: In a free market system, high profits tend to decline thanks to what one economist called “the erosive forces of imitation, competition, and expropriation.” Rivals copy the leader’s winning ways, new companies enter the market, consulting companies spread best practices, and employees move from company to company.
Another study, by Anita McGahan at Boston University, examined thousands of U.S. companies from 1981 to 1997 and found a similar pattern. Companies were placed into one of three categories based on their profit performance during the first three years (measured as operating income to total assets) — high (the top 25 percent), medium (middle 50 percent), and low (bottom 25 percent). Then they were tracked over the next fourteen years. How much movement was there among groups? Of the high-performance companies, 78 percent were still in the high group at the end of the study, while 18 percent dropped to medium, and 5 percent fell all the way to low. Of the medium performance companies, 81 percent were still medium at the end, while 10 percent improved to high and 8 percent fell to low. As for the low performance companies, 78 percent were still in the bottom group at the end, but 20 percent moved up to medium and 2 percent moved all the way to high. These findings show that performance is not random but persists over time, yet there is also a tendency to move toward the middle, a clear regression toward the mean.