November 2004 The Problem of Induction "Custom, then, is the great guide of human life. It is that principle alone, which renders our experience useful to us, and makes us expect, for the future, a similar train of events with those which have appeared in the past." -- David Hume   ICHARD NIXON FAMOUSLY DECLARED, "We are all Keynesians now." Thirty-three years later, we may not all be Keynesians, but it seems everyone is an economist -- especially investors. As the up, down, and generally flat 2004 comes to a close, most traders -- whether stocks or bond -- are taking their cue from the economy. Falling oil prices send stocks up while rising prices boost bonds. The Fed's avowed belief in the strength of the U.S. economy supports both stocks and bonds. Everyone anticipates the Labor Department's monthly employment report. In a sense, the focus on macro-economics is a welcome change from the Internet IPO froth and hyped analyst reports of the old equity bubble days. Nevertheless, the current emphasis on economic numbers raises problems of its own. This year's presidential election subjected each release to every sort of political spin. Data, which seems so objective -- often to the point of being exceedingly dry -- was twisted to support each party's own agenda. Careless traders ran the risk of investing on political belief, not solid market analysis. But even putting politics aside, investors should be cautious in interpreting economic data, not because of anyone spinning the numbers but because of two basic yet fundamental issues. These aren't new but they have gained importance given the economy's newfound prominence in investors' analysis. Not What You Measure, but How You Measure The first concerns how economic factors are measured. Two researchers (or economists for that matter) probing the same data can come up with strikingly different results if they employ distinctly different procedures. This pops up in a number of well-known economic series and was the basis for debate in the recent political campaigns. Measures of inflation are a prime example. Do you think prices are rising? Sure you do, and to a certain extent this belief is borne out by the Bureau of Labor Statistics' Consumer Price Index (CPI). Through October, reported year-over-year inflation was up just short of 2.5%, about double the rate of a year ago. Chart 1 CONSUMER PRICE INDEX and FED FUNDS RATE Source: Baseline Despite three rate increases since June, real short-term interest rates (defined here as the Fed Funds rate minus inflation as measured by the CPI) still remain negative. But in a broader context, even a 2.5% increase is right in line with the inflation rate from the past ten years and well below the 3.1% long-term average. Based on this, inflation, as the Fed tells us, is not a problem. That's probably small consolation when you're filling your car with $2/gallon gasoline or paying for college or child care when they've risen by 11% and 5%, respectively. Skyrocketing healthcare costs don't seem affected, either. So what's up? You feel rising prices but the CPI doesn't seem to reflect them. It's all in the way the BLS measures the CPI. First off, housing represents a major portion of the index. That's probably appropriate since housing costs represent a considerable portion of most consumer's monthly expenditures. What may not be appropriate is how the BLS measures it. Many consumers own their homes, so it would seem reasonable to estimate their housing costs based on their mortgage payments. But everyone has a different rate and payment, and even those are periodically refinanced. Estimating these would be a daunting, if not impossible task. Rather than divide the U.S. population into homeowners and renters and then estimating average mortgage costs, the BLS simply treats everyone as renters. Housing costs are simply the BLS's estimate of what it would cost homeowners and renters alike to rent their homes. When the economy slowed, rents fell in many parts of the country. To be sure, mortgage rates did, too, and many homeowners lowered their monthly housing costs by refinancing. Yet for most homeowners with a fixed-rate mortgage, housing costs remain relatively stable, regardless of the short-term fluctuations in rents and the overall economy. With housing representing about 40% of the CPI, it's quite likely this method of measurement is understating actual inflationary pressures. (Of course this will reverse itself when rents again begin to rise, but that doesn't mean there isn't a short-term weakness in the process.) Prices of automobiles and appliances are subject to a different problem. This one applies to any product that is periodically updated and improved. Automobiles are good illustrations. Compare your previous car to your current one. If your older vehicle was more than five years old, it probably didn't have side-impact airbags. Your new car probably does. If your old one had a CD player, keyless entry, or automatic climate control, it was probably because you paid extra for them. Now they're considered "standard" features and included in the base price of the car. So if the price of your new car was greater than the cost of your old one, did it really cost more? If you wrote the check or are still making the payments you certainly think so, but the BLS isn't sure. Why? Because it's not appropriate to compare the base price of the two cars if the newer one includes features that were "options" on the older vehicle. Sure, you may not have had side-impact airbags on your previous car, but now you do so you may actually be paying less for the overall vehicle. In essence, you're getting more for your money. There's a certain academic logic to these "hedonic" adjustments, but for most consumers, the bottom line is, well, the bottom line. Are sticker prices rising? Obviously they are and it's equally obvious that while consumers may not perceive the benefit of many added features, they certainly feel the price increase. Chart 2 U.S. EMPLOYMENT Source: Baseline The government's numbers show that despite this year's encouraging reports, jobs have still declined over the past four years. Employment is also dependent on the measurement technique. Throughout the 2004 campaign, Democrats pointed out that according to the government's reported figures, payrolls declined by roughly 500,000 during President Bush's term. On the other hand, the household survey showed just under 1.7 million jobs were created over the same period. What's up with that? Once again, it's all in the way the data are measured. The government's employment figures are obtained by surveying employers. That certainly makes sense given they are, as their name implies, employers. Also as its name implies, the household survey queries households, or more precisely, the individuals who inhabit them. Its results differ from the Labor Department's not because people lie but rather because it captures the self-employed and those working for small businesses falling under the government's radar. The employers surveyed for the government's tabulation don't know about these folks. So this is one of those situations where both measures could be accurate. Employers may still be hesitant to rebuild their workforce following the 2001-2002 recession, yet that doesn't mean displaced workers are all unemployed. Certainly some are, but the household survey suggests many are either now self-employed or are working for other small businesses. Perhaps the two surveys need to be considered together. It also suggests the economy may be undergoing a deep-seated change. The U.S. has definitely shifted away manufacturing to a more service-oriented economy. (That's one of the reasons so many production jobs are now being outsourced overseas.) The reallocation of labor from the traditional factory-based employers of the Labor Department's survey to smaller, more service-oriented businesses may be a result -- or a factor. Like the Past? This brings us to the second fundamental problem for investors in interpreting economic data. Using economic series as the basis for forecasts and investment strategies relies on the assumption that the future will be like the past. Chart 3 S&P 500 P/E RANGES Source: Baseline The top of the S&P 500's P/E range has fallen dramaticaly from the 1990's bull market yet the bottom has remained well above that of the early '90s. This isn't anything new to investors -- especially quants. Basic investing approaches involve targeting sectors, industries, or individual securities that performed well during similar periods in the past while avoiding those that didn't. Clearly for this to work, the investor must assume the future will be like the past. Indeed, many of the most calamitous investing errors stem from the belief that, "It's different this time." Nevertheless, the economy is not static. Its evolution may be gradual or dramatic as during the industrial revolution. Could the move from manufacturing to services represent a similar fundamental change? If so, the past may not be so predictive. Consider equity market P/E ratios. Traditionally they've averaged around 15x towards their highs and the upper single-digits at their lows. Presently the S&P 500's forward P/E is just over 17x. Are stocks expensive? Contrary to comparisons with the past, there's a growing feeling among investors that stocks are actually oversold and are due for a year-end rally. There are some good reasons for this including the conclusion of the presidential race, continuing (albeit slower) profit growth, and stronger corporate balance sheets. Although 17x is high compared to the historical averages, it's near the lowest level since the mid-1990s. And maybe 17x actually isn't high compared to historical standards. Industrial firms traditionally carry lower P/Es than those in other sectors. During manufacturing's dominance, this was reflected in historical P/Es. But if the U.S. has now become more service-oriented, perhaps the historical figures are no longer the appropriate benchmark. Instead, the "averages" may need to be ratcheted up to more closely reflect the dominant sectors. Service sectors have traditionally sported higher P/Es than manufacturing so comparisons to market averages of the past may not be appropriate. Chart 4 S&P 500 EARNINGS GROWTH RATE Source: Baseline The Fed tightening has pushed up short-term rates while long rates have declined resulting in a flatter yield curve. This doesn't just apply to the equity market, either. Lately the bond market has also puzzled investors relying on the past for guidance. Starting in June, the Federal Reserve embarked on a tightening campaign. With real rates (Fed Funds Rate minus annualized inflation) in negative territory, the goal is to bring them back to a sustainable level to lessen the potential of inflation. Yet while the Fed has pushed short-term rates up, intermediate and long-term yields have fallen. That's not how it worked in the past. Of course the economy was so global then, either. Perhaps the biggest factor supporting long-term bonds -- especially Treasuries -- is (ironically) the U.S. current account deficit and weak dollar. Imports are running far ahead of exports, particularly since one of our primary trading partners, Japan, is trying to export its way out of recession. The burgeoning U.S. deficit doesn't help, either, since it keeps the dollar weak, periodically spurring foreign intervention. As foreign countries amass weakening dollars, they turn to Treasuries in an effort to receive a better return. Their demand helps prop up long-term Treasury prices holding down yields that move in the opposite direction. Chart 5 U.S. TRADE DEFICIT and JAPANESE YEN EXCHANGE RATE Source: Baseline The dollar has fallen against the Japanese yen as the U.S. trade deficit has grown. This has happened before, but the deficit has never been this great. In addition, long-term yields are often viewed as the bond market's prediction for future inflation. With the fed proactively moving before inflation becomes a problem, the U.S. may ultimately be the global investor's country of choice for the long haul. In the past, the Fed traditionally waited until inflationary pressures were well entrenched before taking any action, only to overshoot the mark and raise the specter of recession. If the Fed sticks to an anticipatory approach, the previous tightening cycles may not be good indicators for the yield curve. History can be a guide to the future only if current economic and market conditions remain compatible with those of the past. Differences -- especially fundamental ones -- lessen the predictive power. Certain Uncertainty Eighteenth Century philosopher David Hume (whose quote appears at the top of this page) wrestled with this problem on a broader scale. He ultimately concluded that any prediction based on past observation is far from certain since it must assume the future will be like the past -- something that cannot be proven. Economic data and historical market trends -- regardless of how objective they may appear -- cannot provide the definitive guidance many investors seem to believe. There's simply to guarantee conditions and results are comparable or that the economic data truly reflects current circumstances. It's understandable that traders act on the release of every economic report, that's what they do: trade. But long-term investors need to keep economic reports in the proper perspective. They are, after all, just another bit of data to be considered in the investment process. We may all be economists now, but successful investors need to be more. E-mail your comments. Search this site! Just enter you key word or words:   Get current quotes or follow your own custom portfolio, courtesy of E-Line Financials:   Search:TickerName Company DataQuickQuote     Homepage Return to Top