Beating the market by looking at a calendar seems totally inconsistent with what would normally be expected. But that's the nature of a "calendar anomaly."
The "January effect," discovered in the early 1980s by Wharton Professor Don Keim, is the best-known calendar anomaly. It states that small stocks significantly outperform larger ones during January.
This anomaly has occurred in 81% of the years since 1925 but not in any of the last five years. Widespread publicity has most likely led to its demise.
Wharton Professor Jerry Siegel, in last year's best-selling "Stocks for the Long Run," identifies other such anomalies: September is the worst month for investing; Monday is the worst day and Friday the best; the first half of a month is much better than the second.
Calendar anomalies exist in other areas, including business cycles. The author coined the term "turn of decade" or "TOD" effect in the Nov. 19, 1990, issue of the Fort Lauderdale Sun-Sentinel ("Economic Lag Fits Turn- of-Decade Pattern"). The TOD effect refers to the unusually high likelihood of a recession's beginning in years that end in 9 or 0.
This anomaly has a remarkable track record in predicting recessions: 80% for the last five decades and nearly 80% for the 14 since 1854. Official recession dates from the National Bureau of Economic Research are available only since 1854.
Four of the last six recessions were TOD recessions. One of the two exceptions was during 1981-82, shortly after the 1980-81 TOD recession.
The 11 TOD recessions on record include the Great Depression and the 1899-1900 "turn of century" recession. Of the three turns of decade without recessions, two - 1879-80 and 1949-50 - were clearly expansionary periods.
Some economists, however, have argued that a recession was, in fact, under way in the December 1939-March 1940 period. This recession (not recognized by the National Bureau of Economic Research) would increase the TOD effect's predictive accuracy from 79% to 86%.
Of course, recessions occur at times other than the turns of decades. The economic research bureau has identified 31 recessions since 1854, and slightly more than one-third of them, 35%, were TOD recessions.
If recessions were totally random occurrences, it would be expected that 20% would happen in years ending 9 or 0. TOD recessions thus occur at a rate nearly twice that suggested by chance alone.
The TOD effect might be driven by the practice of treating decades as discrete time capsules with specific characteristics, for example, the Roaring Twenties. This decade mentality is likely to be even more powerful for centuries and, especially, millenniums.
The TOD effect may also be explained by uncertainty among consumers and businesses as they enter a new decade. "Why should I buy a soon-to-be outdated 1999 model when the new and improved year-2000 models will be coming out?" Would you rather own Windows 99 or Windows 2000?
Uncertainty is a fact of life, but doubts seem much greater as we enter new decades. This uncertainty, especially when aggravated by unexpected events, may translate into reduced confidence in the economy.
With about two-thirds of our economy driven by the economic and market psychology of consumers - the business sector is second in importance-such reduced confidence in a new decade might reduce consumption and investment.
This scenario would set into motion the classic economic cycle of reduced aggregate demand, resulting in lower corporate profits, output, employment, and so on.
Not all TOD recessions, however, follow this demand format. Some have been caused by supply factors, for example, threats to the oil supply, or other economic shocks, for example, wars. These other factors work independently and simultaneously by further reducing consumer and business confidence.
Though there is much we do not know about links between uncertainty and lack of confidence, some major sources of uncertainty during 1999-2000 might reduce U.S. consumer and business confidence substantially, bringing on a TOD recession:
The year-2000 computer problem: Ed Yardeni of Deutsche Bank Securities predicts at least a 60% chance of a year-2000 recession based solely on the Y2K computer problem.
The year-2000 presidential election: Putting aside the impact of the current impeachment trial, there will be considerable uncertainty over the outcome and economic implications of the election.
The bursting of the stock market "bubble": Even if most small equity and fund investors are in it for the long haul, their confidence in the market may have been shaken by the recent 20% selloff, despite the subsequent rebound. Some economists believe a sustained market drop of 10% to 20% would cause a recession, but others say it would take a 30% to 40% drop for that to happen.
Professor Siegel reminds us that almost without exception stocks turn down before recessions, noting that 93% of all recessions were preceded or accompanied by 8% declines in the total stock returns index. Remember, though, that not all bear markets cause recessions.
The new euro and European Union: Most of the complicated transition into this new currency and mammoth monetary union will occur during 1999- 2002. Uncertainty about it could slow economic growth there.
Such a slowdown could reach the United States via the recession "virus"- the fact that each of our last four recessions was associated with contractions in most of the world's 11 major countries. The Financial Times has estimated that 40% of the world is now in or on the brink of recession.
The Japanese and broader Asian meltdown: Japan's current recession, the worst on record, is part of a larger and potentially expanding slowdown throughout Asia. Problems in Brazil, though closer to home, are likely to have a lesser impact on our economy.
Likely credit crunch: Federal bank regulators have been preaching about the need to tighten relatively lax underwriting terms before the next recession.
This jawboning, like that preceding the 1990-91 "rolling recession," may end up tightening the credit spigot much more than it should - and at exactly the wrong time, when the economy may be contracting. More pronounced in its economic impact may be the credit crunch developing in key capital markets such as junk bonds.
The "X" factor: The wild card or "X" factor refers to some unpredictable shock that may be enough to tilt the economy into recession.
If the economy keeps growing until February 2000, it will have set a record for the longest economic expansion on record, topping the 106-month Kennedy/Johnson expansion of the '60s. We may witness such a record, but it will likely be short-lived.
The TOD effect makes it 80% certain that a recession will begin in 1999 or 2000. The earliest any TOD recession has begun is June of the "9" year, but the most frequent, and also average, starting point is January of the "0" year. If that holds true, we would not break the '60s growth record.
The last turn-of-the-century recession began in June 1899 and lasted until December 1900. There was a recession from June 1869 to December 1870- starting about a year after President Andrew Johnson's 1868 Senate impeachment trial.
Most recessions last about 18 months, but the last four TOD recessions were only half that long, on average. The most recent recession, 1990-91, lasted only eight months and was moderate by most standards.
Many questions are unanswered about the TOD effect. Still, like the January effect, TOD has an impressive track record. It is difficult to argue with nearly 150 years of business cycle data and patently unwise to bet against something with a likelihood of nearly four out of five.
This is especially the case with a potentially more powerful and simultaneous turn of the century and turn of the millennium effect, when uncertainty and a potential lack of confidence may be even greater.
