Professor: "You'll see that the software allows you to determine what variables play the most important roles in forecasting. You can then rebalance what weight you want to put on say, oil prices, for future GDP growth estimates in the formula." Student: "What if the variables the model says are unimportant now in predicting the future become important in the future, or if your estimates for the inputs are way off?" Professor: (Pause) "You have to make assumptions that some sort of consistency will take place into the future -- that what worked in the past will work in the future. Frankly, forecasting works better at predicting the future when the future follows the same path; it's the turnarounds that forecasting has trouble with." Years ago in my "economic forecasting" class, the preceding conversation took place. The implications of this point didn't sink in at the time. Everybody makes forecasts. Some are a little more scientific, using software, surveys and data. Others just make guesses, educated or otherwise. Often the "professional" forecasts are no better than the guesses. It's a running joke that economists have predicted 12 of the last five recessions. Most business decisions are made based on some forecast, formal or not, of what the future will look like. Buildings going up around the country exist because the builders think they can sell or rent the units at prices that make the venture profitable. None of the backers are forecasting sharp declines in rental income or sales prices, or else they wouldn't be building, any more than you would invest in the stock market today if you thought earnings in the S&P 500 would be 30% lower next year. Fund investors have something in common with managers and executives. They look at what has happened over recent years, and they forecast that behavior into the future. With investors, if the S&P 500 were up an average of 15% a year for the last five years, then it is a far better investment than bonds if bonds were up only 7% a year for the last five. With executives, if air miles flown per person have gone up 10% a year for five years, then they need to start buying planes and adding gates to ramp up for the expected future traffic. It's the exact same error whether an investor makes it with his $75,000 portfolio or a CEO makes it with a $5 billion loan.