Wishful thinking vs. Reality: Investigating Analyst's Forecasting Accuracy
Financial analysts, especially those working for large multinational banks and investment funds, often seem to be perceived as gurus of the capital markets who are able to forecast a market’s future direction. The media is filled with analyst’s expectations, opinions, earnings- and cash-flow-forecasts, as well as buy- and sell-recommendations. Considering their prominence and ubiquity, it is safe to assume that analyst forecasts are the basis of countless investment decisions worldwide.
Do analysts live up to their reputation? Are they able to make accurate predictions about future stock market movements? Numerous academic studies create a different impression. According to Imhoff and Pare (1982), analysts were not able to make more accurate predictions than simple econometric models, a finding confirmed by Wild and Kwon (1994) between 1981 and 1986. Dreman and Berry (1995) go a step further by calling into question the theoretical validity and practical relevance of the most commonly used valuation models. According to their findings, the vast majority of valuation models do not produce meaningful and scientifically relevant output. More current studies, such as one by Kwag and Stephens (2007), come to similar conclusions. They find that the absolute prediction errors between 2000 and 2003 –depending on country and sector– reached values of up to 50%!
Especially noteworthy are the results of Dreman (1998, p.2), who investigates 500,000 analyst earnings forecasts made three months before the company’s report date. The data for more than 1,500 public U.S. companies reveals an average forecast error of 44%! In this context, it is interesting to note that –contrary to what one might expect– the forecast error has not been reduced by technological advancements. In 2008, StarCapital extended and confirmed Dreman’s (1998) research. The investigation of more than 1.5 million consensus estimates over a time frame of more than 33 years showed that the absolute earnings forecast error 12 months before the report date was more than 30%:
Dreman (1998) investigates 500.000 analyst forecasts made between 1973 and 1996 (data from A-N Research Group and I/B/E/S, we present the results until 1984). In 2008, StarCapital AG repeats Dreman’s (1998) study using 1.5 million consensus estimates (consisting of roughly 10.5 million individual estimates) made between 1985 and 2008 on the basis of I/B/E/S and Worldscope.
The previous sections have shown that financial analysts –on average– do not seem to be able to make accurate predictions of future earnings on a single company level. Are the results on a more aggregate level, namely an equity index, more promising? Each year, German newspapers ask financial analysts (mainly from banks) for a year-end-prediction of the DAX30 index. Even though it is clear that no one can realistically expect a precise forecast, it should be safe to assume that financial experts are able to predict the general direction of the markets, e.g. that the DAX30 closes on a lower level in twelve months than today. Unfortunately, the results speak a different language:
Sources: handelsblatt.com, welt.de, boerse.de, faz.net, manager-magazin.de, focus.de
The forecast error is the difference between the estimated and actual price change of the index in a given year.
Based on the average forecasted year-end-value of 11,844 and a mean forecast error of 18%, one should not be surprised by year-end-DAX-values between 8,828 and 12,658 points. These results are unsurprising, since the future is neither foreseeable nor predictable. It is safe to assume that financial analysts will never be able to forecast unexpected events such as economic crises, wars, natural disasters or technological advancements and discoveries.
List of References
Dreman, D., Berry, M. (1995, May-June). Analyst Forecasting Errors and Their Implications for Security Analysis. Financial Analysts Journal, 51 (3), 30-41.
Dreman, D. (1998). Contrarian Investment Strategies : the next generation : beat the market by going against the crowd. New York: Simon & Schuster.
Imhoff, E., Pare, P. (1982, Fall). Analysis and Comparison of Earnings Forecast Agents. Journal of Accounting Research, 20 (2), 429-39.
Kwag, S., Stevens, A. (2007, Spring/Summer). Serial Correlation of Analyst Forecast Errors in 12 Asia-Pacific Markets. Journal of Applied Finance, 29-39.
Wild, J., Kwong, S. (1994, October). Earnings Expectations, Firm Size, and the Informativeness of Stock Prices. Journal of Business Finance & Accounting, 21 (7), 975-996.