Analysts' Recommendations: Evidence of a Portuguese Investment Bank

1. Introduction

There is still a lot of controversy on the forecasting ability of analysts. On one hand, if markets are efficient in its semi-strong (or even strong) form, in the sense of Fama (1970), there would be no ground for the existence of research departments as it would be impossible to implement a profitable strategy on the basis of the publicly available information. Yet, some authors claim that evidence of analysts' forecasting ability in itself should not be interpreted as a violation of market efficiency if one cannot implement that strategy effectively. In other words, finding that research analysts play an important role in disseminating information may be consistent with market efficiency; only evidence of effective trading strategies on the basis of public information, such as research reports or analysts' recommendations, should be accepted as contradictory evidence. Recent research (see Wermers, 2000, as an example) suggests that the performance of active management is not superior to a passive strategy due to trading costs. Our research can thus also inform on the value created in active management done on the basis of stock picking skills.

On the other hand, in the last few years, there is growing suspicion on the information value of analysts' recommendations (particularly for sell side-analysts) motivated by anecdotal evidence on lack of independence of research departments due to pressures by other investment bank departments such as brokerage or M&A (Mergers & Acquisitions).

The main purpose of this study is to evaluate the effectiveness of several trading strategies built on the basis of the recommendations produced by a team of research analysts in a Portuguese investment bank. Analysts identify undervalued assets for which they issue a buy recommendation. These investment recommendations are regularly published.

All recommendations refer to stocks traded in markets for which the investment bank offers brokerage services. Analysts also issue recommendations on other kind of securities, but we exclude them from our sample.

We design simple trading strategies on the basis of these recommendations. Stocks are bought at the time of the recommendation disclosure, held for a certain period and then sold. We create different strategies for an array of holding periods and analyze the results over time and for subgroups of stocks.

The rest of the paper is organized as follows. Section 2 provides a brief literature review on analysts' recommendations. Section 3 presents the methodology, explains the different recommendation-based trading strategies and describes our data. Section 4 presents our main results and section 5 concludes.

2. Literature Review

The literature on analysts' recommendations has focused on three different questions: (1) the type of analysis used by financial analysts to evaluate stocks; (2) the stock price reaction to the analysts' recommendations (3) the value of analysts' recommendations as effective tools for stock selection from an investor perspective.

Our paper addresses this last question. In a competitive and rational world, investors will only follow analysts' recommendations if the expected benefits are greater than the cost of advice, in other words, when analysts' recommendations are expected to have (informational) value. Financial theory tells us that the most economically rational benefits extracted from an investment recommendation are the positive excess returns following recommendations. Moreover, analysts' recommendations have (informational) value if the analysts have superior or inside information on the financial asset, and/or if the advice service is cost-free .

Even if our central approach is from an investor point of view perspective, our evidence could also shed some light on the impact of this public information (published investment recommendations) on prices and therefore inform about semi-strong or strong market efficiency in the sense of Fama (1970). However, the implications of our study in terms of the impact on stock prices and on market efficiency are limited because these recommendations are disseminated to a small number of investors .

2.1 Evidence on Analysts' Recommendations

The debate over the value of analysts' recommendations to stock selection is not settled and the existing empirical evidence that supports the hypothesis of real superior returns from investing on the basis of analysts' recommendations is not consensual.

The seminal article on investment recommendations was written by Alfred Cowles III (1933) who studied investment recommendations of 16 financial services companies, 25 financial periodicals and The Wall Street Journal editors. Cowles showed that recommended stocks had, on average, a negative performance when compared against a market benchmark, and concluded that investment recommendations didn't add (informational) value.

Until Womack (1996), there was little evidence on whether analysts' recommendations would yield abnormal returns. In spite of some findings suggesting that recommended stocks had positive excess returns, there were criticisms of sample bias or imprecise data. Womack (1996) looks at stock prices' daily reactions to changes in the 14 biggest U.S. brokerage house analysts' recommendations and finds statistically significant positive excess returns from investments on recommended stocks. However, these excess returns show strong mean reversion in the six months following the announcement. Their main focus is to determine the impact of changes in analysts' recommendations on stock prices (and evaluate semi-strong form of market efficiency) rather than to assess the usefulness of these recommendations from an investor's perspective.

Yet, Jaffe and Mahoney (1999) conclude that common stock recommendations made by investment newsletters do not outperform appropriate benchmarks (control firms). Moreover, there is no evidence of performance persistence, when performance is measured by abnormal returns.

Barber, Lehavy, McNichols and Trueman (2001) show that buying (selling short) stocks with the most (least) favourable consensus recommendations, together with daily portfolio rebalancing and a timely response to recommendation changes, would allow a monthly abnormal return of 0.75 percent. The data used in their paper included over 360 000 investment recommendations from 269 brokerage houses and 4340 analysts, from 1985 to 1996. The authors conclude that these results