Article Preview
Top1. Introduction
In this study, I track the earnings per share (EPS) and sales projections of equity analysts to explore the precision of the whole analyst industry, rather than the performance of an individual. I also provide statistical measures of analysts’ consensus of EPS, sales, rating, and price target zone. I attempt to provide answers to three main questions: whether the projections of equity analysts are accurate; whether the standard deviations of the projections are too large to allow investors to find the general consensus opinion; and whether analysts share systematic bias.
The meanings of answering these questions are threefold. Firstly, the quality of analysis determines the profitability of investment recommendations and directly affects returns. Barber, Lehavy, McNichols, and Trueman (2001) prove such important role of equity analysts by showing that purchasing or selling short stocks with strong buy or sell consensus recommendations yields annual abnormal returns of at least 4%.Untiedtand Pidun (2013) prove that in corporate portfolio management, passive portfolio managers rely on the quality of analyst to a large extent. Kunz (2011) further documents that analysts are subject to their cognitive abilities in terms of perception and interpretation of the firm measurements.
Secondly, analyst opinions affect corporate capital structure and market environment. Gilson, Healy, Noe, and Palepu (2001) conclude that financial analysts with industry expertise can improve capital market intermediation function. Chang, Dasgupta, and Hilary (2006) provide evidence that firms covered by fewer analysts are less likely to issue equity compared with debt. At an earlier time, Moyer, Chatfield, and Sisneros (1989) support the role of analyst monitoring as an efficient device for controlling agency-related costs of debt and equity and as a response to the information demands of investors. Hall and Stanley (2012) also suggest that diversity and governance policies appear to have played only a modest role in explaining corporate performance, while analysts’ opinions can significantly affect the performance.
Thirdly, analyst projection interferes with the future equity returns by setting market expectation. Bhojraj, Hribar, Picconi, and McInnis (2009) show that the firms exceeding analyst forecasts can generate a short-term stock price benefit compared with firms that underperform forecasts. Kasznik and McNichols (2002) document that abnormal annual returns are significantly greater for firms meeting expectations and the market assigns a higher value to firms that meet expectations consistently. Fuchs and Sarstedt (2010) address that narrow samples in empirical management research are sufficient proof of the divergence between market expectation and practical implementation.
In the U.S. equity market, the pool of equity analyst is highly diversified in terms of expertise and experience. Numerous recommendations issued by stock analysts lead to the most common question: are they reliable? In fact, this question includes three concerns: the precision, the standard deviation, and the distribution pattern of analysts’ projections. Precision refers to the quality of analysts’ projections as a whole. However, even though the mean of analyst projections is accurate, a large projection standard deviation might confuse investors who are unsure about which individual analyst opinion they should follow. Furthermore, abnormal skewness and kurtosis of analyst opinions represent analyst irrationality and larger risk of following such opinions.
Some previous studies address the issue of analyst projection precision and conclude that analyst’s models produce better forecasts than naïve time series models, martingale, and submartingale models (Brown and Rozeff, 1978). Consistently, Unlu and Yan (2009) find that changes in aggregate analyst recommendations forecast future market excess returns and changes in industry-aggregated analyst recommendations predict future industry returns. However, these studies focus on the comparison among models, rather than track the projection and the true accounting value.