CSR makes for more efficient markets
February 18, 2013 | 2 min read
In their paper “Corporate Social Responsibility and Earnings Forecasting Unbiasedness” published in May 2012, the three researchers from the University of Roma Tor Vergata, wish instead to “emphasize that an interesting unexplored dimension, where it is possible in principle to evaluate the relationship between CSR and risk, is that of analysts’ forecasts on earnings.”
A complex database
To enter this ‘unexplored dimension’, the authors used data from several sources. They examined U.S. analysts’ forecasts of U.S. firm earnings from 1997-2004, using the Institutional Brokers’ Estimate System ( I/B/E/S) to determine forecasted and actual earnings per share (EPS); CSR scores from the most widely-adopted CSR scoring standard, the Kinder, Lydenberg and Domini Research & Analytics, Inc. (RiskMetrics-KLD); and data on corporate stock prices from Datastream. Merging the three data sources led to a total of 162,070 observations, with 4,325 unique analysts and 629 unique companies as part of the sample.
“By comparing this information, and calculating the absolute value and the dispersion of the earning forecast error for high/low CSR oriented firms it is possible to check how CSR affects an ex-post measure of risk and uncertainty represented by the distribution of the deviation between ex-ante analysts’ forecasts and actual ex-post released corporate earnings,” explain the researchers about their methods.
CSR: predictable or unpredictable forecasting?
The RiskMetrics-KLD ranking is based on eight CSR domains in firms: community, corporate governance, diversity, employee relations, environment, human rights, product quality and controversial business. Firms ‘earn’ a positive point for all applicable CSR strengths they evidence, and a negative point for each CSR weakness.
Becchetti, Ciciretti and Giovanelli (whose paper wasproduced as part of the Sustainable Investment Research Platform funded by MISTRA, the Swedish Foundation for Strategic Environmental Research) argue that“there are at least two reasons why CSR (and more specifically RiskMetrics-KLD scores for social responsibility) should affect the earning forecast error.” One is increased transparency in reporting; the other is that “if CSR is assumed as minimizing transaction costs with stakeholders (and reducing litigation), it consequently tackles an important source of risk….thereby reducing uncertainty and variability of earnings.”
Top and bottom scores
Ultimately the research team looked at the 20 companies at the top, and the 20 at the bottom, of the low-CSR-weakness fields as indicated by RiskMetrics-KLD domains.
“A crucial aspect of our findings is that CSR contributes to make financial markets efficient as unbiasedness and efficiency are (in almost all specifications) not violated in the subsample of the top 20 percent (lowest CSR weaknesses) companies, while they are in the bottom 20 percent CSR companies. Corporate Social Responsibility has been generally considered in the literature as something unconventional with respect to mainstream financial theory postulating maximization of shareholders wealth and supporting the efficient market hypothesis. What we actually document with our research is that this perception is wrong: CSR seems indeed to bring markets closer to efficiency since it significantly reduces the earning forecast bias and the variability of analysts’ forecasts,” they conclude.
Becchetti, Leonardo, Rocco Ciciretti, and Alessandro Giovannelli. Corporate Social Responsibility and Earnings Forecasting Unbiasedness. SSRN Scholarly Paper. Rochester, NY: Social Science Research Network, May 4, 2012. doi:10.2139/ssrn.2050870.
with attribution (and link, if online) to www.tias.edu.
To be cited as: “CSR makes for more efficient markets”, Lloyd Kurtz, www.tias.edu, February 18, 2013.