I just read the introduction of a paper that compares commercial and academic risk measures (Price, Sharp, and Wood 2010). And the winner is (drum roll….) commercial risk measures in almost every test. Here’s the abstract:

Although a substantial body of academic research is devoted to developing and testing risk proxies that detect or predict accounting irregularities, the academic literature has paid little attention to commercially developed risk measures. This is surprising given the general consensus that accruals-based risk proxies in the academic literature are very noisy (McNichols [2000]). We compare the commercially developed AGR risk proxy with proxies from the academic literature to determine which measure best detects and predicts accounting irregularities. We find that AGR outperforms academic risk measures in all head-to-head tests for detecting and most head-to-head tests for predicting Securities and Exchange Commission enforcement actions (AAERs), egregious accounting restatements, and shareholder lawsuits related to alleged accounting improprieties. Incorporating commercially developed risk proxies into future research may yield interesting insights beyond what academic proxies have provided to date.

This looks like a good paper, not only because of the comparison it performs, but also because it gives a good overview of the state of the art in risk measures, both commercially and academically. I’m looking forward to reading the rest of the paper.