A 2006 study led by Claes Fornell claimed that an investment strategy built on changes in the American Customer Satisfaction Index (ACSI) produced persistent, market-beating returns with lower risk. Fornell, who helped create ACSI, recommended going long firms whose ACSI scores beat rivals and rose by at least two points and shorting those that trailed and fell by two points; the paper reported outperforming the S&P 500 every year tested. That specific two-point cutoff and the combination of level-plus-change rules were central to the claimed anomaly.
Three independent reexaminations in 2009 dismantled the result. Ittner, Larcker, and Taylor showed that using standard multifactor risk adjustments, t-tests, and out-of-sample checks eliminated the abnormal returns and that the original findings were highly sensitive to the arbitrary cutoff and portfolio construction. Jacobson and Mizik found the effect concentrated in about ten computer/Internet firms during the 1995-2006 dot-com era, and a fuller replication they had prepared was reportedly pulled after complaints. A separate direct replication similarly failed to confirm abnormal returns. Additional reporting revealed Fornell had traded on nonpublic ACSI releases in 2003 and the university restricted personal use thereafter. An ETF later built on ACSI exposure has underperformed, consistent with the failed replications and with evidence that any positive associations were narrow, time-bound, or artifacts of data mining.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.