By LARRY SWEDROE
Since capital flows impact security prices, an interesting question is whether non-fundamental correlated demand — driven by mutual fund investors’ tendency to chase Morningstar ratings — creates systematic style-level price patterns. Itzhak Ben-David, Jiacui Li, Andrea Rossi and Yang Song, authors of the November 2020 study Non-Fundamental Demand and Style Returns, sought the answer to this question by examining flows based on changes in Morningstar ratings. They exploited the change made in June 2002 when Morningstar began benchmarking funds against peer funds their style (3x3 style boxes for large, mid and small; and growth, core and value). Their full sample period was 1991-2018.
Findings
Following is a summary of their findings:
— Monthly fund flows to mutual funds strongly depend on Morningstar ratings.
— A rating upgrade results in a surge in mutual fund flows. The flow increases lead to contemporaneous stock price appreciation and subsequent reversals.
— Throughout the sample period, five-star funds received flows that amounted to 2 to 3 percent of their assets under management (AUM) per month on average — the AUM of five-star funds increased by about 25 to 40 percent over one year. In contrast, one-star funds experienced outflows of -1.5 to -2 percent of their AUM per month on average. The patterns did not differ before and after June 2002—mutual fund investors chased Morningstar ratings regardless of the rating methodology.
— Before the 2002 reform, the most upgraded styles (i.e., styles with the highest recent changes in ratings) drew large fund inflows, and their returns exhibited momentum and subsequent reversals — the most upgraded style attracted approximately 15 percent more flows than the most downgraded style over the subsequent 12 months. Opposite patterns were observed for the most downgraded styles.
— The impact of a ratings change on fund flows primarily took place over the first 12 months.
— Stocks in the top decile of past rating changes outperformed stocks in the bottom decile by about 20 percent over the subsequent 12 to 18 months. However, the cumulative return difference between the two groups of stocks indeed reverted over the 36-month horizon.
— After the 2002 reform, ratings became evenly distributed across styles, and therefore ratings-chasing flows were distributed across the entire spectrum of styles. Because ratings-chasing investors stopped applying price pressure on a small subset of winning styles, style-level momentum and reversal largely disappeared, demonstrating the powerful influence of Morningstar ratings.
— Price pressure effects were stronger in stocks with higher mutual fund ownership—ratings-induced flows generate substantial price pressure at the stock level. The price effect before June 2002 was stronger in the style portfolios constructed by stocks with higher mutual fund holdings and relatively muted in those constructed by stocks with low mutual fund holdings. There was no effect in any of the terciles after June 2002, as the feedback loop was broken.
— Because Morningstar fund ratings depend on past returns, the pre-2002 rating methodology exacerbated the positive feedback at the style level. Their rating reform led to a decrease in the profitability of style momentum—long/short style momentum strategies generated a considerable return of 70 to 90 basis points per month before June 2002 and became unprofitable afterwards.
— As ratings became more homogeneous across styles due to Morningstar’s reform, so did flows and returns. Ratings-induced price pressure could explain on average 10 to 30 percent of the variation in monthly factor returns before 2002. However, the explanatory power dropped precipitously afterwards.
Summary
Ben-David, Li, Rossi and Song demonstrated that a seemingly unimportant change in Morningstar ratings created a long-lasting impact on the allocation of investors’ capital across styles. That, in turn, altered the cross-sectional variation of style returns, style-level momentum and widely-used return factors.
They concluded: “While style-level return chasing may have taken place on a smaller scale before our sample period, the introduction of fund ratings — which coincided with the exponential growth of the retail mutual fund sector — made it easier for investors to chase style-level returns and thus potentially amplified flow-induced feedback trading at the style level. The fact that style return dynamics changed dramatically after the reform highlights the importance of non-fundamental demand in shaping systematic returns.”