Examining Mismatched Stocks: A Novel Source of Predictable Price Pressure

A recent presentation at the Chicago Quantitative Alliance Fall Conference presented a novel study of mutual fund rebalancing behavior and the effect on stocks that they held.

One SSRN research paper by Cameron Peng and Chen Wang[1] takes this a step further by investigating overall factor characteristics of funds, the stock-level factor characteristics of individual fund holdings, and whether matches or mismatches between the two predict anything about subsequent stock returns.   

In a nutshell: At the quarter end, value stocks held by growth funds, and growth stocks held by value funds, tend to underperform. One possible explanation that they discuss is target rebalancing, where the selling pressure on mismatched stocks outweighs the buying pressure from other funds.    

Fund Factor Demand vs. Stock Factor Characteristics 

Although the paper does rely on reported fund quarter-end holdings, it takes a different approach to classifying the “factor demand” of funds. Peng and Wang regress the funds’ monthly returns against standard factor returns from Kenneth French’s data library over a five-year rolling window and calculate factor loadings for each fund. Their approach avoids using the fund’s self-reported investment objectives or the reported quarter-end holdings, as they may not capture the intra-quarter trading behavior of the fund and may be subject to “window dressing.” It also allows for an objective numeric measure of factor demand for any fund with five years of return data. 

Peng and Wang double-sort individual stocks into 25 (5x5) portfolios based on the stocks’ individual characteristic (i.e., the stock’s book-to-market (B/M) ratio) on one dimension, and the holdings-weighted factor demand of the funds that hold that stock (i.e., the factor demand of the funds that hold that stock, weighted by held shares). They form the portfolios at each quarter end and hold for one quarter, and backtest this from Q2 1980 to Q4 2018. Their 5x5 cap-weighted portfolio results are shown below. 

Figure 1: Annualized Portfolio Return % (Value Weighted) 

Source: Peng and Wang, “Factor Demand and Factor Returns,” SSRN, Oct. 3, 2021 

The corner portfolios highlighted in green represent “matched” stocks (i.e., high (low) B/M stocks primarily held by funds that track high (low) B/M returns). Red corners, on the other hand, are “mismatched” stocks—high (low) B/M stocks primarily held by funds that track low (high) B/M returns. Well-matched stocks tended to outperform mismatched stocks in their study. 

A Quick Look at Last Quarter 

Let’s take a quick look at how this theory played out last quarter using some slightly different methodology. While Peng and Wang used mutual fund returns and factor returns to calculate each fund’s factor demand, we’ll look at the institutional-level-reported 13F holdings that were held on June 30, 2022. We do this rather than those at the fund level as an alternate take, with the idea that institutions with multiple funds may rebalance among funds internally and ultimately reduce the pricing pressure implied by the study. For each 13F available, we calculate the book-to-price (B/P) ratio for each equity stock held, and aggregate up to the institution level, weighted by market value of the equities. We then do a 5x5 double sort of the stocks, by the B/P quintile of the stock itself, and the holdings-weighted B/P of the institutions that held them. 

For curiosity’s sake, here were the five largest companies alongside where they landed in the 5x5: 

Figure 2: Five Largest Stocks for June 30, 2022 

Source: FactSet (Data as of September 30, 2022) 

All of them are growth stocks with low B/P ratios, which puts them in stock quintile 1; they are also primarily held by institutions with low aggregate B/P ratios. These stocks are all well matched with their owners—three of the five had positive returns, making them overall positive on average. Let’s now look at ten large mismatched stocks. 

Figure 3: 10 Large Stocks, Mismatched with Their Holders on June 30, 2022 

 Source: FactSet (Data as of September 30, 2022) 

Note that quintile 1 indicates stocks/institutions with a low B/P ratio, and 5 indicates high B/P ratio. As expected, most of these stocks did underperform over the past quarter. 

Finally, let’s look at the overall results for each of the 25 portfolios were over the past quarter. These hypothetical portfolios are equally weighted and only include equities with market cap over $2 billion. The returns are also demeaned by stock quintile to account for the fact that most stocks were negative (growth stocks in particular) over the quarter. Note also that these results are for one quarter, is highly not statistically significant, and is solely presented for the sake of interest: 

Figure 4: Equal-Weighted Portfolio Total Return, Demeaned by Stock Quintile

Source: FactSet (Data from June 30–September 30, 2022) 

Final Thoughts 

While our analysis shows that past performance is not a guarantee of recent results, there are other observations and things to consider before conducting your own research and potentially altering your research process. Some practical considerations may make this difficult to incorporate into an active trading strategy.  

It’s important to note that Peng and Wang’s academic paper[2] simply aims to measure the impact of fund rebalancing and factor loadings and considers the holdings as of quarter end and the subsequent returns from that point. Since 13F holding reports are not available to the market as of the quarter-end date and typically aren’t filed until over two months after quarter end, the impact of the rebalancing may already be played out by the time the filings are available to inform a trade. After accounting for reporting lags, subsequent research may be needed to test this as a strategy.  

[1] Peng, Cameron and Wang, Chen. “Factor Demand and Factor Returns,” Oct. 3, 2021. Available at SSRN: https://ssrn.com/abstract=3327849 or http://dx.doi.org/10.2139/ssrn.3327849. 

[2] Peng and Wang, 2021.

This blog post is for informational purposes only. The information contained in this blog post is not legal, tax, or investment advice. FactSet does not endorse or recommend any investments and assumes no liability for any consequence relating directly or indirectly to any action or inaction taken based on the information contained in this article.