“Juicing,” according to the authors, is a strategy whereby mutual funds purchase stocks prior to stock dividend payments in order to artificially increase fund dividend yields. The authors examine the implications of juicing for alternative explanations of investor demand for dividends; they also examine the potential impact of juicing on fund returns.
What’s Inside?
According to Miller and Modigliani (Journal of Business 1961), firm dividend policy should be irrelevant to investors. Subsequent theories acknowledge what investment professionals often see in practice, which is that at least some groups of investors demand dividend-paying assets.
The authors test the clientele and catering theories of investor demand for dividends by examining an equity mutual fund strategy that the authors call “juicing.” When a fund juices, it purchases stocks before the ex-dividend day, collects the dividends, and then sells the stocks in order to artificially increase the fund’s dividend yield.
How Is This Research Useful to Practitioners?
The authors examine dividend-paying equity mutual funds and find that in 7.4% of fund years, funds paid dividend distributions more than twice the size of the distributions implied by their quarterly holdings. The authors also find that the tendency to have high excess dividend ratios is persistent; funds that juice in one year are much more likely to juice in other years.
Juicing is associated with significant capital inflows. Funds that have an excess dividend ratio greater than 1.38 (the authors’ chosen threshold for juicing) receive on average an additional 6.8% in annual flows relative to comparable nonjuicing funds. Funds that have an excess dividend ratio greater than 2.00 (an extreme version of juicing) receive an additional 12.2% in flows relative to nonjuicing funds.
Juicing seems to appeal to investors who are less sophisticated. The authors find that juicing is significantly less likely for funds with an institutional share class and more likely for funds with higher expenses. The authors note that a greater prevalence of juicing among retail funds is also inconsistent with a dividend—specifically tax—clientele effect because retail fund investors are more likely than institutional fund investors to pay an income tax on dividends.
According to the authors, the higher capital inflows and lower investor sophistication associated with juicing funds are more consistent with the psychological desire for dividends posited by Baker and Wurgler’s (Journal of Finance 2004) catering theory of dividends.
The authors find that juicing mutual funds do not have higher returns relative to comparable nonjuicing funds and by some measures perform worse. They also show how juicing can be costly to investors. Juicing funds have higher portfolio turnover than comparable nonjuicing funds, and they can also generate additional annual tax costs for investors.
How Did the Authors Conduct This Research?
For 1990–2011, the authors obtain data on US domestic equity funds from CRSP and primarily quarterly mutual fund holdings data from Thompson Reuters. They impose filters to eliminate erroneous holdings observations. In addition, they require that each fund’s total market value of holdings be equal to at least 90% of total net assets. The authors restrict their analysis to funds for which the actual paid annual dividend yield is greater than 0.5%, resulting in a sample of 2,224 funds with a total of 9,418 fund-year observations.
The authors construct an excess dividend ratio, which measures the ratio of the actual total annual dividends received by the fund to the annual dividend levels implied by the fund’s holdings. They generate a simulated bootstrap distribution of excess dividend ratios and select an excess dividend ratio of 1.38 as a rough cutoff to identify juicing funds because it is the 99.5th percentile’s median value.
The authors use logit regression when the dependent variable is categorical (e.g., dummy variable Juice138); otherwise, they use linear regression. They selectively log transform independent variables and use t-statistics to test for significance.
The authors test for robustness by recalculating their descriptive statistics on juicing using multiple alternative assumptions about how to attribute dividends. They also repeat their analyses of the determinants and persistence of juicing for the full set of funds rather than just those with a dividend yield of more than 0.5%.
Abstractor’s Viewpoint
The authors effectively leverage equity mutual fund information to shed light on questions regarding investor demand for dividends that cannot easily be answered by focusing solely on the dividends of publicly traded companies. They test the merits of two contrasting theories, but in the testing they also generate findings with implications for mutual fund performance.