A major component of the Mutual Fund scandals of 2003 was the allegation that certain investors were allowed to engage in the late trading of mutual fund shares. Under the forward pricing rule, trades in U.S.-based open-ended mutual funds are required to be priced at the next net asset value per share (NAV) calculated after an order is received. 1 For funds that calculate NAVs once per day at 4 PM Eastern time (the vast majority), orders received before 4 PM should be priced at the NAV calculated on the day of the trade while trades received after 4 PM should instead be priced at the next-day net asset value. 2 Late trading occurs when investors place trades after 4 PM but still receive the 4 PM price. Late traders can use information revealed after 4 PM to guide their trades: buying fund shares when their current value is greater than NAV and selling when the reverse is true. Doing so allows them to earn expected abnormal returns at the expense of the fund’s long-term shareholders. 3

The independence of editorial content from advertisers' influence is a cornerstone of journalistic ethics. We test whether this independence is observed in practice. We find that mutual fund recommendations are correlated with past advertising in three personal finance publications but not in two national newspapers. Our tests control for numerous fund characteristics, total advertising expenditures, and past mentions. While positive mentions significantly increase fund inflows, they do not successfully predict returns. Future returns are similar for the funds we predict would have been mentioned in the absence of bias, suggesting that the cost of advertising bias to readers is small.

We use mutual fund recommendations to test whether editorial content is independent from advertisers’ influence in the financial media. We find that major personal finance magazines (Money, Kiplinger’s Personal Finance, and SmartMoney) are more likely to recommend funds from families that have advertised within their pages in the past, controlling for fund characteristics like expenses, past returns and the overall levels of advertising. We find little evidence of a similar relationship for mentions in the New York Times or Wall Street Journal. Positive media mentions in both newspapers and magazines are associated with significant future inflows into the fund while advertising expenditures are not. Therefore, if we interpret our coefficients causally, a large share of the benefit of advertising in our sample of personal finance magazines comes via the apparent content bias. The welfare implications of this apparent bias are unclear, however, since our tests suggest that bias does not directly lead publications to recommend funds with significantly lower future returns than they might have recommended in the absence of any bias. In selecting funds to recommend, magazines overweight past returns relative to expenses, and as a group their recommendations do not outperform even an equal- weighted average of their peers. Nevertheless, this approach leaves magazines with large numbers of funds with high past returns from which to select, and so bias towards advertisers can be accommodated without significantly reducing readers’ future returns. Interestingly, the recommendations of Consumer Reports, which does not accept advertising, have future returns comparable to or below those of the publications which accept do advertising.