Liquidity and investor confidence in the turn-of-the-month regularity

ABSTRACT

Ogden (1990) offers a compelling explanation for the ubiquitous turn-of-the-month (TOM) seasonality. He hypothesizes and shows that the clustering of payment dates at the end of the month results in a stock return regularity that is related to increased liquidity and monetary policy. This article introduces investor behaviour into Ogden’s TOM liquidity hypothesis where higher TOM returns depend not only on the availability of increased liquidity but also on investors’ willingness to invest new funds. The empirical evidence is consistent with the argument. When confidence is high, investors’ willingness to invest the increased liquidity results in a TOM regularity. But when confidence is low, a TOM regularity is absent as investors park the increased liquidity. This additional measure of investor confidence provides a more complete explanation of Ogden’s liquidity hypothesis.

KEYWORDS:

• Turn-of-the-month
• seasonal pattern
• anomaly
• confidence

JEL CLASSIFICATION:

• G12
• G14

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ McConnell and Xu (2008) argue that if Ogden’s hypothesis is correct, then trading volume and/or flows of funds into equity mutual funds should be higher, on average, for the TOM over the ROM. They find no evidence of higher trading volume at the TOM and no discernible pattern in net fund flows across the sample period. However, the arguments of both Ogden (1990) and this article do not necessarily imply higher trading volume at the TOM across the entire sample period. Both hypotheses imply that only in months in which liquidity and/or confidence is high will investors commit new funds to the stock market at the TOM, resulting in an increase in returns. In the months of low liquidity and/or low confidence, investors may not invest in the stock market. Hence, one should not expect to see higher trading volume and/or fund flows across all months.

² Naturally, there are other heuristics for making this categorization. The 25th and 75th percentiles were chosen because they capture distinct differences between high and low confidence regimes. They also capture expected differences. Specifically, the high confidence months tend to be between mid-1987 and the end of 1989, and mid-1996 to mid-2001. The low confidence months are at the end of 1973 to mid-1975, between April 1979 and March 1983, and mid-1990 to the end of 1993.