Data and Supplemental Materials

SAS Code to perform the GRS test:

Materials to accompany "Seasonal Variation in Treasury Returns" (formerly titled "Opposing Seasonalities in Treasury Versus Equity Returns"):

  • Appendices to accompany the paper.
  • Additionally, below I provide data for the Onset/Recovery variable defined in this paper. Use of the Onset/Recovery variable is recommended for settings where you expect seasonal risk aversion to lead to opposite signs on the dependent variable across the fall and winter periods. That is, it is appropriate for studying "flow" variables (as opposed to "stock" variables). For example, use of the Onset/Recovery variable is appropriate to use when considering security returns. (Daily equity returns are expected to be negative in the fall, as risk averse investors shun risky assets. Then daily equity returns are expected to be positive in the fall, as investors resume their former risky holdings. Similarly daily Treasury bond returns are expected to be positive in the fall and negative in winter.)
    • The Onset/Recovery variable (OR) is available in raw text format in monthly or daily form.
    • The Onset/Recovery variable (OR) is available as a SAS dataset in monthly or daily form.
    • Note that in the daily data files, "Julian" refers to the day of the year (from 1 to 365, or 366 in leap years). For non-leap years, simply delete the 366th observation.

Materials to accompany "Seasonal Asset Allocation: Evidence from Mutual Fund Flows":

Materials to accompany "Seasonal Cycles in Bid-Ask Spreads":

Below I provide data for the Incidence variable defined in this paper. Use of the Incidence variable is recommended for settings where you expect seasonal risk aversion to lead to the same sign on the dependent variable across the fall and winter periods. That is, it is appropriate for studying "stock" variables (as opposed to "flow" variables). For example, use of the Incidence variable is appropriate for use when considering bid-ask spreads. When considering bid-ask spreads, one expects seasonal depression to lead to a spread between dealer bid and ask quotes that is wider during both the fall and winter seasons than during the rest of the year. As such, the Incidence variable, which has the same sign across the fall and winter periods, is appropriate.
  • The Incidence variable is available in raw text format in monthly or daily form.
  • The Incidence variable is available in SAS dataset format in monthly or daily form.
  • Note that for the daily data, "Julian" refers to the day of the year (from 1 to 365, or 366 in leap years). For non-leap years, simply delete the 366th observation.

Materials to accompany "Winter Blues: A SAD Stock Market Cycle":

  • Appendix containing sensitivity checks and a table of gains to various pro-SAD trading strategies
  • If you are interested in estimating the effect of seasonal risk aversion on equity returns, I recommend that you use the Onset/Recovery variable we introduce in our "Opposing Seasonalities in Treasury Versus Equity Returns" paper instead of the obsolete sine wave and Fall dummy variable specification that we employed in our original "Winter Blues..." paper. It yields virtually indentical results without relying on an ad hoc dummy variable.

Materials to accompany "Volatility Forecasts, Trading Volume and the ARCH vs Option-Implied Volatility Tradeoff":