Treasury rates no longer predict returns: a reappraisal of Breen, Glosten and Jagannathan (1989)

Research output: Contribution to journalArticleResearchpeer-review


Breen, Glosten, and Jagannathan (1989) show that the negative relation between excess stock returns and Treasury bill rates is economically important. From 1954 to 1986, the predictive ability of interest rates facilitated a trading strategy that generated average returns at least on par with a buy-and-hold market investment but with significantly lower risk. The services of a portfolio manager using this predictive model to invest justified a management fee of nearly 2% per annum. Using currently-available data, we can nearly perfectly replicate Breen et al.’s (1989) key findings in sample. However, the success of Treasury bill rates as a predictor of equity returns appears to be specific to the time period studied. When the same methodology is applied out of sample from 1987 to 2018, there is little statistical or economic evidence of predictability. Additional out-of-sample analysis of G20 countries shows only sporadic support for the notion that interest rates predict equity returns.
Original languageEnglish
Number of pages16
JournalCritical Finance Review
Publication statusAccepted/In press - 2019


  • Return predictability
  • Treasury bill rates
  • Trading strategy
  • out-of-sample forecasts

Cite this