# Monthly average rate of discount, 3 month Treasury bills, Sterling Vs 1 Month

I have been advised by one of my professors to use UK treasury bills for a risk free rate when calculating expected returns for stocks.

I wanted to know whether I should be using 1 month T-Bills or 3 month T-bills.

1. Which is better to use and why?

2. My lecturer mentioned that the risk free rates I download will be annualized percent figures (although % notation may or may not be there), so I would need to convert them to daily rates before I can you use them in regressions. Does anyone know where I can obtain this data? I found the 1 month and 3 month T-bills data on the Bank of England website. I need to know how to get the daily data for these T-bills.

• Is it not the case that the shorter the maturity the lower the risk? In any case, it is unlikely the UK government will ever default on its 1M or 3M T-bills (the last time it was rescued by the IMF) so they probably have a very similar risk. I would rather use whichever has more data. – luchonacho Aug 17 '17 at 20:28

Answer to your first question Take a look at what Aswath Damadoran uses for his empirical work. He's been teaching applied corporate finance for a little while at NYU. These are his slides.

• Datastream
• Bloomberg

Also try FRED database

There are different oppinions on which rates to use for calculating expected return (I assume you mean throu CAPM?).

Most, however, do agree that it should be long-term risk-free, so 1 and 3-month T-Bills are going to be way to short a timespan. Usually you would go with 10 to 30 year government bonds.

My professor in B-School recommended 10-year government bonds for a similar case.

• I agree with you, 10 to 30 year would be much more ideal.However I am dealing with a 2 year window (2015-2017). I am working on the FTSE all shares index returns before and after Brexit. With estimation window from the start of 2015 up till 1 month before the referendum and then the second window is post referendum up till now. I am working with daily data. – DaveSwans Aug 17 '17 at 12:12
• If you are dealing with historical figures - then I guess you should use historical rates for 10 year bonds back in 2015 to 2017. Consider if your 2-year timeperiod was in 1994-1995, then using todays rates would be wrong - same thing here, expect the historical figures are not that old. – ssn Aug 17 '17 at 12:16