R: Trouble using SharpeRatio from the PerformanceAnalytics package

C

Cinji18

I'm confused and having issues using `SharpeRatio` from the `PerformanceAnalytics` package. After all was said and done, I can't seem to get my number to match the sharpe ratio posted on Yahoo Finance.

Here's my procedure (I'll try to keep it simple):

> getSymbols("XLU", src = "yahoo", auto.assign = TRUE, return.class = "xts")
> head(XLU)
XLU.Open XLU.High XLU.Low XLU.Close XLU.Volume XLU.Adjusted
2007-01-03 36.41 37.10 36.41 36.82 4229300 23.33392
2007-01-04 36.86 36.94 36.67 36.77 2958400 23.30223
2007-01-05 36.67 36.67 35.99 36.11 3784500 22.88396
2007-01-08 35.97 36.22 35.93 36.09 2415000 22.87128
2007-01-09 35.94 36.18 35.87 36.12 2342100 22.89030
2007-01-10 35.99 36.20 35.95 36.13 1918100 22.89664

> adj <- Ad(XLU)
> head(Ad(XLU))
XLU.Adjusted
2007-01-03 23.33392
2007-01-04 23.30223
2007-01-05 22.88396
2007-01-08 22.87128
2007-01-09 22.89030
2007-01-10 22.89664

> monthlyRTN <- monthlyReturn(Ad(XLU))
> monthlyRTN <- monthlyRTN["2016-08/2019-08"]
> head(monthlyRTN)
monthly.returns
2016-08-31 -0.055448911
2016-09-30 0.003613012
2016-10-31 0.008981693
2016-11-30 -0.054218282
2016-12-30 0.048991012
2017-01-31 0.012559332

> SharpeRatio(monthlyRTN)
monthly.returns
StdDev Sharpe (Rf=0%, p=95%): 0.2386500
VaR Sharpe (Rf=0%, p=95%): 0.1508941
ES Sharpe (Rf=0%, p=95%): 0.1304109

The 0.23865 is nowhere near the 0.63 posted on Yahoo Finance under Risk Statistics under 3 years. What am I missing?

Thanks.
 
You are comparing different periods.

Cinji18 said:
I'm confused and having issues using `SharpeRatio` from the `PerformanceAnalytics` package. After all was said and done, I can't seem to get my number to match the sharpe ratio posted on Yahoo Finance.

Here's my procedure (I'll try to keep it simple):

> getSymbols("XLU", src = "yahoo", auto.assign = TRUE, return.class = "xts")
> head(XLU)
XLU.Open XLU.High XLU.Low XLU.Close XLU.Volume XLU.Adjusted
2007-01-03 36.41 37.10 36.41 36.82 4229300 23.33392
2007-01-04 36.86 36.94 36.67 36.77 2958400 23.30223
2007-01-05 36.67 36.67 35.99 36.11 3784500 22.88396
2007-01-08 35.97 36.22 35.93 36.09 2415000 22.87128
2007-01-09 35.94 36.18 35.87 36.12 2342100 22.89030
2007-01-10 35.99 36.20 35.95 36.13 1918100 22.89664

> adj <- Ad(XLU)
> head(Ad(XLU))
XLU.Adjusted
2007-01-03 23.33392
2007-01-04 23.30223
2007-01-05 22.88396
2007-01-08 22.87128
2007-01-09 22.89030
2007-01-10 22.89664

> monthlyRTN <- monthlyReturn(Ad(XLU))
> monthlyRTN <- monthlyRTN["2016-08/2019-08"]
> head(monthlyRTN)
monthly.returns
2016-08-31 -0.055448911
2016-09-30 0.003613012
2016-10-31 0.008981693
2016-11-30 -0.054218282
2016-12-30 0.048991012
2017-01-31 0.012559332

> SharpeRatio(monthlyRTN)
monthly.returns
StdDev Sharpe (Rf=0%, p=95%): 0.2386500
VaR Sharpe (Rf=0%, p=95%): 0.1508941
ES Sharpe (Rf=0%, p=95%): 0.1304109

The 0.23865 is nowhere near the 0.63 posted on Yahoo Finance under Risk Statistics under 3 years. What am I missing?

Thanks.
More...
 
GRULSTMRNN said:
You are comparing different periods.
More...

Can you elaborate? Even after I use endpoints() to subset end-of-month prices before computing monthly returns, I got the same monthlyRTN, and thus, same SharpeRatio as before.
 
Cinji18 said:
I'm confused and having issues using `SharpeRatio` from the `PerformanceAnalytics` package. After all was said and done, I can't seem to get my number to match the sharpe ratio posted on Yahoo Finance.

Here's my procedure (I'll try to keep it simple):

> getSymbols("XLU", src = "yahoo", auto.assign = TRUE, return.class = "xts")
> head(XLU)
XLU.Open XLU.High XLU.Low XLU.Close XLU.Volume XLU.Adjusted
2007-01-03 36.41 37.10 36.41 36.82 4229300 23.33392
2007-01-04 36.86 36.94 36.67 36.77 2958400 23.30223
2007-01-05 36.67 36.67 35.99 36.11 3784500 22.88396
2007-01-08 35.97 36.22 35.93 36.09 2415000 22.87128
2007-01-09 35.94 36.18 35.87 36.12 2342100 22.89030
2007-01-10 35.99 36.20 35.95 36.13 1918100 22.89664

> adj <- Ad(XLU)
> head(Ad(XLU))
XLU.Adjusted
2007-01-03 23.33392
2007-01-04 23.30223
2007-01-05 22.88396
2007-01-08 22.87128
2007-01-09 22.89030
2007-01-10 22.89664

> monthlyRTN <- monthlyReturn(Ad(XLU))
> monthlyRTN <- monthlyRTN["2016-08/2019-08"]
> head(monthlyRTN)
monthly.returns
2016-08-31 -0.055448911
2016-09-30 0.003613012
2016-10-31 0.008981693
2016-11-30 -0.054218282
2016-12-30 0.048991012
2017-01-31 0.012559332

> SharpeRatio(monthlyRTN)
monthly.returns
StdDev Sharpe (Rf=0%, p=95%): 0.2386500
VaR Sharpe (Rf=0%, p=95%): 0.1508941
ES Sharpe (Rf=0%, p=95%): 0.1304109

The 0.23865 is nowhere near the 0.63 posted on Yahoo Finance under Risk Statistics under 3 years. What am I missing?

Thanks.
More...
Unless you can see how Yahoo finance is calculating the Sharpe Ratio, don't waste your time trying to match their value. Are they using the risk free rate of return in the calculations? Are you (doesn't look like it)? As @GRULSTMRNN mentioned, what is your return period vs theirs?
 
fan27 said:
Unless you can see how Yahoo finance is calculating the Sharpe Ratio, don't waste your time trying to match their value. Are they using the risk free rate of return in the calculations? Are you (doesn't look like it)? As @GRULSTMRNN mentioned, what is your return period vs theirs?
More...

I left out the risk free rate in this example because I wanted to keep it simple for the forum, but I do use it in my actual script. Right now, I'm using monthly returns. I have no idea how Yahoo calculates it. I am using Yahoo to check and confirm I am computing things correctly in my script, but it sounds like that is futile.
 
If you have no idea how they calculate it and what assumptions they make then wanting to match numbers is a useless pursuit. Even a different risk free rate completely distorts results. You may also improperly scale monthly results to annualized figures,not saying you do, just voicing a hunch. Figure out their comoutional assumptions then we can further help, otherwise asking further will not get you answers

Cinji18 said:
I left out the risk free rate in this example because I wanted to keep it simple for the forum, but I do use it in my actual script. Right now, I'm using monthly returns. I have no idea how Yahoo calculates it. I am using Yahoo to check and confirm I am computing things correctly in my script, but it sounds like that is futile.
More...
 
GRULSTMRNN said:
If you have no idea how they calculate it and what assumptions they make then wanting to match numbers is a useless pursuit. Even a different risk free rate completely distorts results. You may also improperly scale monthly results to annualized figures,not saying you do, just voicing a hunch. Figure out their comoutional assumptions then we can further help, otherwise asking further will not get you answers
More...


Thank you. At least that gives me some next steps.
 
Cinji18 said:
The 0.23865 is nowhere near the 0.63 posted on Yahoo Finance under Risk Statistics under 3 years. What am I missing?
More...
I think that the 0.63 is an annual value. So you can't compare it to the monthly value you calculate. It would be better, as a first step, to use the daily returns to calculate an annual Sharpe ratio and see whether that is more in line with your desired Yahoo target value.
 
HobbyTrading said:
I think that the 0.63 is an annual value. So you can't compare it to the monthly value you calculate. It would be better, as a first step, to use the daily returns to calculate an annual Sharpe ratio and see whether that is more in line with your desired Yahoo target value.
More...

Thanks. I will check to see if that is the case.
 
Cinji18 said:
I left out the risk free rate in this example because I wanted to keep it simple for the forum, but I do use it in my actual script. Right now, I'm using monthly returns. I have no idea how Yahoo calculates it. I am using Yahoo to check and confirm I am computing things correctly in my script, but it sounds like that is futile.
More...

What is you risk free rate you are using? Is your base currency USD or which other currency?

Rhetoric question: what RFR would you use for EUR or JPY as base?

And why in the world is Yahoo Finance reporting so, so important to you?
 
You must log in or register to reply here.
Back
Top