@AC I remember you mentioned that the fund risk rating in Dozens invest are calculated based on their historical volatility in the virtual meeting between community members and Dozens. At the time of the meeting I didn’t think too much about it, but now I’m actually thinking about it, the method itself is not bad, but judged by the outcome of the risk ratings, Dozens might have done something really wrong with the method.
Let’s go back to the example I mentioned in the virtual meeting - the US Treasury Bonds (Vanguard USD Treasury Bond UCITS). I remember I said that this fund is going to have equity-like volatility and bond-like return, and you corrected me saying that the fund risks are rated based on their historical volatility, which means a fund like what I’ve described wouldn’t have the risk rating 2 of 5. This sounds sensible without thinking too much about it. However, as soon as I checked the index the fund’s KIID, I realised that the fund only had 4 years of history returns (since 2017), and they are all positive without too much swings. This immediately made me worried about how did you calculate the volatility of the fund.
My first reaction to that is, you could have calculated it based on the available history of the fund itself - 4 years of data, which will be very very wrong. To proof whether I’m right about the assumption, I pulled the data of the intermediary US Government bonds data from here (ITT in the spreadsheet), which is the data closest to the index the fund tracks. I have also pulled the exchange rate data from here in order to convert the data to British Pounds for UK based investors.
After analysing the data between 1950 and 2020 - 71 years of data, I got the following results:
- Standard deviation in USD is 7.2 and in GBP is 12.6 - three-quarters higher than in USD
- Max 1 year drawdown in USD is -4.9% and in GBP is -7.9% - more than three-fifth worse than in USD
- Max drawdown (over 1 or more years) in USD is -4.9% and in GBP is a staggering -12.3%
- Yet, the 71 years annualised nominal return in USD is 5.64%, and in GBP is 5.67% - only 3 basis points difference!
Off-topic: unless the investor has future USD liabilities, I cannot think of a single reason for someone in the UK to invest in the currency unhedged USD government bonds. The investor would be taking uncompensated currency risk if they do that, and the result would be exactly what I had described - equity-like volatility and bond-like return.
Now, let’s get back to the topic. I don’t think Dozens will rate a fund at risk 2 of 5 if you knew the fund’s value had dropped nearly one-eights in the past. If my assumption is correct and you are calculating the volatility based on short term data, I would worry many “low risk” funds on the Dozens invest shelf are suffering from the same issue. If the above is true, I would strongly recommend you to reevaluate the risk levels of all funds in Dozens invest which has a risk rating at 4 or below using a much longer time frame. 30 years is a good starting point. An appropriate alternative should be used for funds with a short history. Many funds may be riskier than what you had initially thought.