12 Comments
User's avatar
Martin Schwoerer's avatar

yes, you have my attention, so I'll be watching this one as time goes bye. Good catch!

(It's too young and unproven to spend any serious money on, though...)

By the way, as I always like to simplify, here's my version. CAGR 12.98%, max drawdown -5.41%.

https://tinyurl.com/2tdckdek

Expand full comment
Jaewon Jung's avatar

Cool, thanks for sharing!

Expand full comment
Martin Schwoerer's avatar

A propos of not much, I tried out some other MF ETFs, some of which started back in 2012 or 2015.

For example, WTMF, FMF and DBMF. Results were generally OK, with MARs not much lower than 1. However, performance in '15, '16 and '22 was slightly or somewhat negative in all cases.

As this data does not include performance results from all business cycles (e.g. a deflationary recession), I remain sceptical, but we'll see how this plays out!

Expand full comment
Jaewon Jung's avatar

Yeah, I have a similar sentiment toward it. One difference of KMLM is that it doesn't invest in the equity market.

Expand full comment
okhi2u's avatar

I recall we had tried to get you to setup a synthetic KMLM based upon old historical data. That might help see if this would have worked better over a longer time period or not. The index it's based on goes back to the late 80s or 90s I think. From memory I think you tried, but there was something fishy about the data?

Expand full comment
Jaewon Jung's avatar

Yeah, I actually pinged them and got the index data. But the discrepancy between the index time series and KMLM price time series was quite noticeable when I looked into them (even for recent days). That's when I gave up on the idea of extending it with the index data.

Expand full comment
okhi2u's avatar

Might be worth a shot asking if they can explain why what you see is happening?

Expand full comment
Jaewon Jung's avatar

I already did, but got no response.

Expand full comment
hydromod's avatar

Would the discrepancy muddy the signal significantly in the backtest?

Expand full comment
Jaewon Jung's avatar

I didn't do any numerical analysis, but in my eyeballing, it was big enough differences (two-digit percentage deviations in daily returns) and made me feel uncomfortable to stitch the data.

Expand full comment
hydromod's avatar

FYI there's an interesting podcast episode where the KMLM approach is described in some detail at https://thealgorithmicadvantage.com/podcast/021-dave-aspell-of-mt-lucas-management/

Expand full comment
Jaewon Jung's avatar

Thanks for sharing!

Expand full comment