A Problem Connoisseur

The ever estimable Lew Burton mentioned on Twitter a while back that many large respected institutions base their decisions on wooly research.

Higher-ups don't appreciate being told they're wrong.

With large asset manager / bank resources, Lew would be open to having every assumption cross checked and debated - uncommon and admirable.

Being super anal is super assuring to your clients.

And yourself!

Even if you don't fully believe in yourself, at least you can believe in your work.

I have my super anal moments - off blog mostly - because this site sucks up enough energy as is, but Lew's tweets led me to think about something else.

Being a problem connoisseur.

If I had large resources, I'd hope to have a large number of problem connoisseurs.

Would they look for the needle in the hay stack? Perhaps, but they wouldn't do it very systematically.

Choosing great tasting problems is probably the most important skill anyone can have. It takes experience and self confidence, however.

For example a while back I looked into a problem to do with geometric returns.

I spent days coming close to a solution. Far too much time wasted. Did this show a good taste in problems or just a way to waste precious time?

Not sure.

The answer was interesting.

Simple arithmetic returns [(t1-t0)/t0] result in a wickedly different formula for volatility.

What's more interesting is that volatility loses its 'physical interpretation'. I.e. if the current price is $100 and the volatility is 10%, you can't equate the volatility to $10 anymore, as the units are lost in the calculation.

The upshot, you should only use logarithmic returns in financial calculations because they are additive and operate as statistical formulae expect.

I squirm every time I see a CAGR, because every number quoted in the analysis apart from CAGR is wrong.

Likely only slightly wrong, but nevertheless.

If you are in the equity business during calm periods don't sweat it; but if you are in the money market fund business and every basis point counts you should probably worry.

That reminds me of a certain pedantic central bank that prided themselves on consistently beating their money market benchmarks. Of course they only beat them because the calculations were slightly wrong.

It's interesting how being so utterly pedantic can lead you to being incredibly myopic. I have so many stories about this (mainly from Asia).

By all means be pedantic but having a nose for a tasty full bodied problem is far more salient.