Journal 2

aj165602

I finished my first journal at the stage where I am comfortable using the native CloudQuant lists. I think it's now a good place to start phase 2.

The main departure is that, as is common in the financial economics literature, I will start using a double-sort methodology to purge the portfolio of size / beta risk, while maintaining dollar-neutrality.

We don't have easy access to market cap. data, but I have found that the stored beta values are very effective, so in the first instance I will sort the universe in two along that dimension.

aj165602

I've started building a single-factor model, and while I no longer actively manage beta, I can still print it to the console screen. When the Detailed Trading Report is back up and running, the report also measures beta risk.

Starting with a single factor that has quite a large beta of -0.14, I am pleased to see that the portfolio beta is reduced to -0.05 by using a double sort.

At first, I am going to attach equal weights to each sub-universe.

aj165602

Unfortunately, this ended up not offering any advantage over the way I was originally doing it (as described in the old journal), and the computational time goes up significantly.

Nevertheless, an interesting exercise that didn't take too long.

It looks like the business model of CloudQuant has changed towards being an alternative data showcase, so it also seems like a sensible place to conclude using the CloudQuant platform. Best wishes to all.