The Concept of Correlated Risks

[Parsad]

This probably belongs in the crypto thread but still a very good example of how institutions (old and new) forget or underestimate correlated risk.

 

https://finance.yahoo.com/news/winklevoss-twins-big-mess-crypto-205838099.html

 

I think Buffett and Ajit Jain's greatest strengths compared to virtually all other investment managers or insurance managers is this complete understanding around, and conviction to avoid, excessive correlated risk.  

 

One of the most amazing things I've ever seen was when Berkshire acquired GenRe, Buffett instantly began dismantling thousands and thousands of derivatives contracts...avoiding the blowups seen in 2008. 

 

He also unwound their position in Fannie Mae and Freddie Mac years ahead of everyone else as well, after they watched Fannie and Freddie start to do things where they had forgotten about correlated risk and excessive leverage.  

 

In 70 years of investing, I don't think Buffett has ever underestimated this one thing!  Thus the reason why there are very few down years and those are usually insignificant.  Cheers!

[UK]

1 hour ago, Parsad said:

This probably belongs in the crypto thread but still a very good example of how institutions (old and new) forget or underestimate correlated risk.

 

https://finance.yahoo.com/news/winklevoss-twins-big-mess-crypto-205838099.html

 

I think Buffett and Ajit Jain's greatest strengths compared to virtually all other investment managers or insurance managers is this complete understanding around, and conviction to avoid, excessive correlated risk.  

 

One of the most amazing things I've ever seen was when Berkshire acquired GenRe, Buffett instantly began dismantling thousands and thousands of derivatives contracts...avoiding the blowups seen in 2008. 

 

He also unwound their position in Fannie Mae and Freddie Mac years ahead of everyone else as well, after they watched Fannie and Freddie start to do things where they had forgotten about correlated risk and excessive leverage.  

 

In 70 years of investing, I don't think Buffett has ever underestimated this one thing!  Thus the reason why there are very few down years and those are usually insignificant.  Cheers!

 

https://www.wsj.com/articles/crypto-lender-genesis-lays-off-30-of-staff-11672939434?mod=hp_lead_pos4

 

Massive crypto lender Genesis Global Trading Inc. laid off 30% of its staff and is considering filing for bankruptcy, according to people familiar with the matter, the latest sign of financial turmoil at the crypto lender. The layoffs weren’t confined to one department and were across the company, some of the people said. Genesis has 145 employees left after Thursday’s layoffs.

[Dynamic]

Late to this thread, but I totally agree.

 

Berkshire pays a lot of attention to correlated risks. I think Charlie Munger is another strong influence on this matter and I also get the impression that Todd Combs had this deeply ingrained.

 

I believe the attitude to risk management was actually one of Warren's key criteria in assessing his perspective investment manager hires, and I always interpreted it to include understanding and avoiding excessive correlated risks. By that token it is probably also very true of Ted Weschler, but I just haven't read or heard things from Ted to be able to assess it yet.

 

I'm certain that Warren Buffett and Ajit Jain have a keen focus on limiting correlated risks in insurance and reinsurance policy exclusions and contact terms, though they're willing to underwrite substantial well defined risks that don't endanger Berkshire's financial strength and independence if they're adequately compensated.

 

I try to take the same approach to my portfolio, not avoiding positions with large upside but a real potential of total loss, but ensuring they are limited in size and uncorrelated and cannot cause our ruin.

[SharperDingaan]

Portfolio theory uses the variance-covariance between stocks in a portfolio, the expected return of each stock, and the risk-free rate to determine the optimum portfolio and the individual stock weightings within that portfolio. Traditionally, the variance-covariance matrix has been determined from a rolling history of the most recent 60 months of data. 

 

In the 16 years since the Great Recession of 2006, there has been so much ANNUAL fundamental and material change in the rolling history, that the approach has become useless. If you hadn't realized/adjusted for this, your 'optimized' portfolio has been anything but 'optimal' - and for quite some time. Crypto portfolios being a prime offender.

 

Today, the variance-covariance matrix is determined from option volatility, and looks entirely forward; history is a very minor component. However the same portfolio, looks very different to what it used to look like, and the more change in the rolling 60 month history - the greater that difference. Post Beijing Olympics and the domestic rollout of e-CNY, a portfolio using the 60 month average, and a high weighting of BABA has underperformed - largely because it hadn't recognized the change that e-CNY created.

 

The reality of course is that when the sh1te hits the fan, all correlations tend to 1.000. However, what is unstated; if that if you win big, you also end up having to take equity in the counter party which would otherwise fail. Minority ownership can be a bastard if you don't have scale. 

 

You still have to think.

 

SD 

 

  

Edited February 12, 2023 by SharperDingaan