What has your hit and miss ratio been?

[sinner]

There are different types of risk: idiosyncratic, secondary, systemic.

Diversifying a bunch of stocks within a sector allows you to harvest volatility differences between stocks in that sector, but you remain exposed to secondary and systemic risk.

Diversifying a bunch of sectors allows you to harvest volatility differences between sectors in a given (international or national) market, but you remain exposed to systemic risk.

Diversifying a bunch of asset classes allows you to harvest volatility differences between asset classes and does protect well from systemic risk.

In each superior case you reduce overall portfolio volatility versus its inferior at the cost of beta returns, but compensated by low volatility alpha returns (i.e. high win rate trades - the US Permanent Portfolio linked above "delivered positive returns over 98% of periods since 1970").

If you think about it, it also protects you from bubble risk, as you will never hold your largest allocation for any given investment when the market is rising (i.e. valuations falling) - admittedly this does depend on your rebalancing strategy, whether it's price or time or volatility based.

Here's another blogpost I've posted about before, on the difference between convex and concave trading strategies and when each outperforms

http://cssanalytics.wordpress.com/2...et-allocation-lessons-from-perold-and-sharpe/

The first takeaway is that there is no uniform winning strategy in all market conditions. Each strategy has a particular regime in which is it likely to shine.

 

[sinner]

Turnkey Analyst (http://alpha.turnkeyanalyst.com) has a backtesting tool for some simple and academic long only portfolios going back to 1978. Below I highlight a few different ones:


ew_index == Equal Weight 6 asset class (SP500, MSCI EAFE, MSCI Emerging Mkts, FTSE NAREIT, GSCI, 10Y US Treasury rolling) basket with annual rebalancing to equal weight
ew_index_ma == Above with (monthly evaluation) trend following rule
mom_ma == Above with trend following rule and (monthly) weighting asset classes by momentum
risk_parity_mom_ma Above and weighting classes by both momentum and volatility


In this case I think the results are pretty self explanatory, in this case rebalancing wasn't useful for the GFC (since most of the assets in the portfolio are 'risk assets') but it did a good job of protecting from the tech bubble.

 

[sinner]

Another similar example which shows protection from secondary risk, SP500 (which is a market weighted index) versus the equal weighted (annual rebalance) CRSP universe (unfortunately this test only goes to Dec 2011)

 

[Ves]

Thanks Sinner. The GesaltU blog has made for some good reading when you have linked to it in the past. This time is no exception.

 

[DeepState]

(unfortunately this test only goes to Dec 2011)

Hi. Popping in quickly. Just want to fill a data hole. Below is an extract of performance from Fernholz's (Originator of Stochastic Portfolio Theory that went on to become re-bal) firm. It includes rebal over SP500 that shows modest outperformance for the three years to 31 March 2014 which overlaps the missing data.

InTech, which has been in existence for 25 years, has never closed a portfolio for reasons of performance so this is all their accounts that were ever formed and had external assets in them. It covers different universes, including global equities and small cap equities. Thus, no selection/survival or hindsight type bias exists in this data. All records are live composites. The firm has ~USD50bn. Not one single strategy has underperformed buy-hold in the period since inception. Before advisory fees, outperformance of buy-hold is around 2%pa as previously indicated as indicative for the strategy in general. Those which had access performance of around 1% per annum were run to tighter bounds, which limits misweighting, and hence called "enhanced".

Given the length of the history, essentially all sorts of market conditions have been encountered.

Enjoy.



Disclaimer/Disclosure: I do not endorse InTech, it's officers, subsidiaries or associates. I have had a sandwich lunch with their CEO and a one course meal with their President of International Division and received Christmas cards and presents of nominal value. Otherwise I have no financial relationship with them whatsoever. This is not advice nor a solicitation to invest in these products. Please do your own work.

 

[rb250660]

I have had a sandwich lunch with their CEO and a one course meal with their President of International Division and received Christmas cards and presents of nominal value.

Who cares.

 

[DeepState]

Who cares.

It's a joke dude, despite it being factually accurate.

However, in reality, standards of disclosure have reached the point where meals are itemised and the value of the meals are also disclosed when people are interviewed in newspapers and conflict of interest registers in professional organisations providing advisory services. So given I've been asked about disclosure here and there etc. I went OTT or, more accurately, met those standards.

Personally, I wouldn't care either. It's not like the tandoori chicken wrap ($4) or seafood risotto ($25) were out of this world. Touche.