Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Quantopian Lecture Series: Position Concentration Risk

One very central concept in finance is that of diversifying. This notebook goes over exactly why diversification can reduce risk, and how mediocre components can add to a whole better than the sum of the parts. The main takeaway lesson is that you want to be invested in many many assets when running a strategy.

All the lectures can be found here: https://www.quantopian.com/lectures

Disclaimer

The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by Quantopian. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. No information contained herein should be regarded as a suggestion to engage in or refrain from any investment-related course of action as none of Quantopian nor any of its affiliates is undertaking to provide investment advice, act as an adviser to any plan or entity subject to the Employee Retirement Income Security Act of 1974, as amended, individual retirement account or individual retirement annuity, or give advice in a fiduciary capacity with respect to the materials presented herein. If you are an individual retirement or other investor, contact your financial advisor or other fiduciary unrelated to Quantopian about whether any given investment idea, strategy, product or service described herein may be appropriate for your circumstances. All investments involve risk, including loss of principal. Quantopian makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances.

3 responses

Delaney.

Just several simple questions:

From practical point of view:
How big should be the number many many if assets are in the same asset class and in different asset classes?

How big should be the number many many to already diversified assets (DIA, QQQ, MDY, SPY, VTI)?

May be it is enough just use several assets, may be just two but negatively correlated from different asset classes (SPY, TLT, GLD, IYR...)?

It will be very useful to attach several backtests with default slippage and commissions on full market cycle showing practical results.

Practical approach in my opinion is to

  1. develop a simple market (or highly complex) timing model as this a positive return stream (this is our first asset)

  2. run quick sim and find single assets that have low correlation to the market and positive return generating (maybe real estate, some bond funds, any beta neutral etf) - hopefully quantopian can get us foreign data soon here and futures data

  3. develop as many beta neautral strategies as possible that are uncorrelated with eachother

  4. a true hedgefund wouldnt stick to just securities, imo. If you could find a pro gambler who wins 55-60% against noobs at the kiddie table - fund him because hes not correlated with the market most likely - scale might be a problem though

I forsee a heavy dose of kalman filters and cost to short :)

In practice large funds use thousands of assets, especially through long-short equity strategies. In practice of course the number of assets you can trade is limited by your capital base, so you may only be able to afford to trade 100. If that is the case, then it is always preferable to try to minimize correlation between the assets. In fact with fewer assets often you can get lower pairwise correlations, as the more assets you choose the harder it is to find assets that aren't correlated with something you're already holding. This is why hedge funds that are not correlated to any others are so attractive, even if they have poor Sharpe ratios. Of course it's not safe to choose a low number of highly uncorrelated assets, as you open yourself up to high position concentration risk.

In general the approach should be like this:

Taking into account transaction costs and my capital base, how many assets can I afford to trade?
Set this is the number of assets you will trade. Select assets based on your pricing model, but also based on diversification and correlation considerations.

Here is an example of how this might work:

  1. Determine you can afford 200 assets.
  2. Develop a value pricing model using fundamental metrics.
  3. Price all assets via this model, sort into a ranking.
  4. Run some correlation reduction and sector neutrality heuristics on the top 500 ranked assets. This could be a clustering analysis. Pick the top 100 scored with correlation and sector neutrality in mind.
  5. Repeat 4 for your shorts.

I don't have time to put up some backtests, unfortunately. Lots of work to do on a few other projects right now.