Wednesday, November 14, 2012

US portfolio recommendation (from 12, November 2012)

The portfolio recommendation is based on two low-volatility strategies: a long-only minimum-variance portfolio and a “130:30” minimum-variance portfolio, which is long 130% and short 30%.

These strategies use advanced Optimization and Statistics techniques to hedge against the estimation risk of the associated models. As a result, they attain consistently better risk-adjusted returns than market indexes, as these portfolio recommendations show.

For more details about the implementation of these strategies, please read the following post: Some efficient low-volatility portfolios: the minimum-variance policy

The long-only and the 130:30 low-volatility portfolios recommended for this week, with their corresponding weights, can be found in this file: US_weights_20121112.csv

Although I recommend a portfolio composition every two months, it is desirable to maintain this composition for a quarter year, and then rebalance with the new composition.
The current long-only portfolio composition contains 21 stocks and has changed a bit respect to the previous quarter (three stocks have been purchased and one has been sold). The turnover is 29% (due to this change and the portfolio growth). On the other hand, the 130:30 portfolio contains 58 stocks and the corresponding turnover is a bit larger: 46%.
Regarding the performance, over the last year (52 weeks), the long-only strategy attained a volatility of 9% (versus 15% of the S&P 500). The volatility of the 130:30 strategy is even better: 8.5%.

The weekly 95%-VaR of the long-only portfolio was 1.8% (versus 3.7% of the S&P 500). The corresponding VaR for the 130:30 portfolio was 1.5%.

The last year annualized Sharpe ratio of the long-only strategy was 1.15 (after proportional transaction costs of 40 bps were discounted). On the other hand, the SR of the 130:30 strategy was 1.6. Finally, the SR of the S&P 500 was 0.69 over the same period.

In the next figure, you can see the compounded return over the last 52 weeks of the three considered portfolios.

Both low-volatility portfolios attain better returns than those of the S&P 500.

But let add information about the risk. The next graph shows the risk-return space for the three considered portfolios.

The red point represents the mean return and volatility of the long-only portfolio over the past 52 weeks. On the other hand, the green point represents the 130:30 portfolio, and finally the blue point represents the S&P 500 index over the same 52 past weeks.

We can see the two low-volatility portfolios have better mean returns than that of the S&P 500, and also their volatilities are better. In this case, we say the low-vol portfolios dominate the index.

I have computed the same risk-return space for every week over the last year, using the same 52-weeks historical method to estimate the mean returns and the volatilities. The long-only and 130:30 portfolios attained a higher return than that of the S&P 500 (77% and 83% of the time, respectively). Moreover, the volatility of both low-vol portfolios was always less than that of the S&P 500.

As a summary, the low-volatility strategies dominate the market index most of the time, showing they attain consistently better risk-adjusted returns.


  1. Hola Javier,

    Great blog by the way. Can you shed light on the difference, if any,
    between Low-Vol vs low-Beta?
    I have just finished reading "Benchmarks as limits to artbitrage:
    understanding the low-volatility anomaly" (Baker, Bradley & Wurgler,
    FAJ, 67(1), 2011) and in it they say "Beta and volatility are highly
    correlated" (p. 47).

    From their tests, they say that Beta is closer to the "anomaly" than
    volatility, at least in large cap stocks as fund managers are usually
    overweight large-caps because of their benchmarks (the article's
    causal explanation).

    Secondly, what do you think of the study done by Timothy B. Riley "Dissecting the Low Volatility Anomaly" which states that low-vol has not been performing in the last decade?
    He says the anomaly is explained by indiosyncratic volatility, not total volatility.
    Not totally sure what it means:
    1. That there is another unknown factor at play?
    2. Last month volatility?

    1. Dear Bruno,
      Thanks for your comments.
      Regarding your first comment, I think this discussion is useful:
      Regarding your second question, who knows? :)
      This paper sheds some light:
      Best regards.

  2. Respected Sir,

    Let me first congrats & thank you for your wonderful post .sir my name is amit,from india, i am a NanoTech.professional,i also wanna to learn analytics for trading,please guide me how to start from scratch as i am novice in this field .your little guidance will change the life of many layman like me.

    sir i can offer free on/offline assistance to any assignment.

    Warm Regards

    1. Amit,
      Thanks for your kind comments. You can start by reading the papers I recommend in this blog. All the relevant information you need is inside these papers…
      Best regars.

  3. My cousin recommended this blog and she was totally right keep up the fantastic work!

    Point Estimation Training