I want to know if anyone has tried to create a market-neutral portfolio. Since P123 does a pretty good job picking stocks, why not neutralize the effect of a down market by shorting the S&P index or buying puts. Of course there is a cost but it seems to me the reduction in risk would be worth it. You think?
Werner Ganzz (a retired engineer) has a great free web site that focusses on hedging mutual fund and stock portfolios. The man is a real genius in this area:
That is a great link - thanks. Very interesting comments about small caps.
My logic has been that a good ranking system should allow for pairs trading. If your system truly has a correlation between rank and return, you should be able to make risk-neutral profits in a sector or industry by buy high-ranked in the industry and selling low-ranked in the industry.
I like pairs better than market-neutral because that way not only are you managing beta, or market risk, but also some of the volatility specifically related to industry or sector risk.
I’ve tested my ranking systems in all the sectors, although haven’t had the time to test at the industry level. My 40% p.a. long strategy (w 30 stocks) results in about a 10% return and almost a complete elimination of beta with pairs trading.
But, I haven’t implemented because I am looking for larger returns and am young enough to accept more risk. Risk, to me, has to be defined within a time horizon. Over a 25 year time horizon, volatility can be overcome. Over a 2.5 year time horizon, it may not.
I am hedging my port by shorting RUT (Russell 2000) futures.
Using different simulations, i found that hedging ports by shorting approx. 2/3 of its value often results in an almost market neutral portfolio. Of course hedging reduces annual gains, but this can be compensated by leveraging the investment using margin.
I found that for most ports hedging the market risk combined with some leveraging, the original annual performance can be achieved, while drawdowns are cut in half.
Lets look at the GARP Smallcap port as an arbitrary example:
http://www.portfolio123.com/port_summary.jsp?portid=29106
This port has an annual gain of 29% (as of 12/1/05) and a max drawdown of -25% (tracking daily values).
Hedging 66% of the ports value by shorting RUT you get an annual performance of 18% and a max drawdown of -8%.
You can now leverage this market neutral port by factor 1.6 (60% margin) and get the original annual performance of 29% with a max drawdown of -12.6%.
Using 50% margin, you end up with 37% anual perf and a max DD of -14.5%.
Above numbers include the 5% margin required for the futures and margin financing cost.
For me it´s also important how long it takes to recover from a drawdown. In the above example the unhedged port needs 278 trading days to recover from its worst DD, the hedged port needs 178 trading days.
This approach certainly is not the perfect hedge, but it´s quite easy to adopt.
I download all simulations of portfolio that i use for real money in excel, and combine the results.
I have annual gain of 130%
and a maximum drawdown of about -17%
When hedging with shorting 1 contract of russell2000 future and if my portfolios worth the same amount as the contract (about 70 000$)
My annual gain became 115%
Maximum drawdown -13%
The optimal result i found is :
- Shorting 1 contract of russell 2000 future
- Portfolio worth 1.66 * amount of contract
My annual gain became 122%
Maximum drawdown -9.5%
Azouz
Day Trading Systems
azouz and all,
maybe i forgot to note that one should try to keep the hedge-ratio constant as you port value growths. depending on the invested value this may not possible with 70K value futures, so options or leveraged ETFs might be an alternative.
hees,
what is your margin financing cost? do you have a good cost of carry benchmark (ie 1year treasuries, etc) for both futures and borrowing for equities?
Current margin financing cost is 4.25% + 1% = 5.25%
Margin requirements for one RUT contract are $3500 (5%).
For futures cost of carry is negative if you short them. So hedging is free as beer 
I attached an excel sheet which helps you to simulate a portfolio, which is hedged against an index.
It includes data for the public ports GARP Top Ranked, Diversified Low Priced Growth, Balanced Mid Cap, Value w/Momentum and index data for Russell 2000, S&P500, Nasdaq.
For all above ports hedging reduces max DD and volatility, while annual return is little affected.
The sheet allows you to use it with your own simulation results (performance->download->total-column). You can easily alter hedge ratio, margin/leverage, financing cost, contract size, etc.
I also have a few questions to the community:
Could anyone spot errors in the calculations, assumtions or values?
If hedged ports offer superior risk/reward ratios, shouldn´t we suggest to hedge ports and adjust exposure as best practice?
How can i calculate sharpe ratio in excel, to be comparable with the pf123 values? (Problem with compounding if transforming to the same frequency).
Which risk/reward ratios do you achieve with your 100+% ports if you hedge+leverage them?
Do you have any questions, i can answer?
hedge_demo.xls (1.1 MB)
hees,
Thank you for posting your spreadsheet. How did you arrive at 66% as the percentage of your portfolio to hedge? How does the spreadsheet support your statement that your annual gain is reduced by 15% and your drawdown by 4%?
Regards,
Sam
The 66% is a rule of thumb ratio which worked well with most simulations. Select the index with the highest beta as your hedging vehicle. The spreadsheet shows the beta between the selected sim and index. This beta is a good starting value for a hedge ratio.
You can fine tune this ratio by changing it´s value and looking at the resulting beta, annual return, deviation and max DD values.
I think simply hedging by a ratio which results in an almost zero beta port is a good deal. Other hedge ratios can result in a slightly better risk reward ratio but might be fitted for the bear market in 2001/2002.
Regarding the statements of annual gain reduced by 15% and drawdown by 4% you have to ask Azouz 
As I mentioned you can paste your own simulation results in a spreadsheet column and see for you self how hedging works on your selected sims. Just drop a message if i may assist you.
When i hedged all my portfolios with one russell2000 contract, my annual gain was reduced by 15% and drawdown by 4%.
Its all deponds on your portfolios.
I did like hees, i make a spreadsheet with all my data plus russell2000 data and found these results.
Some questions re hedging a portfolio:
-
If one perfectly hedges a portfolio 1:1 with a long portfolio and a short device(option, future or Rydex type short fund) how does one make net money?
-
When hedging, are you constantly hedging, or do you attempt to do it with some type of market timing? That is, you implement your short hedge when you become concerned about the market using whatever technique you like.
-
Should one use hedging to reduce volatility (standard deviation re sp500) or…another standard?
-
When designing a portfolio vs the short hedge should alpha be used in the analysis? There is another form of alpha, NCalpha that uses standard deviation which might be better for hedging.
-
If one wanted to use a Rydex fund, should one consider one that tends to lose at a similar rate to sp500 or twice the rate(reduce the amount you purchase). One advantage of the fund may be that you do not have expiration dates.
thanks for any comments
Jim Hill
Did you you keep the hedge ratio constant as your port increased in value? If you bought just one contract, your hedge ratio decreases as your value grows and protection against market risk fades.
In the long term, pf123 ports seem to have a higher annual return than the overall market. Nonetheless in the short term the ports correlate with the market (betas of .50 to .80). If you hedge against market risk, you invest in the difference between the development of the market and your port. The resulting investment probably has lower annual returns than the port, but it also has lower max DDs and a lower volatility. In theory, risk reward ratios of such an investment should be better than those of an unhedged port. Further one could extend the part of invested money or even trade on some margin and get a higher annual return for a given risk tolerance in comparison to an unhedged portfolio.
Since i don´t know anything about the direction of the overall market for sure (otherwise tomorrow i would be a zillionaire), i prefer constant hedging. I tried to to use some market timing signals (http://www.timingcube.com) to adjust the hedge ratio with the market sentiment and this resulted in a better performance. But i´d fear to fall into a survivorship bias trap if i trust signals from one of those timing signal sellers.
In the first place one can use hedging to reduce volatility and max DD. I use it to construct an investment with a better risk reward ratio and decide how much money to put on it.
While alpha says nothing about the risk you take, NCalpha is return in excess of the market´s return normalized by standard deviation.
This measure can be usefull independend from hedging. Comparing funds, it tells one wether a fund has greater returns since it is exposed to greater market risks (beta>1). Hedged portfolios should result in better NCalpha and sharpe ratio values.
The money one puts on the hedging product can not be invested in the portfolio. Therefore a double inverse fund is the better choice (2x leverage) leaving 66% of one´s money to be invested long. Shorting with futures is even better since these investments are leveraged by factor 20 -even more money (95%) remains on the long side.