Equity long-short strategy, anyone?

[zsng99]

Hi,

I’ve just completed my dissertation for my MSc programme recently. It is a research on the equity long-short strategy using simple fundamental analysis on value-glamour stocks. My research shows that by going long potential winners amongst value stocks and shorting potential losers amongst glamour stocks produces an average return of about 50% per year from 1998 to 2005.

I believe that the strategy could still be improved with more fine-tuning. Coupled with some technical analysis and the right position-sizing technique, the return should be even greater.

I was wondering if my dissertation research would be able to help me land a job in the trading sector. Or if my work is good enough to attract the attention of potential employers? Would anyone out there be interested in my work at all?

Please, I seriously need some advice/feedback from u guys.

 

[tomorton]

50% per year is good enough for anyone. But, what's the worst drawdown? Would it wipe you out? If you don't get wiped out, 50% per year will make you very rich, why do you need a job?

Seriously, you only tested 7 years, and you think that will get you a job?

 

[zsng99]

50% per year is good enough for anyone. But, what's the worst drawdown? Would it wipe you out? If you don't get wiped out, 50% per year will make you very rich, why do you need a job?

Seriously, you only tested 7 years, and you think that will get you a job?

Worst year earns 23.53%, best is 70.99%. I need a job coz I only have the trading idea but without any funds to back myself up. Furthermore, this strategy requires access to some database, which I understand is quite costly. Not to mention is the limits to abitrage problem associated with the strategy.

By the way, the strategy was tested for 8 years, not 7. Anyway, the initial testing period was 11 years, but I'll have to shorten it to 8 years because I'll have to run some regressions using 3 years of historical data due to some unexpected results using the conventional methodology.

I don't really know I will get a job with my research, that's y I post on this forum asking for advice.... Thanks for your advice anyway.

 

[fifty2aces]

Sorry if this comes accross as insulting, but are you aware of the risks of curve-fitting/over-optimization? Have you tested the system on out of sample data?

 

[zsng99]

Sorry if this comes accross as insulting, but are you aware of the risks of curve-fitting/over-optimization? Have you tested the system on out of sample data?

Yup... U could be right. It might just as well be over-optimisation, data snooping etc..
Unfortunately, testing the strategy out of the sample data is just beyond the scope of my dissertation. Would really like to test if it's still working in recent years eg. 2006-2008.

 

[Charlton]

Process - rather than conclusions

Hi,

I’ve just completed my dissertation for my MSc programme recently. It is a research on the equity long-short strategy using simple fundamental analysis on value-glamour stocks. My research shows that by going long potential winners amongst value stocks and shorting potential losers amongst glamour stocks produces an average return of about 50% per year from 1998 to 2005.

I believe that the strategy could still be improved with more fine-tuning. Coupled with some technical analysis and the right position-sizing technique, the return should be even greater.

I was wondering if my dissertation research would be able to help me land a job in the trading sector. Or if my work is good enough to attract the attention of potential employers? Would anyone out there be interested in my work at all?

Please, I seriously need some advice/feedback from u guys.

First, congratulations on completion of your dissertation.

I think, to help you land a job in the trading (or any other) sector, you need to illustrate the uniqueness of your talents. What is it about your research, its content, the way you conducted it, the lateral thinking and your analysis skills that shows that an employer should choose you above anything else.

It's difficult to ascertain from what you have told us. The conclusion of your work is hardly earth-shattering - it's probably what one would expect i.e. that a fundamentally strong stock would gain value over a number of years and one that is a glamour stock, with weak fundamentals would lose value over the same period. There are all sorts of questions that I would be asking such as how did you define each of these categories. What is the basis of your simple FA ? Was it sufficiently robust to determine suitable candidates ? How did you choose candidates from the total population ? How do you define "potential" winner and loser etc. Perhaps the ways that you explain this to an employer would have greater significance than the actual results of the research.

You mentioned the improvements that you would make in your research. I think you should spend some time considering these in depth and how you would explain them to a potential employer.

Thus, in short, I would concentrate on the process you went through and think it throught from an employer's viewpoint - then, get out there and start applying for positions aimed with those explanations.

Good Luck

Charlton

 

[theknifemac]

Yup... U could be right. It might just as well be over-optimisation, data snooping etc..
Unfortunately, testing the strategy out of the sample data is just beyond the scope of my dissertation. Would really like to test if it's still working in recent years eg. 2006-2008.

What data do you need to test the strategy, is daily OHLCV sufficient or do you need additional fundamental data ?

Mac

 

[fibonelli]

During the period 2000-2005, value stocks generally outperformed growth stocks both in the US and globally. Your test period is far too short as the value/growth cycle is typically between 7-10 years.

If your research if to have real merit, I recommend that you start the test from, say, 1982.

Also, test the strategy between 1966 to Dec1973 and then Jan1974 to 1982.

 

[zsng99]

What data do you need to test the strategy, is daily OHLCV sufficient or do you need additional fundamental data ?

Mac

Nope, daily OHLCV data is not enough. I would need yearly and quarterly data on the balance sheet and income statement in order to test the strategy, e.g. earnings, depreciation, book value, long-term debt, etc....

 

[zsng99]

During the period 2000-2005, value stocks generally outperformed growth stocks both in the US and globally. Your test period is far too short as the value/growth cycle is typically between 7-10 years.

If your research if to have real merit, I recommend that you start the test from, say, 1982.

Also, test the strategy between 1966 to Dec1973 and then Jan1974 to 1982.

Yes yes. Totally agree. I was planning to cover the period from 1985 to 2005 initially but due to time constraint, my supervisor advised that I cover only for a period of 10 years.

Bear in mind though, that value stocks substantially underperform glamour stocks in 1999 but the strategy still yields a return of 41% in that year.

 

[zsng99]

First, congratulations on completion of your dissertation.

I think, to help you land a job in the trading (or any other) sector, you need to illustrate the uniqueness of your talents. What is it about your research, its content, the way you conducted it, the lateral thinking and your analysis skills that shows that an employer should choose you above anything else.

It's difficult to ascertain from what you have told us. The conclusion of your work is hardly earth-shattering - it's probably what one would expect i.e. that a fundamentally strong stock would gain value over a number of years and one that is a glamour stock, with weak fundamentals would lose value over the same period. There are all sorts of questions that I would be asking such as how did you define each of these categories. What is the basis of your simple FA ? Was it sufficiently robust to determine suitable candidates ? How did you choose candidates from the total population ? How do you define "potential" winner and loser etc. Perhaps the ways that you explain this to an employer would have greater significance than the actual results of the research.

You mentioned the improvements that you would make in your research. I think you should spend some time considering these in depth and how you would explain them to a potential employer.

Thus, in short, I would concentrate on the process you went through and think it throught from an employer's viewpoint - then, get out there and start applying for positions aimed with those explanations.

Good Luck

Charlton

Hi Charlton,

Thanks for the advice, it's really helpful. Really appreciate it.

Regards,
Joe

 

[jackypw]

Investment banks have already been using the Long-short strategy for large portfolio management. They are using the similar fundemental analysis as you mentioned but more complex. I think your work will definitely land you a good job in this sector. Try to get contact with people in the sector of IB and see if they are interested in you?