Friday, April 11, 2008

How Could I Produce an Alpha of 9%?

A recent discussion on Roger Nusbaum's blog typified the diametrically opposed positions of those who think it is easy to beat the market and those that think it is impossible. I'm targeting an alpha of about 9%, so how do I think I can produce it (apart from pointing at my recent track record)? There are three main potential sources:

1. Active trading: 2-4%. 2% means earning the same amount in trading as last year. One of my annual goals is to beat that number. 4% would be doubling last year's result, which is, realistically, the best result I can imagine at this stage.

2. Passive Alpha: 2-4%. About 40% of my portfolio is dedicated to what I call "passive alpha" investments. These are actively managed funds and other financial companies which I believe can produce significant risk adjusted returns. I assume they could attain 5-10% each. Multiplied by the portfolio share that is 2-4%. 5-10% is not just hypothetical. TFSMX has an alpha of 8%. Berkshire Hathaway has been credited with an alpha of 10%. Man Financial has averaged at least 10%. And so on.

3. Timing and Security Selection: 2-4%. These numbers are purely hypothetical. But let's assume that my portfolio beta was 0.5 for the first two months of the year and I then increased it to 1. I would have avoided half the losses in the first two months of the year by timing. This assumes that the markets are relatively benign for the rest of the year and I timed in the right not the wrong direction. The MSCI lost almost 8% in January and February, while my portfolio lost around 2% in total (both in USD terms). Therefore, avoiding 2-4% of losses here through timing sounds reasonable. Of course, if I never changed the beta upwards then this result would be purely due to low beta. Hopefully, some of my few industrial stock selections will add a little value too.

To explain the timing effect, let's imagine that the market goes down for six months of a year at 10% a year and goes up the other six months at 10% per year. Also imagine that the investor has a true beta of 0.5 when the market is going down and 1.0 when it is going up. If we use a regression to estimate a constant beta for the whole period, we'll come up with the average: 0.75. Then in the declining six months my predicted market return will be -7.5% p.a. but I'll in fact only lose -5% p.a., while in the rising part of the year my predicted return will be 7.5% but I will in fact gain 10% p.a. The investor's alpha from this source will, therefore, be 2.5%.

The average of each of these categories is 3% and adding them all up we get to 9%. Of course "alpha" technically is the average excess return over a period of reasonable length. Looking at just one year is probably stretching the concept. Maybe, I should just say a "risk-adjusted excess return of 9%". But it's easier to say "alpha" :)


Anonymous said...

Given a long-term average return on stocks of 10% or so, you think you can add 9%?

So you can average 19% per year? Highly doubtful, I would say. It can be done, but not very many people have managed it over the long term. That's up there with Soros.

Bigchrisb said...

In your shoes, I'd be asking myself two questions:

1. What is the opportunity cost of alpha? If you ahceive alpha by being a full time trader (at the cost of a regular job), then you need to benchmark it as appropriate. If I was in your shoes, and I could make $40k/year of alpha, that would be a pretty poor trade-off against my $100k salary. Indeed, a $60k/year worse trade-off. Perhaps you should be benchmarking against the MCSI and the opportunity cost of your time?

2. You were very fortunate in the timing of your return to Australia. Currency effects over this period have significantly offset your losses in USD terms. However, this is an impact that you are unlikley to repear with good timing, and a change that had significant other costs (costs of moving). I would be very wary of using this period for forward predictions of my future peformance.

Those two things said, I'm impressed by your confidence in effectivley withdrawing from the workplace with a portfolio of ~$430k. I'm intending to have about ten times this before I feel connfident of taking the same step.

mOOm said...

Anon - Soros, Simons and the like returned 35% a year. That was after their fees. So I'm not aiming anywhere near that.

Chris - very good comments. I earned $US75k in my last year as a professor. So it is a pay cut. I doubt I'd make $US75k right away here though. I plan at the end of this first year to assess my forward direction. My wife is of course working at the moment. I've actually been thinking of doing an analysis in Australian Dollars and did some preliminary calculations - I'll have to do that now :). Another factor that is probably having an effect is the money I expect to inherit and which I've discussed on here my role in managing. If I include that, my net worth is 4 times higher. I think close $2 million currently in total is certainly a good number given our modest lifestyle ambitions. I don't need $4 million. Yes, there is no guarantee on that number but it allows me to take a little more risk now to see what can work out. At the end of this first year I will reassess and decide what I want to do going forward.

mOOm said...

I did the analysis using Australian Dollar returns for myself and the MSCI. Alpha is lower as expected at 7.2%. The main puzzle is that it hasn't increased as much over time as the USD based ones - they show negative values back in the 1990s while the AUD one was positive then already.