Archive

Posts Tagged ‘performance measurement’

Absorelative performance

January 13th, 2009 5 comments

No, it’s not a typo… “Absorelative” is the term I’ve coined for what we’ve been tasked with at work for the year 2009. Our mission is to earn an absolute return, unless the market moves up, in which case we’re expected to deliver a relative return – so we’re looking for a mix of absolute and relative performance, or “absorelative performance”. Basically, this means that if everyone loses money, we must not. But if everyone makes money, we too must make money.

To better understand the implications of this, it’s important to understand how money manager performance is measured, and how managers are compensated. Basically you can measure how your investment manager is performing in one of two ways: absolute return, or relative return. Absolute return is the total return of your portfolio. If you give someone $1 million to invest and they give you back more than $1 million at the end of the year, then they have earned a positive absolute return. Likewise, if you get less than $1 million, they have earned a negative absolute return. The fact that other investments (such as a passive index fund) could have earned 30% or lost 30% is completely irrelevant – your manager gets paid if he/she makes money. Relative return, on the other hand, depends on a benchmark. Suppose you want your manager to invest in large U.S. company stock. An appropriate benchmark could be the S&P 500 – an index of the largest 500 U.S. companies, which  you can invest in yourself – it’s pretty easy and cheap. So, you only want to pay your manager if he/she is able to beat what you can earn yourself – that is, beat the S%P 500. If the S&P 500 earns 25% and your manager earns 28%, then he/she has earned a 3% relative return. If he/she earns 22%, he/she has earned -3% relative return – he/she has still made money, but he made less than the benchmark, and so shouldn’t get paid. Clearly, the best compensation structure for you depends on what you’re looking for – if you want exposure to a particular asset class and are open to its risk and return, you want to set up a relative compensation structure – only paying your manager for picking those names that are better than average. However, if you’re looking to minimize risk and just want positive growth, an absolute structure is probably a better fit – an often small, but positive, return with very little chance of losses.

Certain types of money managers typically have certain kinds of compensation. Hedge funds can go long and short (that is, they can buy assets they think will make money, and borrow and sell assets they think will lose money, expecting to repurchase the sold asset later at a cheaper price to repay the loan, and keep the difference between the sale and repurchase prices), and so typically employ strategies that can profit in up markets and down markets. These strategies should make money no matter what, and so are called absolute return strategies. As you would expect, since hedge fund managers employ absolute return strategies, they are typically compensated on an absolute return basis.

Traditional money managers (think of insurance companies, mutual funds, etc) aren’t allowed to go short, and so can only buy whatever asset class they are managing (bonds for bond managers, stocks for equity managers). Thus, if you’re a long-only equity manager, and stocks are falling, your portfolio will fall too since all you can do is buy the falling asset class. The performance of these strategies is tied to the benchmark, and so they’re called relative return strategies, and the manager is compensated relative to some benchmark.

Getting back to the situation at work… Allstate is an insurance company, and as such we’re a long-only shop. Until now our performance has been measured on a relative basis. In 2008, equity markets fell sharply, as did our portfolios. By definition, if we beat our benchmarks we should still get paid, and if not, we shouldn’t. As you can imagine, the investors who give us capital may not be thrilled with this setup. On one hand they were looking for exposure to equities (and by that I mean the risk and return that comes with equities), and should be glad that we were able to provide a return greater than, albeit still negative, the benchmark they had selected. On the other hand, when 35% of your portfolio disappears, it’s hard to be excited and pay your money manager for making you one third poorer than you were at the start of the year. So, going into 2009 things changed – the investors decided that they didn’t want to take this much risk, and that they didn’t want to pay their managers if they lost money – all the ingredients of an absolute performance setup. (As a side note, I would have argued that if the investors aren’t comfortable with the risk of equities, the simpler, more direct solution is to simply not invest in equities. They should invest in a safer asset class. However, this is a suicidal argument for me since if the investors choose not to invest in equities, I’d be unemployed!)

Earning an absolute return in itself is not a big issue – see my upcoming post on negative beta. The problem comes in when the investor wants the relative dimension as well – when the investor says “don’t lose money, but if the market is up x%, you must be up about x% too”. The issue here is that in order to get absolute return, we have to eliminate all beta exposure (market beta, size beta, style beta, sector beta, etc). This way, no matter what happens in the markets, our stocks earn only alpha – their unique return after accounting for all systematic risk. This is a great positive in down markets – in fact zero beta is what allows us to not lose money (assuming we’re right about our stock predictions – our alpha). However, not having this beta exposure in up markets means that your portfolio is missing out on systematic return. That is, if everyone feels more confident about things and stock markets recover, you’ll miss the recovery because it is market-wide compensation for positive sentiment, not company-specific compensation. Thus, you will under-perform the market in up markets.

So, given that we can’t earn absolute return by neutralizing beta since that also eliminates our return in up markets, the only choice we’re left with (as far as I can see) is to be in stocks in good times (so that we’re earning alpha and beta), and be in cash (or cash equivalents) in bad times so we’re not losing money tied to our stocks’ betas. This strategy is called market timing, and it’s something that no manager that I’m aware of has ever been able to do consistently and successfully. Basically you have to know more often than not when the market will go up so you can buy stocks, and when it’ll go down so you can sell stocks.

Now, my friends, you understand the dilemma for 2009. In the absence of a crystal ball that tells us what will happen in the markets tomorrow, how do we know when to switch from stocks to cash? If you have any ideas, or see something I missed, please leave a comment and educate me. In the mean time, I’m going to roll my dice – even means market, and odd means cash!