23 Mar 2015

"Oh, There Goes Gravity" - Eminem and Mean Reversion

Adam Geha

23 Mar 2015


It's not often in the course of an investment career that you have occasion to quote Eminem, but there's a lyric from his song 'Lose Yourself' that often springs to my mind when I'm contemplating an investment.

Snap back to reality, oh there goes gravity.

Not exactly the most profound of lyrics, I know, but it has literally made me and saved me millions of dollars.

In the course of my investment career, I estimate that I've reviewed financial models for over 1,500 commercial property deals (and invested in only about 60 of them). When considering which deals to go for and which to let go, one of the questions I've made a habit of asking is: "If we buy this property at the proposed price, how much money does the vendor make ?"

I ask this not to feed the black dogs of envy and schadenfreude but because careful analysis of EG's track record over the past 15 years has taught me that the deals that performed best for us tended to be the deals where the vendor had recorded a mediocre or even negative return.

And the corollary also holds true: the deals that performed worst tended to be the deals where the vendor was making a small fortune by selling to us. The correlation was far from perfect but the nexus was real and it was noteworthy.

In other words, returns cannot fly high forever. Gravity ultimately asserts itself - and moreover, the longer gravity is defied, the more likely and more savagely it is to re-assert itself.

Gravity in the investment literature is known as "mean reversion", the tendency over time for a data series to revert to its long term average.

For example, if a secondary Sydney CBD office asset generates 15% annual ungeared returns for 8-years (without a major refurbishment), mean reversion dictates that it is likely to generate 3% annual returns over the next 8-years or so. This is because the long term average annual return for such assets has tended to be approximately 9% per annum.

This is why EG's risk management software (PRISMS) now makes it mandatory for our cap trans team to estimate the (ungeared) annual return to the vendor on every proposed acquisition. If the vendor's return is significantly above the historic norms for that asset class/strategy then our PRISMS software immediately assigns a higher risk to the deal. It doesn't mean that we can't do the deal, but by applying a risk loading to the deal, PRISMS effectively makes it harder for us to do deals where the vendor is getting rich by selling to us.

So it's become part of EG folklore that when a deal leader reports that a vendor is making a fortune by selling to us at the prevailing market price, the whole room breaks out into knowing laughter and one or more of us is apt to sing the refrain (however ineptly, with apologies to Eminem):

Snap back to reality, oh there goes gravity.

Returns cannot defy gravity indefinitely. I knew that intellectually for years but it took a talented rapper to help me turn it into an axiom habitually practiced.