Rental yields and bottoming out of the property market

A number of commentators have noted that rental yields in Ireland have crept up, from levels as low as 2% (or even less, in top locations such as Ballsbridge) to between 3.5 and 4%. Before the Celtic Tiger era, rental yields were much higher, even more so before Ireland joined the Euro. This is leading a number of commentators to see the creeping up of rental yields as the ominous sign of more bad things to come for the property market. But here is a thought: rental yields were so high before the adoption of the Euro because (a) inflation expectations were higher and (b) Ireland was still seen (and was) an ‘emerging country’ with the typical high risk-reward profile. Also, we were coming from decades of abnormaly high yields on financial and non-financial assets, an aspect of what financial economists studying stocks call the “equity premium puzzle”. The current level of the rental yield might be just high enough to signal ’fair’ valuation of houses. I came to this conclusion by looking, essentially by chance when working on other things, at data on returns on US stocks in the construction sector. It makes sense to me that the ‘betas’ of these stocks broadly correspond to the ‘betas’ of investing in houses, as it is reasonable to presume that margins in the construction sector would move in sync with property valuations. This essentially allows us to overcome the problem that we do not have sufficiently long time series of quality data on property investments. Well, these construction stocks have a ‘market beta’ (loading on the CRSP excess-return) of 1.26 and a ‘squared market beta’ (loading on the squared market return, a measure of coskewness) of 0.88. These betas sound reasonable to me, as they would suggest that the value of houses comes down fast when the economy slows down (high market beta) but it also offers some stability when things get really bad (positive squared market beta, i.e. positive coskewness). Using a price of market covariance risk of 2 (essentially, a relative risk aversion of 2, which is reasonable) and a price of coskewness risk of 15 (see various papers of mine for why a much higher value does not really make sense), we get a market risk premium of 4.32% and and a squared market risk premium of -0.81%. When multiplied by their corresponding betas, the risk premia add up to an overall risk premium for investing in construction/property of around 4.73%. Add to this a REAL risk-free rate of 0.7% per annum (the long run average) and subtract a conservative estimate of long-term average real GDG growth of 2% and you get 4.73% + 0.7% – 2% = 3.43%, which is very close to the rental yield at which many properties in good locations in Dublin are trading (of course, after you ask for the by now customary 10-15% discount on the asking price). In less prime areas, the ‘market beta’ would be a bit higher, say 1.4, while the ‘squared market beta’ would be a bit lower, say around zero. This would bring the rental yield higher to something around 4.5-4.75%. Again, this is the level around which investment properties in not rock-solid locations are stabilizing.