11 August 2015
There seems to be a veritable cottage industry based on the failed predictions of rising interest rates – remember 2011, or 2013 and the end of QE3? Perhaps we should get interest rate predictions right before worrying about the consequences (if any).
While at times the correlation between the long bond rate and listed real estate prices is high, over the longer term, there is little empirical evidence to support any relationship between nominal interest rates and real estate values.
This is not surprising since there is no theoretical basis for a link between nominal interest rates and commercial real estate values.
Remember the Bond crash of 2011 and 2013? Neither do we!
In March 2011, leading global bond managers argued the end of Quantitative Easing (QE2) in July 2011 would result in soaring interest rates. The argument was ‘Who will buy the bonds if not the Federal Reserve?’
In 2014, as the end of QE3 approached, the consensus estimates for the year-end 10-year US bond yield was 3.4% (rising from 3.0% in Jan 2014).
What transpired was the opposite of expectations. Bond yields fell significantly after QE2 in July 2011 and again in 2014. Why does the market get these forecasts so wrong?
Our best guess is most commentators believe market forces determine interest rates. While this is true along the curve, the curve in turn simply reflects the anticipated future cash rate. And the cash rate is entirely determined by the monopoly issuer of the currency – the central bank. In short, as with any market dominated by a monopolist, interest rates are anything but a free market.
The best prediction method for long-term bond rates is to attempt to predict the monthly cash rate for the next 10 years as determined by the currency issuer – the central bank. Given the anemic recovery of the US economy at near zero interest rates, it was a brave person to suggest the average cash rate over the next 10 years would be 3.4% after QE3.
First we need to be clear about exactly what we are discussing. Cash rates or long bond yields?
The chart below shows the rolling 6-month capital gain of holding 10-year US treasuries versus the rolling returns of listed US REITs. If US REITs were sensitive to long bond yields, we should observe a strong positive relationship (higher bond prices i.e. lower yields = higher REIT prices). While there appears to be good correlation since 2012 (correlation around +0.6), there is little evidence to support a long-term relationship between long bond prices and listed real estate performance (correlation -0.37). Even excluding the traumatic events of the financial crisis (February 2008 to June 2009), the correlation between bond and listed property prices is -0.17.
The relationship between listed REIT price performance and short-term interest rates is even less convincing, suggesting bond yield movements are more relevant.
And in Australia, the data is similar. There is a good correlation recently, but nothing significant over the long term.
Why is there no significant long-term negative relationship between interest rates and listed real estate prices?
Real estate, like most asset classes, offers a total return comprising of income and capital gain. Over time, capital gain reflects the growth in underlying income (rent).
For real estate to capture growth, the owner (generally) provides ongoing capital to maintain the competitive position of the real estate, which may include tenant incentives, commissions, and general ‘stay in business capex’. Not every dollar collected as rent remains in the landlord’s pocket.
In addition, growth can be enhanced by expansion, redevelopment or re-positioning to higher and better use. Outperforming inflation growth can also be achieved by buying assets well below adjusted replacement cost, buying assets where the demand from the underlying tenant is expected to outperform the economy, or where supply is restricted by natural or planning barriers. Rising inflation also adds to growth, sometimes offset by the rising nominal interest rate.
In contrast, bonds offer a total return equal to their yield to maturity. That’s it.
Comparing a total return (bond yield) with only part of a total return (real estate yield) is clearly not comparing apples with apples.
Over short periods, there is good evidence listed real estate has some correlation to the movement in long-term interest rates. For traders, it makes sense to worry about listed real estate pricing and bond yields.
For long-term investors, however, factors which drive total returns do not include the need to predict nominal interest rates, which is good news since it appears to be a very difficult thing to do.
As a simple thought experiment – if its costs $30,000 per sqm to build an office tower in Sydney by the year 2025, office values will be around $30,000 per sqm, irrespective of interest rate levels.
In a world where ‘Fed watching’ has become a full time profession for many, it is easy to become distracted with discussions about interest rates. In this environment, we remind ourselves of the following:
The content contained in this article represents the opinions of the authors. The authors may hold either long or short positions in securities of various companies discussed in the article. The commentary in this article in no way constitutes a solicitation of business or investment advice. It is intended solely as an avenue for the authors to express their personal views on investing and for the entertainment of the reader. In particular this newsletter is not directed for investment purposes at US persons.
If you enjoyed this article, please share to help others find it. If you’d like the next article delivered to your inbox, select 'subscribe'.