Walking the dog

April showers and May … flowers?

The weather is now ideal for a May stroll, or even a random walk, as April showers fade from our macro view. Positive news on the Chinese and US economies, a buoyant equities market, and a Brexit extension have all brought more sunshine into the outlook. Our macro deck is filled with financial data series – and many real estate metrics – that appear to be random walks. Global stock market indices, bond markets, and even property prices do not always revert to a stable long-run average. Nobel prize-winning financial economist, Eugene Fama, described such time series as “having no memory.” This makes them inherently unpredictable using even the most advanced econometrics.

Fortunately, this does not mean it is impossible to predict real estate prices or rents, thanks in part to the special case when random walks are cointegrated with each other. Think of this like going for a ramble with your dog. The path you take might be random, but it would be a safe bet that you and your dog would stay close to each other along that unpredictable path, albeit with variations depending on whether you are using a leash. Cointegrated time series behave much this way. For example, while recently listed Lyft and Pinterest’s stock prices (see page 13) are a random walk, analysts can price information on New York exchanges to predict the price movement of Rakuten shares in Tokyo, which has stakes in both companies. If market prices do not reflect this, then it can create an opportunity for arbitrage.

For real estate investors, the key cointegrated pair of time series are real estate values in the public and private markets: page 3 in this month’s deck. The difference between these series – the NAV premium or discount – has varied significantly over time, but has been near parity over the long-run. However, since October 2016, the implied value of real estate based on public REIT and real estate security pricing has been below the values implied by transactions in the private real estate market. This significant global NAV discount had some observers wondering if this particular “leash” had snapped entirely in two, especially for office and mall REITs.

LaSalle’s conviction that public and private real estate values must be similar over the long run – based on their cointegrated relationship – has been borne out so far in 2019. In early April 2019, the NAV gap – based on the EPRA/NAREIT global index – reached its narrowest level in 31 months. The way the gap has been resolved is meaningful. Private values did not decline to meet the public market, albeit with some exceptions, such as for US mall retail. Nor did public and private meet in the middle. Instead, public values have recovered and the gap has been resolved largely in favor of private real estate values. Underneath these headline figures, there continues to be significant variation in NAV discounts and premiums across individual countries and sectors – leaving many opportunities for LaSalle’s Securities team to make accurate, high conviction predictions of relative value changes.

We recommend that investors focus their forecasting efforts on three key areas where high conviction predictions are possible: cointegrated data series like those described above, data series that are not random walks but are mean-reverting (like the vacancy rates shown on page 23 of this month’s deck), and real-time information predictive of the near future – such as when competitive projects are scheduled to break ground. Technology and improved data frequency is changing how we do these kinds of forecasts. Not unlike taking a Fitbit or smart watch on our next ramble, real estate investors benefit from the collection of more granular and geolocated data.