How much is that building?

Want to know how much you’ll pay for a share of stock? That’s easy – just look up its price on any of the hundreds of websites that provide stock prices in near real time.

Want to know how much you’ll pay for that six-story office building your company has been eyeing as a potential investment? That’s not so easy, because commercial real estate is an illiquid asset.

Figuring out what price you should pay and what kind of return you might earn on that building if you sell it sometime in the future is even tougher. Real estate is not bought and sold frequently enough to provide real market-based price data, like the kind investors have for most stocks and bonds.

This is a big problem. Commercial real estate represents a significant portion of overall asset value in the economy. In 2010 there was an estimated $3 trillion of institutional quality commercial real estate in the U.S., and in 2015 there was about $2.5 trillion in institutional and non-institutional quality commercial real estate debt — most of it tied to individual properties. In other words, commercial real estate represents trillions of dollars in equity and debt on the economy’s balance sheet.

And the problem of pricing commercial real estate extends to other illiquid assets, too, including residential real estate, private equity and other alternative investments.

Jacob Sagi, a finance professor and Wood Center in Real Estate Distinguished Scholar, set out to tackle this problem. He published his finding in “Asset-Level Risk and Return in Real Estate Investments,” which won the best paper award at the 2017 Utah Winter Finance Conference.

A traditional pricing model, which works well for frequently traded assets, like stocks on a stock market, finds that the longer an asset is held, the greater the range of possible values it could have when sold. This is based on the “random walk” theory, which says stock prices take a random and unpredictable path. One important implication of this model is a mathematical relationship that ties the range of likely asset values over a short holding period to the range of likely asset values over a long holding period.

But when Sagi applied this model to real-world information on commercial real estate sales supplied by the National Council of Real Estate Investment Fiduciaries (NCREIF), it didn’t fit the data. Commercial real estate prices don’t appear to follow the random walk. Variations in the return on commercial real estate investments held over longer holding periods appear too small, based on what we see from shorter holding period transactions. Even more puzzling, the data actually suggest that short holding periods — a year or less — are more likely to yield abnormally high returns.

“If you approach this data as you would when looking at stocks, it makes absolutely no sense,” Sagi says. “Either there’s a problem with the data or you need a new theory.”

So what’s going on? Does it really make sense to buy commercial real estate and then sell within a year to produce higher average returns? Intuitively, it seems unlikely that would work on a consistent basis. So how might investors get a better grip on commercial real estate prices, risk and return?

A different selling process

Sagi developed a new model to make sense of asset-level risk and return for real estate investments. To do so, he considered unique characteristics of real estate transactions.

First, they are typically facilitated by a broker. When it comes to buying and selling real estate, there isn’t a market of practically unlimited buyers willing to make you an offer, as there is when stocks, commodities or other liquid assets trade. The number of potential buyers a seller can transact with in a given time period is limited, and each buyer and seller will value the property based on individual attributes, such as their cost of capital.

Second, commercial real estate transactions take a long time to execute. Once an offer is made, there is a lengthy period – usually several months – while the buyer conducts due diligence. The buyer will inspect the property. Frequently during this process, the transaction price that buyer and seller had agreed to is renegotiated. Due diligence reveals facts about the property that can lead to a change in the price the buyer is willing to pay — perhaps the HVAC system is on its last legs and will require an expensive replacement soon, for example.

And third, during due diligence, commercial real estate is “locked up.” That is, even if another potential buyer comes along while due diligence is being conducted, the seller cannot engage with that buyer. Instead, the due-diligence process runs its course until buyer and seller agree to complete the transaction or the buyer calls off the deal by paying a small proportion of the purchase agreement price.

Sagi’s model accounts for these attributes, providing investors a new way of assessing the value of commercial real estate. The model also fits the data from NCREIF, suggesting it accurately describes what’s going on in the real world of commercial real estate transactions.

Implications of the model for investors

The model has some important implications for commercial real estate investors, whether they own property they plan to eventually sell or are hunting for property to acquire.

In the model, owners periodically receive bids for their property from investors, but gains are only possible if the bidders’ valuation (net of transaction costs) exceeds the value the owner has for the building. Since it takes time to discover whether a bid corresponds to a realizable transaction, and not all buyers and sellers will have complementary valuations, the number of opportunities to sell (or buy) a particular piece of commercial property is limited.

This dynamic introduces substantial transaction risk – the possibility that a potential sale at a given price will fail to take place. For a liquid stock, the transaction risk is essentially zero if the given price is below the current market value of the stock. For example, if Apple (NASDAQ: APPL) is currently trading at $200 per share, then even for large trades there is little risk that a limit order to sell at $199.99 will fail to be consummated. By contrast, based on the NCREIF data, the odds of a transaction at a “fair” market value is just 12 percent in a given quarter. (Fair market value, in this case corresponds to the appraisal price of the property, which in turn is based on prevalent current transactions for similar properties.)

An owner targeting an expected transaction at fair market value should also expect that it would take about a year for the right buyer to come along. The NCREIF data Sagi worked with didn’t have time-on-market data, but “days on market” statistics from Costar, a commercial real estate data provider, appear to confirm the model’s predictions.

For an owner with a high cost of capital who may want to sell a property quickly, that means listing the property at a lower price (and higher cap rate) to increase the odds of getting a sale. To increase the chances of a successful sale to more than two-thirds in a given quarter, the seller would be forced to discount the asking price 25 percent compared to prevailing market prices. This insight is helpful, Sagi says, when pricing mortgages or commercial-mortgage backed securities, particular in foreclosures where the lender has a high cost of capital or lacks the expertise to manage commercial property while selling it.

Owners who want to sell for higher prices will find themselves in the opposite situation. The more patient they can be, the better the chance they have of selling for a relatively high price.

So why is it that the random walk prediction fails in commercial real estate – i.e., short holding period variation are too high relative to those of long holding periods? The transaction risk disproportionately amplifies the range of possible returns from shorter holding period. In fact, the shorter the holding horizon, the more dominant the transaction risk. By contrast, for longer holding horizons, the transaction risk is eventually “amortized” away.

And that quirk that Sagi found in the data – that investors who sold after holding a property for a year or less often saw above-market returns? The model suggests that they were just lucky: The ones that bought cheap and sold dear. If such an investor hadn’t first encountered a desperate seller and shortly after encountered a buyer willing to pay above-market prices, they wouldn’t have held the property less than 12 months, and the figures showing above-average returns wouldn’t have ended up in the NCREIF data.