Even still, Smith and his predecessors based their probability estimates around the infamous Gaussian bell curve that has become the darling of stats professors worldwide. It wasn’t until a new breed of mathematicians and economists came to prevalence that Smith’s techniques for calculating risk were also seen as too stringent to fully understand the messy world around us. One of the standouts of this new clique in the economist camp is Benoit Mandelbrot. He is a polymath who applies his studies in fractional geometry to try to understand “the art of roughness” of physical phenomena and “the uncontrolled element in life.”
Mandelbrot wrote an influential paper titled How Fractals Can Explain What’s Wrong with Wall Street back in 1999. In 2008, when the world’s credit system was melting down, this essay was rehashed to help describe what are now known as black swan events. “The mathematics underlying portfolio theory handles extreme situations with benign neglect: it regards large market shifts as too unlikely to matter or as impossible to take into account,” he wrote. He goes on to explain that, “It is true that portfolio theory may account for what occurs 95 percent of the time in the market. But the picture it presents does not reflect reality, if one agrees that major events are part of the remaining five percent.” He then uses a maritime example to illustrate his point because, after all, even economists love a good folksy fable: “An inescapable analogy is that of a sailor at sea. If the weather is moderate 95 percent of the time, can the mariner afford to ignore the possibility of a typhoon?”
Now is the part of the story where I relate things to the COVID-19 pandemic, the black swan event of the day. But calling this downturn a black swan event (or even unprecedented for that matter) goes against exactly what Mandelbrot was preaching. Yes, a lot of the details of the COVID-19 virus and the world’s response was difficult, if not impossible, to predict. But, viruses are not new. There have been plenty of examples of infectious diseases wreaking havoc on cities and states, just never in the globally fused, internet fueled era that we are living in now.
The commercial property industry was particularly exposed to the shutdown. Retailers were forced to close, causing what is looking to be a domino of bankruptcies. Office workers are working from home, many of which are considering not returning. Many tenants are unable or unwilling to pay their rent, all the while governments around the world are imposing eviction moratoriums. As property companies worry that many of their investments might see a significant drop in value, many in the industry are considering new ways to underwrite properties to be prepared for a similar (but totally different) event in the future.
When we talk about risk, we are usually referring to its impact on value. Correctly valuing a commercial property is one of, if not the, most important consideration for anyone in the property industry. While there is quite a bit of standardization in the appraisal industry, the nuances are the secret sauce for property investors, asset managers and acquisition teams. A number of methods are used in conjunction to try to land on a realistic value for a property. Some methods use comparable properties, or comps. They find recent sales and try to infer how the different characteristics of a property in question would compare. This strategy has a few critical flaws. First, buildings are unique, so trying to compare two of them can leave out important factors, location being the most obvious one. Second, commercial property don’t sell that often. For many small metros there might only be a few sales a decade for large properties. This makes it almost impossible to infer much about value for a building in what can be a completely different market. Even in markets with a lot of turnover traditionally the COVID-19 pandemic has brought transactions to a near complete halt, creating a situation where everyone is waiting for sales to start back up again in order to find a new supply/demand equilibrium.
Our unique environment and associated circumstances require much more due diligence—and different considerations for specific property types, markets, and submarkets.
Andrew Lines, Valuation Partner at CohnReznick
The other methods for valuing a building are more akin to what the rest of the financial world does to price assets. It uses a discounted cash flow method that adds up all of the total projected revenue and subtracts the total future costs. It then discounts future revenue based on a discount rate, this discount rate is where the risk calculation comes in. “The discount rate, in essence, is the thermometer of investor expectation,” says Andrew Lines, Valuation Partner at the advisory firm CohnReznick. The discount rate is usually determined by the “risk free rate,” usually ten year U.S. Treasury bonds, and a risk premium. These premiums are usually calculated based on historical averages and often applied across the board to all property types and markets. “Our unique environment and associated circumstances require much more due diligence—and different considerations for specific property types, markets, and submarkets,” Lines explained.
So many factors contribute to this one number that becomes calculating it almost seems impossible. Most professionals in the property industry simplify the process by using some variation of a weighted average. Optimistic, pessimistic, and neutral outlooks are determined, and the probability of each is weighted into a final calculation. This attempts to put the world’s uncertainty into a few predictable, precise scenarios—something that economists like Mandelbrot have warned against for decades.
One of the problems with calculating risk this way is that it only takes into account predictable assumptions. The unpredictability of the COVID-19 crisis is exposing these flaws. For example, many models will take into consideration a low vacancy rate, but I am guessing that almost none factored in a large scale stoppage of rent payments like we are seeing now. Another problem is that many of the assumptions that the property industry uses are based on the market, not the underlying financials. Many hot markets had seen building prices get bid up to historically low cap rates, the ratio of operating income to the buildings purchase price. Like with any market, strong demand often signals value, which creates more demand and adds air into the kinds of financial bubbles that tend to burst rather than deflate. As more and more investors were willing to pay top dollar for high-demand real estate, property prices started to get to a point that was untenable compared to the risk they carried—at least, compared to the unknown risk that wasn’t being represented in the industry’s models.
There are some lessons that can be learned from investors in non-publicly traded companies, commonly referred to as Private Equity. Much like properties, the businesses in which they invest are unique and illiquid. To help them understand how to discount the company’s projected cashflow in their models they use detailed tables that inform them of the proper premium or discount that should be applied. These tables are made for all kinds of risk, regulatory risk, supply chain risk and, yes, even pandemic risk. Some also use advanced modeling techniques like Monte Carlo simulations that calculate outcomes of every possible scenario and can help visualize the probabilities compared to outcomes. So why are these same techniques not common in modeling for commercial properties?
To learn more about this I talked to L.D. Salmanson, co-founder of the real estate data management and analytics platform Cherre. “Very often the property industry focuses predominantly on the current cost of capital—the loan from the bank—to determine the discount rate, and assign a constant value to represent other, systemic, or market-oriented risks,” he said. “This is overly simplistic since there is so much that can change over time with regard to our understanding of these risks and our ability to calculate them. It is a group of variables and not a constant, and it is different for every market, sector, or perhaps, even property.” His company has created a knowledge map for every property, tenant, owner and lender in the entire country to help them have a more granular understanding of how the market is changing. He hopes that we will soon see fund managers and lenders work more closely with academia to help them better understand and capture these risks in modeling, something that other sectors of finance have been doing for decades.
So, if we view the discount rate as the fluid, undulating set of coefficients that it really is, financial models will need to adapt accordingly. I spoke with Arik Kogan, vice president of financial and investment solutions at MRI Software, which provides management and modeling tools for the property industry. He predicts that this pandemic will force investors to be more nimble with their modeling techniques, “I think we will see organizations that have relied on low-level specific Excel modeling of individual assets will now look to assess whether they have the kind of insight and nimbleness they need at the portfolio level.” He says the speed with which this crisis emerged has shown the importance of having the ability to quickly change models, even for previously unforeseen events like an anchor tenant going bankrupt or a complete freezing of the commercial mortgage industry.
It isn’t just property owners that need to update their modeling techniques. One of the most exposed organizations during this downturn are lenders. Many have had to grant forbearance for loans on properties that are not collecting enough rent to cover their debt obligations. The concern is that these loans, which often have strict covenants that require certain debt to value ratios, might require more capital to be put in by borrowers whose devalued buildings now do not meet their ratios. This could trigger a liquidity crisis that could exacerbate the losses already being felt by property owners and operators.
What has bothered me about lending from a research perspective is the fetish they have for thresholds—we are basically using the same techniques we did thirty years ago.
Richard Green, Chair, The Department of Real Estate Development at USC
Richard Green is the chair of The Department of Real Estate Development at the University of Southern California. He has long been a proponent of a more advanced way to underwrite commercial real estate lending. “What has bothered me about lending from a research perspective is the fetish they have for thresholds—we are basically using the same techniques we did thirty years ago,” he told me. He sees a world where lenders put important factors like loan to value, debt coverage ratios, amortization period, term, sponsor track record and credit of tenant into a regression analysis. From there each transaction or loan would be able to have a “scorecard” that would show how far it varies from other deals. Over time, this would be able to predict risk of loans outside of the model and help create a much more sophisticated process that would, at times, challenge popular opinion. “The gold standard for office tenants used to be Arthur Anderson, for retail it was Sears and J.C. Penney’s. The industry assumption that big corporate tenants are always less risky than smaller businesses obviously needs to be re-thought,” Green said.
One of my main takeaways from Mandelbrot’s work was that risk exposer over time isn’t linear, it is exponential. The further off we try to calculate probabilities of certain outcomes, the harder it becomes. In the 90s J.C. Penney was able to raise $500 million dollars through a 100-year bond. Now, just under one third of the term through the loan, the company is now filing for bankruptcy. Even if the store was able to survive the e-commerce revolution and this global pandemic, imagine how much more disruption might be in store for it from now until the end of the century. I expect that, with the accelerating pace of change that we are seeing, we won’t see another issuance of a 100 year corporate security for a long time, if ever.
As Madelbrot explained, by trying to wrestle fuzzy real-world probabilities into exact calculations you can expose yourself to the unknown unknowns that regularly occur, albeit at irregular intervals. The property industry needs to adapt a new mindset and deploy new tools to help understand properties, and their risk premiums, on a much more individualistic level. We need to keep in mind that there is no such thing as a perfect model and be ready to change our entire calculation methodology as new information comes to light. We also need to remember that, no matter how much we know, there is always another surprise, one that no amount of calculation can predict. As Mandelbrot explains in his essay, again bringing back the nautical reference he loved so much, “The new modeling techniques are designed to cast a light of order into the seemingly impenetrable thicket of the financial markets. They also recognize the mariner’s warning that, as recent events demonstrate, deserves to be heeded: On even the calmest sea, a gale may be just over the horizon.”