The Impact of Recession on CRE Debt

RMA Journal, The, March, 2001 by George J. Pappadopoulos

A significant performance decline in the stock markets, slower GDP growth, weaker consumer spending, and high energy prices have led many to revise their expectations for a recession. The level of concern has increased and a growing number of investors are deliberating the impacts of an economic downturn. Certainly, recessions are bad for real estate. Reduced demand for space drives vacancies upward, and NOIs deteriorate. Investors then price those cash flows more conservatively, and values fall. The critical question then for a portfolio lender, or bond buyer, is the extent of these declines. Is there sufficient protection in a mortgage structure to weather these declines?

Obviously, there is no simple answer. The investment's resilience depends not only on the particular loan structure, but also on the impacts of recession on the specific market and property type collateralizing the loan. Clearly each market (MSA/property type pair) has a unique level of exposure. Factors such as industry mix, extent of overbuilding, supply/demand response rate, property type, drivers and extent of the recession all play important roles in determining the degree of deterioration. How then do these variables affect individual mortgages?

To explore this question, we directly compare the impacts of recession on two different whole loan structures in two different markets. While we provide only a sample of this recession analysis, it behooves debt investors to run simulations such as these on their own portfolios, which are uniquely characterized by specific loan structures and market exposures.

Combining market research on the path and the volatility of income and value growth with the loan structure creates a rigorous context for forecasting whole loan and loan pool performance. We roll each loan forward through a particular forecast of market behavior and observe a time series of the key ratios of the loan--LTV and DSC. At the mortgage level, default is viewed in the context of current, localized market pricing information, and is based on contemporaneous measures of LTV and DSC ratios.

In other words, we measure the extent to which the borrower's economic position changes over time. At origination, the original loan structure provides downside protection--the borrower has more than sufficient cash flow and equity. But now we can see how this level of protection is impacted by changes in real estate and capital market fundamentals. These ratios offer a measure of just how thin our original protective "cushion" might become.

Using this approach, we compare two loans that are part of the collateral for a security in the market today (see Figure 1). The chart on the left illustrates a 77% LTV, 1.18 DSCR apartment loan in Los Angeles, while that on the right outlines a 75% LTV, 1.26 DSCR office loan in Washington. By using base case and severe recession projections for NOI and value for the collateral underlying each loan, the charts compare the loan-specific impact of the different sets of projections on the critical loan ratios.

The first thing to note is how much larger the differences in recession and base case performance are for the Washington loan versus the Los Angeles loan. By comparing the red and green DSCR lines (and/or the red and green LTV bars) of each chart, it becomes clear that a recession retards the strong NOI and value growth expected for apartments in Los Angeles, but the deterioration is nowhere near as severe as that for Washington office.

There are two key reasons for the difference. First, apartments generally fair better than office during a recession. Although apartments are negatively impacted by the lower levels of household formation that occur during a recession, they benefit from reduced homeownership rates. In addition, the Los Angeles apartment market is much earlier in its cycle than Washington office. Los Angeles has not been as strongly impacted by heavy new supply as Washington. This supply in Washington office is fostered, and currently supported, by strong demand growth-absent a recession, Washington office is expected to perform quite well. Unfortunately, in a recession the demand goes away but the supply remains, and the end result is a more severe impact.

What this means for the specific loans is that the seemingly more protective loan structure placed on Washington office (DSCR of 1.26 vs. 1.18, and LTV of 75% vs. 77%) could be more problematic than the Los Angeles loan's structure in a recession. Exhibit 1 outlines improving ratios for both loans under the base case scenario (the green DSCR lines and LTV bars), but the red recession lines and bars reach very sensitive levels for the Washington office loan. In fact, both ratios cross the dangerous level of 1, a situation in which the borrower must continue to make debt payments yet has no equity in the collateral.

LTV and DSC ratios are helpful measures, but they actually fall short of providing a complete assessment of a loan's exposure to market risk. A more appropriate measure would also address marker volatility--our level of comfort regarding the available cushion. Iris the marker-specific volatility of NOIs, and values, that describes the likelihood that a given level of movement in NOI or value (enough to ear away our cushion) will occur. Probability theory tells us that the larger the required movement relative to its standard deviation, the lower the likelihood that problematic movement will occur. Therefore, our risk measure at each point in time is simply the quotient of each time period's NOI or value cushion (numerator), and the market standard deviation of NOI or value (denominator).