If your bank is lending on a property with a variety of units available to lease, there is a chance that all those units might be leased up and there is a chance that none of the units will maintain their lease over the life of the loan. The reality is, the outcome is likely somewhere in between. The average banker would look at one set of cash flow and calculate their debt service coverage off a base case using a set of assumptions. The good news is you are not an average banker. Otherwise, you wouldn’t be reading this. The better way to conduct investor real estate analysis is to create a base case and then a set of scenarios around the base case. In this article, we look at how to conduct a probabilistic analysis on a commercial real estate property and also provide your bank with a model to sharpen its underwriting and to offer to your potential borrowers in order to add more value to the relationship.
When it comes to underwriting, there are two main types. The more common is “deterministic analysis” where the underwriter looks at the fundamentals of the property to see that the set of expected cash flow is sufficient to meet the underwriting criteria of the bank. While the approach is sound, it has its shortcomings.
Take the loan below with an average 1.33x debt service coverage. Would you approve it?
For starters, bankers don’t know if the presented base case is likely or not. What is the probability of achieving that case? Is it conservative and assured, or is the projected debt service coverage a fantasy?
The base case likely only considers one outcome and ignores thousands of others. For those banks that do a best and worst case scenario, that is a step in the right direction. However, you still don’t know the likelihood of each outcome.
Further, interdependence between inputs is likely ignored, and the model presents an oversimplified version of reality with little predictive power.
The better way to underwrite a multitenant commercial real estate building is to look at a spectrum of outcomes using probabilistic underwriting. This is also known as a Monte Carlo simulation and means that inputs are modeled using a range of possible values according to a given distribution. In the model below, we will assume a “uniform” distribution of outcomes where all possibilities are equal, but we could also use a lognormal, power or, the most common, a normal distribution.
A Monte Carlo simulation entails letting a program, such as Excel, randomize a set of input variables and then looking at the outcome. This one iteration. In our sample model, bankers can input the total number of units, occupancy, revenue, expenses, cap rates and other inputs.
We then let the program choose another set of variables to run randomized iterations. After running hundreds or thousands of iterations, bankers now have a probabilistic set of outcomes. The Monte Carlo output offers a much more comprehensive view than the output from deterministic underwriting. Probabilistic underwriting tells bankers not only what could happen but how likely it is to happen.
Let’s go back to our deterministic underwriting in the Base Case table at the start of this article where the DSC was 1.33. Here, bankers might assume that since the building meets their minimum debt service coverage test of 1.25x that the risk profile is sufficient.
After running one thousand iterations of the base case using a Monte Carlo simulation, we get the below output.
We now know not just the averages, but the variability around the mean for this property. This output now gives us a feel for the realm of possibilities – an understanding that we did not have before. Knowing the likely minimum, maximum and the standard deviation paints us a picture of the output’s distribution or sensitivity of the variables. In fact, we can graph this information to uncover an even deeper understanding of the nature of the risk on this project.
Here, bankers can see that the loan is expected to break the debt service coverage test approximately 325 times out of 1,000, or 32.5%. That means that almost a third of the time, the bank will have to spend extra time on this loan trying to improve performance. Approximately, five percent of the time this loan is statistically will not meet its debt service obligations. Is that risk worth the spread that is being charged? Maybe.
Regardless, this probability distribution is an indication of risk. Bankers need to ask themselves if they feel comfortable with that number. If not, the banker must contemplate how the loan could be better structured to reduce that probability?
This information is invaluable because if the bank has other options, they might choose to underwrite a loan where the managers have demonstrated better expense control and where rent growth was above 2%. In this case, risk drops by more than half and the probability of breaking the 1.25x threshold falls to 12%, with the average debt service going to 1.61x.
Alternatively, because of the debt service sensitivity, the lender may want to increase both the maturity and the amortization to reduce the debt service burden. In this case, the longer set of cash flows increase the profitability of the loan, and while the loss given default would increase, the probability of default would be reduced thereby offsetting the increased risk.
Using This Analysis to Add Value to the Borrower
As valuable as this analysis is for a bank’s underwriting, it is even more valuable to the borrower. Many real estate investors lack the sophistication to run this type of analysis. Bankers can use this model to sit down with investors and walk through the assumptions and the projections.
Borrowers can understand that if they increase the terms of their leases, then they can reduce the revenue boundaries and make the property more stable. Germane to this market, the borrower can also understand how the entrance and expected exit capitalization (cap) rates will impact their investment. If the property is too expensive, this model will help uncover that fact upfront.
Both lenders and borrowers can play with this model to see if a property is more sensitive to the capitalization rate or free cash flow given the loan structure, the property, and the market. By adjusting each variable and looking at the output, sensitivities can be isolated.
If your bank is looking to build out its relationship banking capabilities, having tools like this can help cement your bank into more of a “trusted advisor” role. Few other bankers, accountants or lawyers will be able to run a Monte Carlo simulation for your borrower and having both this understanding and tools can set your bank apart.
A Simple Model
In order to demonstrate probabilistic underwriting, banks can find the model we used for this analysis in our Resource Center, free for you to download and use. Part of the beauty of this model is that it doesn’t require any “add-ins” to generate the Monte Carlo simulation so any bank can use it.
Instructions and documentation are on the first tab with inputs and outputs located on the second tab. Bankers can modify the inputs to run their multitenant investor CRE loans to more accurately understand the risk profile of each property.
Please find the model: HERE
Submitted by Chris Nichols on June 12, 2018