Publications

Introduction

In a recent paper¹, we looked at the effect of forecasts on IFRS-9 and Current Expected Credit Loss (CECL) deductions. This paper goes a step further to explore what lenders can do with deal structures to limit the expected losses. This can be viewed in three ways:

• Approaches to reduce the CECL deductions
• Approaches to reduce the actual risk
• Approaches to arbitrage anyone who does not take these features into account

In the previous paper, we looked at several variations on a commercial real estate deal under three forecasts. In this paper, we take the most-risky combination and show the effect of changing the loan structure. The stylized 'moderately negative' central forecast used was as follows:

• Interest rates, inflation and vacancy rates increase 0.5% per year for three years and then flatten
• Capital values and rental rates fall 5% per year for three years and then flatten

The loan is towards the risky end of the spectrum. It is for a Boston office with an LTV of 75%, initial DSCR of 1.35, floating-rate interest, 20-year amortization and a maturity of 5 years. There are four equal leases with a 1% probability of tenant default, one lease expiring in year three, the others expiring after year eight, and all leases repricing to the market annually.

Analysis

Using the Specialized Finance System's cashflow simulation, we quantified the effect of six variations in the financing structure:

• Reduced loan amounts
• Alternative amortization profiles
• Interest rate caps
• Sweep amortization (i.e., also taking all excess income to pay-down the loan)
• Sweep triggered by covenants on Loan to Value (LTV)
• Sweep triggered by covenants on Debt Service Coverage Ratio (DSCR)

The details of each variation are described later. Let us first look at the overall results. The table shows the net present value of expected loss as a percentage of the initial balance. To give deeper insight, the graphs show the projected annual CECL over 6 years for each of the variations. The annual loss profiles show the balance between market risk, lease risk, and refinancing risk.

NPV of Loss Depending on Financing Structure

Annual Profile of Expected Loss Depending on Financing Structure

Reduced Loan Amounts

As one would expect, reducing the loan amount reduces the risk significantly, and it is interesting to note that, for this deal, as the loan amount reduces, the risk shifts towards the later years.

Alternative Amortization Profiles

The four amortization patterns quantified here are:

• Traditional mortgage (with floating principle and interest (P&I) payments)
• Fixed P&I (whereby the principle payment is reduced if rates rise so that the sum of P&I is fixed)
• Interest only (I/O), and
• Interest roll-up

For this deal, Fixed P&I produces the lowest risk. At the other extreme, the effect of interest roll-up is to eliminate payment defaults during the life of the loan, but at the cost of a large spike in refinancing risk.

Interest Rate Caps

The two variations for interest rate caps are a cap until loan-maturity and a cap extending two years past maturity. Given the 'moderately negative' forecast scenario, with increasing rates, the caps significantly reduce the payment-risk during the life of the loan and the extended cap also reduces the refinancing-risk.

Sweep Amortization

The sweep amortization structures use any excess income to pay-down the loan. The base-case is a traditional mortgage with no sweep. Adding a sweep to the traditional mortgage reduces the peak risk somewhat, but there are still many occasions where the loan defaults because there is not sufficient income to pay the required amortization. The third structure requires payment of interest only, plus a sweep of any remaining cash. This structure greatly reduces the probability of early payment default, with little extra refinancing risk. If such a sweep is acceptable to the borrower, it can be priced with the same risk-cost as the 65% LTV deal.

Covenants

The last two sets of results show the effect of having covenants to require sweep payments only if the LTV or DSCR hit certain thresholds. In both cases, the covenants are only modestly effective because in those scenarios where the deal has deteriorated enough to trigger the sweep, there is little spare cashflow available.

Conclusions

Beyond the details of the property, the details of the financing structure give lenders many opportunities for tailoring the risk profile. The extent of the effect is different for each underlying property and can be quantified with the same tools used for CECL.

With capital and accounting both depending on the measurement of risk, this analysis shows that there are plenty of opportunities for arbitrage and mispricing.

Dr. Chris Marrison
CEO, Risk Integrated
Chris.Marrison@RiskIntegrated.com

1. The Effect of Forecasts on IFRS-9 and Current Expected Credit Loss