Cash flow analytics for CDO transactions.

• Author: Albulescu, Henry C.
• Position: Collateralized debt obligations

The exact capital structure for cash flow CDO transaction, or for synthetic CDO transaction with cash flow components, is determined by modeling cash flow simulations under different assumptions. The aim of this analysis is to show that each tranche can withstand the stresses commensurate with the desired rating.

1. INTRODUCTION

   The exact capital structure for cash flow CDO transaction, or for synthetic CDO transaction with cash flow components, is determined by modeling cash flow simulations under different assumptions. The aim of this analysis is to show that each tranche can withstand the stresses commensurate with the desired rating. Relevant parameters incorporated in the cash flow analysis include:

   * Transaction payment priority and triggers;

   * Intrinsic cashflow characteristics of the assets;

   * Default rate--the expected level of gross defaults;

   * Default timing--when defaults will occur;

   * Default patterns--pattern of defaults that will occur once defaults start;

   * Recovery timing--when recoveries will be achieved after a default occurs;

   * Recovery levels--amount of the recoveries achieved;

   * Interest rate curves--different interest rate paths; and

   * Hedge structures.

   Each of these above parameters is fully discussed further in the article.

   Cash flow analysis is aimed at evaluating the availability of funds for full payment of interest and principal in accordance with the terms of the rated securities. If a transaction has multiple tranches, cash flows will be run for each tranche to assess whether the level of credit support provided is consistent with the rating sought on each tranche. Cash flow analysis also is used for sizing liquidity and other reserves. The analysis takes into account the structural elements of a transaction, including the principal and interest payment allocations; early amortization, "fast pay", or redemption events; excess spread accumulation; and reserve levels.

   CDO cash flow analysis takes into account application of available cash flow to pay down notes based on a transaction's priority of payments. Most arbitrage cash flow CDO transactions are structured with a five-year reinvestment period and a seven-year amortization period. Synthetic cash flow transactions may also have reinvestment periods and amortization periods, but their length may vary. Mostly outside of the US a good number of transactions are also structured as amortizing static pools that just pay down.

   Cash flow transactions typically have coverage tests which function as early or rapid amortization triggers. Coverage tests in arbitrage deals are the "O/C" or par value ratio (for example, collateral principal and cash balances divided by the rated note principal balance must equal at least 115%), and the interest coverage or "I/C" ratio (for example, collateral interest receipts divided by the rated note interest payable and senior expenses must equal at least 150%). When these O/C and I/C ratios are breached, the structure should trigger and the cash flow model should reflect the way in which the transaction is structured.

   Principal payments during the amortization period may either pay down the liabilities if the triggers are breached, or may be reinvested if they maintain or improve the triggers. A structure that does not shut down reinvestment to pay down notes to maintain the trigger O/C and I/C ratios, but rather permits temporary noncompliance and subsequent improvement, may increase risk over time. The reason for this is that each subsequent reinvestment may lead to additional defaults rather than paying down the rated liabilities. If a structure permits reinvestment of these proceeds in order to come closer to compliance for an O/C or I/C test violation, then the cash flow model must accurately model this, and not assume pay down.

   CDO cash flow analysis normally does not take into account collateral prepayments, either from optional redemption of bonds or from unscheduled amortization of loans. Such prepayment analysis would require a complex model that both generated interest rate term structures and forecast corporate bond and loan prepayment rates across yield curve evolutions. In addition, interest rate models are unreliable over more than very short horizons. This, coupled with a portfolio's own deviation from aggregate behaviors, would tend to make this type of modeling exercise less than productive for analytical purposes. Given the uncertainty associated with interest rate modeling, cash flow models also cannot currently account for price depreciation risk on performing assets traded out of the portfolio. Erosion of par through trading is captured in the transaction by the O/C trigger, but cannot be effectively modeled to capture par erosion above where the O/C trigger is set.

   Because cash flow transaction structures vary, a standard cash flow model is not used for all deals. Instead, analysts rely on the transaction-specific proprietary cash flow model prepared by the sponsor or its advisor. Each cash flow model is evaluated to assess whether it accurately reflects the transaction structure, and can measure a variety of risk factors for collateral assets, and debt or equity liabilities.

   Examples of such risk factors include payment terms of the collateral versus debt; interest rate mismatch (for example, fixed-rate collateral versus floating-rate debt); interest rate risks arising from multiple loan indices, payment frequencies, and different amortizing schedules of each asset; and a variety of delinquency, default, and recovery risk scenarios. A qualified, independent accounting firm should review the cash flow model results, and ideally "tie-in and tie-out" the model, in order to verify in writing that the model properly reflects and analyzes the transaction structure and relevant risks. This verification should be performed in accordance with an agreed upon procedures letter. Exhibit 1 illustrates a typical set of assumptions for a CBO/CLO cash flow model.

   In general, the cash flow analysis is based on collateral pool parameters specified by the transaction documents. At closing, the transaction may not have purchased all the collateral that is required. Typically, at closing, the majority of the portfolio has been purchased with remaining balances to be purchased after closing during the ramp-up period. These periods are typically three months, but in certain deals they extended as far as one to two years.

   Cash flow models cannot capture the degree of risk associated with not being able to fully ramp-up. While failures to ramp-up to the required collateral balance or collateral pool characteristics have not happened very frequently in the past, a ramp-up period nevertheless does pose an additional element of risk to investors. If the transaction is a static pool, then all the asset characteristics and actual payment terms may be incorporated in the cash flow models. For deals that are not fully-ramped-up at closing, the transaction is modeled incorporating conservative assumptions that the collateral pool is at the minimum weighted averages allowed by the transaction documents, instead of the typically higher rate available in the market.

   For the most part, CDO transactions are modeled based on the aggregate characteristics of the asset pools, and not using asset specific models. Thus, defaults are taken pro-rata across the asset pools and payments are reduced uniformly as defaults occur. Nevertheless, there are transactions that utilize asset specific cash flow models and have sufficient specificity to permit biasing of default across specific assets. Asset-specific cash flow models are preferred for asset pools where the assets have unique payments characteristics that are not fairly represented by average pool characteristics.

   In modeling the cash flows, the sponsors or their bankers typically perform the modeling. The results of the cash flows typically are the maximum "Break-even Default Rate" (BDR) that the transaction can withstand at each set of input parameters. If the BDR is greater than the Scenario Default rate calculated by the CDO Evaluator (see "CDO Evaluator and Portfolio Benchmarks" section), then that tranche can achieve that rating.

   For example, if the 'AAA' scenario default rate is 30% and the cash flows for that tranche show that the transaction can sustain a BDR of 33% to 39%, then that tranche can be rated 'AAA'. This means that the transaction is sufficiently robust to sustain up to 33% defaults over its life. Since 33% defaults are greater that the 30% expected default rate under the 'AAA' stress, that tranche then can achieve that rating, In general, each tranche has a range of BDR, since multiple cash flow runs...