Why Recovery Timing Matters as Much as Recovery Rate

Recovery rate is the number that gets quoted. It is the number in the teaser, the number in the investment committee summary, and usually the number two investors compare when they talk about a portfolio. On its own it prices nothing, because it says nothing about when the cash arrives or what it costs to collect.

The number everyone quotes

A recovery rate expresses total expected collections as a percentage of the claim. It is a useful summary statistic and a poor pricing input, for the same reason that revenue is a useful summary of a company and a poor basis for valuing one: it omits the two things that determine value, which are timing and cost.

Two portfolios can carry the same expected recovery rate and differ in value by a wide margin. A portfolio of secured exposures where borrowers are engaged and settlements are being agreed produces cash early. A portfolio of similar claims where every file requires enforcement produces the same nominal total, spread across years, net of costs that accrue throughout.

An illustration

The arithmetic is worth doing once, with round numbers chosen for clarity rather than drawn from any portfolio.

Take a claim of 100. Assume both cases recover 40 in total, gross, and both carry a required return of 15 per cent.

In the first case, the 40 arrives evenly across two years. In the second, it arrives evenly across six. Before any cost, the present value of the first is materially higher — the later cash flows in the six-year case are discounted through four additional years of compounding, and at 15 per cent the final year's collections are worth well under half their nominal amount.

Now add cost. Enforcement, legal fees, servicing and property holding costs do not scale with recovery; they scale with time and with the number of files being worked. The six-year case carries roughly three times the duration of cost accrual, so the same gross 40 arrives as a materially smaller net figure.

The result is that two portfolios with identical recovery rates support very different bids. The difference is not a refinement at the margin. It is frequently larger than the difference between a 35 per cent and a 45 per cent recovery assumption on the same timetable.

Where timing comes from

Timing is not a portfolio-level input. It is a property of each file, and it is driven by a small number of observable facts.

The procedural stage the file has already reached is the strongest single indicator. A file where judgment has been obtained and the property is scheduled for auction sits years ahead of a file where proceedings have not been commenced, and the gap between them is not an average — it is a specific sequence of steps, each with its own duration.

Whether the borrower is engaged matters next. Consensual outcomes settle faster than contested ones, and the difference compounds because a contested file also consumes cost throughout.

Whether the borrower is in an insolvency procedure changes the timetable structurally rather than incrementally, because the procedure imposes its own sequence and its own priority ordering on any distribution.

Finally, the asset itself sets the disposal timetable. A residential unit in a liquid urban market clears on a different schedule from a single-purpose industrial building with a small pool of possible buyers.

Costs run on a clock

The point that is easiest to miss is that most of the cost base in a distressed portfolio is a function of duration, not of recovery.

Servicing fees are typically charged as a periodic amount or a percentage of collections, and a file that stays open for six years is serviced for six years. Legal costs accrue as procedural steps are taken, and more steps are taken on longer files. Property taxes, insurance, security and maintenance on repossessed assets run monthly regardless of what the asset eventually sells for. Currency and rate exposure, where the investor funds in another currency, also runs with time.

Modelling costs as a single percentage deduction against gross recoveries — which is common, and quick — systematically understates the cost of the slow files and overstates the cost of the fast ones. That error moves value in the same direction as an optimistic timetable, so the two compound rather than offset.

How we model it

We attach a timetable to every recovery path rather than to the portfolio. For material exposures the timetable is built from the procedural stage of the actual file. For segments underwritten statistically, it is built from the observed stage distribution within the segment, with the uncertainty stated as a range rather than smoothed into a point estimate.

Costs are then split into those that scale with recovery — success fees, brokerage, disposal taxes — and those that scale with time, which are accrued period by period against the files still open. Only after that are cash flows discounted at the investor's required return.

The test we apply is simple. If the timetable slips by two years across the portfolio, does the transaction still clear the return threshold? If a bid only works on the base-case timetable, it is not a bid on a distressed portfolio. It is a bet on the courts.