SecProbe.io

CORRESP
1
filename1.htm

​

January 26, 2023

Division of Corporation Finance

U.S. Securities and Exchange Commission

Washington, D.C. 20549

Attention: Mark Brunhofer and Sharon Blume, Office of Finance

​

 ​

Re:

 Burford Capital Limited

​

 Form 20-F for the Fiscal Year Ended December 31, 2021

​

 Filed March 29, 2022

​

 File No. 001-39511

​

 ​

Dear Mr. Brunhofer and Ms. Blume:

​

Burford Capital Limited (“Burford”, the “Company” or “we”) has today submitted to the U.S. Securities and Exchange Commission (the “SEC”), via EDGAR, this letter setting forth Burford’s responses to the comments of the staff of the SEC (the “Staff”) contained in your letter, dated January 12, 2023 (the “Comment Letter”), relating to Burford’s Annual Report on Form 20-F for the fiscal year ended December 31, 2021 filed with the SEC on March 29, 2022.

The numbered paragraphs and headings below correspond to those set forth in the Comment Letter. The Staff’s comments are set forth in bold, followed by Burford’s response to such comments.

Form 20-F for the Fiscal Year Ended December 31, 2021

Note 2: Summary of significant accounting policies

Fair value hierarchy

Valuation processes

Valuation methodology for Level 3 investments, page 85

​

 1. We acknowledge your response to prior comment 5 and our discussion on January 4, 2023 where you stated your valuation approach is consistent with the principles of the income approach, as described in ASC 820-10-35-24A. As it relates to the initial valuation of your capital provision assets, please respond to the following:

​

 ● Tell us whether when determining the terms you are willing to offer for contractual rights to litigation settlements or court judgments, you perform a probability-based outcome analysis of potential cash outflows to be received under the proposed terms that explicitly applies a discount rate to take into consideration the differences in timing of the potential cash flows. For example, tell us whether your methodology discounts for time value the potential cash inflows that would come from a future litigation settlement based on the expected timing of such a settlement or stage of litigation, and if so, provide a representative example of how this is done.

​

​

​

 ​

 ​

​

​

​

1

 ​

​

 info@burfordcapital.com

www.burfordcapital.com

​

When determining the terms we are willing to offer for contractual rights to litigation settlements or court judgments, we do not use a discount rate for cash flows given the highly uncertain and unpredictable nature of the duration of individual litigation matters. Rather, we arrive at our commercial terms based on our assessment of litigation risk as well as the competitive dynamic at issue. Moreover, our terms usually contain a time-based element which protects our returns in the event of a long duration outcome. We perform initial probabilistic modeling to ensure that, under all realistic scenarios when resolution could occur, we maintain our expected returns given the proposed terms and based on the litigation risk we assess.

​

The primary driver of pricing in the legal finance industry is the risk of loss of the capital deployed. The offer of litigation finance is generally non-recourse. In other words, if the litigation does not produce any proceeds, the counterparty owes nothing to us as the capital provider and there is a total loss of the capital deployed. Because there is an inherent unpredictability in each outcome given that each matter is before a different decision-making tribunal each time, there is material risk of loss in every capital provision asset of the entire capital deployed.

​

In addition to litigation risk, our terms are generally influenced by general perceptions of duration, even though they are likely to be inaccurate at an individual level. For example, we will generally seek higher pricing (whether through an increased multiple and/or a higher back-end share of proceeds) for an international arbitration matter at the World Bank than for a commercial case in US Federal court, because the former is likely to take much longer than the latter, even though there are meaningful differences among individual jurists or tribunals. Nonetheless, those differences in pricing are expressed in blunt terms, such as in different multiples of capital, rather than applying the impact of a time-based discount rate. On the other hand, we would generally not seek different pricing for the same case being brought in US Federal court in the Northern District of Florida, which on average has a very fast docket (12 months at present), or in US Federal court in the Eastern District of New York, which on average has a very slow docket (54 months at present), because our particular case in front of our particular judge may not perform anywhere close to those averages; indeed, we regularly price matters without knowing which court they will be filed in. Further, the annual interest-style rates applied in our deal structure (see below) have not increased from prior levels despite the recent higher interest rate environment.

​

Said another way, in assessing how to price a particular capital provision asset, we do not discount cash flows or create a net present value; rather, we use deal structures that produce varying results depending on duration with a focus on internal rate of return (“IRR”) and return on invested capital and we satisfy ourselves that we will generate an acceptable return under any set of likely durations and outcomes other than a loss. For example, it would be common for us to structure a deal as follows:

​

 o First, the return of deployed capital;

 o Second, a multiple of that deployed capital beginning at 1.0x and rising by 0.5x per annum to a cap of 2.5x, and an annual interest-style rate thereafter (with the rate being set based on risk and negotiating leverage); and

 o Third, a 15% share of net proceeds.

​

This kind of structure recognizes the potential for significant variability in a case’s duration and essentially compensates us for accepting that duration risk; it also makes it inapplicable for us to consider a traditional discount rate approach in our modeling when determining the terms we are willing to offer.

​

We note that our approach to pricing our initial terms in a capital provision rights agreement is consistent with the enormous US contingency fee market, which often does not vary the level of contingent fee based on duration. For example, a

​

​

 ​

 ​

​

​

​

2

 ​

​

 info@burfordcapital.com

www.burfordcapital.com

​

comprehensive study of contingent fees in New York across almost 300,000 cases determined that “[a]ttorneys’ fees were exactly one-third of net recovery in most cases”, without any allowance for the duration of the case’s resolution1.

​

 ● Tell us whether any probability-based outcome analysis you perform when deciding to enter into a capital provision rights agreement can be reconciled to the initial purchase price paid for the capital provision asset. If not, tell us why not.

​

We have not historically performed a reconciliation or calibration of our probabilistic models to the transaction price for the capital provision assets as the models are not designed for, or used in, our assessment of fair value. The purpose of our probabilistic models is to estimate likely recoveries and at what point in a litigation or arbitration recoveries could be achieved. Indeed, we do not apply a discount rate to the output of our models given the challenges associated with predicting litigation duration as discussed above and, when we publish the aggregate output of our probabilistic models, we make it clear that we are only providing that output on a nominal, and not a present-valued, basis. The starting point of our modeling is our entry into the relevant transaction, and we compute probabilistic-weighted profits from that starting point. Moreover, our probabilistic models are already risk-weighted and thus any appropriate discounting of their outputs if undertaken would initially appear to be at a risk-free rate. Discounting our probability weighted outcomes at a risk-free rate (or our weighted average cost of capital) would yield a higher value at the time of transacting than the initial purchase price paid.

​

An example may assist. While our actual probabilistic models are more complex than this, assume that we assign probabilities to a case losing, settling and winning. The probabilistic model’s output (which is its probability weighted computation of those three outcomes) is thus already reflecting the risk of obtaining the output. While we could then discount that output back using a risk-free rate for purposes of time value of money, or even consider the application of a discount rate that includes a factor for our cost of capital, it would not be appropriate to add a further risk premium into the discount rate as the model output has already been reduced for risk.

​

Were we to repurpose our probabilistic models for fair value purposes and calibrate the modeled outcome to the initial purchase price, we could theoretically do so by solving for a discount rate that would result in a net present value of the probabilistic model cash flows that equals the initial purchase price. However, an analysis for purposes of responding to this letter suggests that such implied discount rates would generally be 30% or higher. This is not surprising given that the Company has historically generated returns of around 30% IRR across its portfolio.

​

The Company has prepared an illustration of how the calibration would theoretically work on a simplified but representative example of one of the Company’s capital provision assets that is based on the following basic assumptions:

​

 ● Initial upfront funding of $10 million and no further funding is provided; and

 ● The Company is entitled to its deployed capital back plus a 2.5x multiple return payable only out of any proceeds successfully realized from the underlying litigation event.

​

At the capital provision asset’s inception, the implied discount rate to solve for a present value of the probabilistic model’s expected cash flows that is equal to the transaction price is 31.9%. The illustration then further assumes the following occurs during the uncertain path and duration of the litigation to arrive at an ultimately successful outcome:

​

1 Eric Helland, Daniel Klerman, Brendan Dowling, and Alexander Kappner, Contingent Fee Litigation in New York City, 70 Vanderbilt Law Review 1971 (2017).

​

​

 ​

 ​

​

​

​

3

 ​

​

 info@burfordcapital.com

www.burfordcapital.com

​

 ● Inception to March 31, 2022 – there are no adjudicative events and no revisions to any model assumptions during this period. The inputs for the probabilistic model and fair value for the capital provision asset remain unchanged.

 ● April 1, 2022 to June 30, 2022 – an adjudicative event occurs in the form of a trial win and an award issued in favor of Burford’s counterparty. Both the probabilistic model and the fair value of the capital provision asset are updated to reflect this event, the latter recognizes 60% of the expect profit.

 ● July 1, 2022 to September 30, 2022 - there are no adjudicative events and no revisions to any model assumptions during this period. The inputs to the probabilistic model and fair value for the capital provision asset remain unchanged.

 ● October 1, 2022 to December 31, 2022 – an adjudicative event occurs in the form of confirmation that the Burford’s counterparty’s adversary’s as-of rights to appeal the outcome have lapsed. Both the probabilistic model and the fair value of the capital provision asset are updated to reflect this event, the latter now recognizes 80% of the expect profit.

 ● January 1, 2023 to March 31, 2023 – the cash is settled during this period.

​

The graph below illustrates the calibration at inception and then how the fair values implied by the calibrated model would compare to the Company’s assessment of fair value under the current valuation policy following these assumed events.

​

The results demonstrate the significant difference in fair values resulting from the approach based on a calibrated discounted cash flow model versus the approach the Company, and a market participant, would employ to assess the price they would pay to acquire the capital provision asset in an orderly transaction between market participants at the measurement date. This is because the fair values derived from the former approach would imply that a market participant would be willing to pay significantly more for a capital provision asset as it approaches each adjudicative event simply because of the passage

​

​

 ​

 ​

​

​

​

4

 ​

​

 info@burfordcapital.com

www.burfordcapital.com

​

of time at a robust discount rate, even though nothing more is known about the likelihood of the outcome of that adjudicative event and there remains continuing uncertainty over the timing of when it may be known. This is not the case in practice; using the above example, Burford could not sell a capital provision asset for 30% more than its initial purchase price after a year simply because of the passage of time if no adjudicative event had occurred in the interim. This is why we do not believe using our probabilistic modeling for valuation purposes at this stage of its evolution is appropriate.

​

 ● In response to prior comment 3, you stated that you are a market participant under ASC Topic 820. To the extent you consider the impact of the time value of money in determining the terms for entering into a capital provision rights agreement, tell us whether you believe there would be any basis for not considering changes to the time value of money (for example, market interest rate changes) on the fair value of the capital provision assets each reporting period.

​

Refer in general to response in bullet 1.

​

 ● Tell us whether you factor in other potential risks, such as credit risk or foreign exchange risk, in determining the terms you would offer to initially acquire a capital provision asset, or in subsequent fair value measurements. If not, tell us why you do not believe a marketplace participant would consider this information.

​

Litigation outcomes stand apart from the remainder of the conventional credit universe because they do not arise as a result of a contractual relationship between the judgment debtor and creditor, unlike essentially all other forms of credit obligation. Thus, for example, for a debtholder to recover on a defaulted debt, there are many steps involving notice, a cure period, usually a subsequent judicial or insolvency proceeding that will generally sweep in other creditors and a meaningful risk of the debt being impaired or compromised. By contrast, a judgment creditor has immediate and unfettered rights of action to, for example, seize assets and garnish cash flows. Thus, conventional debt metrics like a decline in credit ratings or a decrease in debt pricing do not generally read across to litigation collection outcomes.

​

Moreover, foreign exchange risk is often dealt with in the damages phase of a case. For example, if a US company is damaged because of some interaction with, say, a Chinese company, the US company’s damages will be computed in terms of its US dollar losses, and if the Renminbi has depreciated in the interim, that will be something taken into account in computing those damages. Additionally, where we invest in a capital provision asset outside of the USA (or when the investment is denominated in other than USD), at each reporting period, the investment value in local currency is translated into USD at the then spot rate (an exit price).

​

Above we described that it is litigation risk that fundamentally drives the initial terms of our transactions. To the extent that this question asks about financial risks other t