Arrow Global Group ARW.L S
December 12, 2017 - 4:36pm EST by
l2kcapital
2017 2018
Price: 378.00 EPS 0.17 0
Shares Out. (in M): 179 P/E 22 0
Market Cap (in $M): 905 P/FCF 0 0
Net Debt (in $M): 1,242 EBIT 109 0
TEV ($): 2,146 TEV/EBIT 14.7 0
Borrow Cost: Available 0-15% cost

Sign up for free guest access to view investment idea with a 45 days delay.

  • debt purchaser
  • going to zero eventually
  • Complex Accounting
  • Aggressive Accounting
  • Short to zero
  • npl
 

Description

 

A quickly deteriorating business model with flawed economics and suspicious accounting
Prices here are based on 391p at time of writing

Introduction of opportunity
Debt purchasers (“DP”) essentially play a game of arbitrage between the rate of return they get on their portfolio investments and the interest rate they raise capital at. However, returns have been compressed, and interest rates only will move in one direction going forward, with further headwinds (regulation, potential accounting scrutiny, cash burn) that could cause the cost of their debt to increase at further issuances. The valuations of such companies only make sense if you believe that future NPLs can be purchased at more attractive prices, similar to what was achieved in the immediate years after the 08 crisis. Given today’s fundamentals, such companies are too levered, consistently burn through cash and are destroying value while facing an increasingly hostile regulatory and macroeconomic environment. The sector uses complex accounting, misleading metrics/ratios and numerous sector-specific acronyms which have misled regulators, accountants, rating agencies, sell-side research and investors. We believe Arrow Global Group plc (“Arrow” or the “Company”) has little to no equity value.

Business model

-          DPs aim to help debtors deal with their debt problems and repair their credit standing with long-term sustainable repayment plans. These are mostly NPLs, already defaulted loans with a mix of paying and non-paying accounts

-          Banks and other credit institutions are less efficient at handling and collecting debts, it is not their core focus. As such, it makes sense for them to sell these to debt purchasers who are better at extracting value via relationships and data/analytics (accurately predicting future collections). Furthermore, such institutions are not keen to give haircuts as it would damage their core business model, whereas DPs factor this into account when acquiring portfolios

-          Establish and maintain good relationships with a variety of debt sellers (banks, FIs, utilities, telco, retailers…) to maintain a pool of portfolio investment opportunities

-          DPs have, to varying degrees also debt collection and asset management businesses (3rd party debt servicing)

Income statement and accounting of debt purchasers explained

Source: Arrow Group Global plc investor presentation.

-          1. Debt purchasers price their portfolio investment/acquisitions based off a gross money multiple

-          2.  The sum of the expected gross (usually, unless specified as net) collections from the portfolio gives the estimated remaining collection (“ERC”), this is usually measured either on a 7 years basis (84M) or 10 year basis (120M), with some residual claims beyond that period

-          3. The collections generated from portfolio per period is then classified into 2 categories: a) income; and b) portfolio amortisation. There are also write-ups/revaluations which will also factor into revenue. Therefore, if the company deems more of its collections from a portfolio to be income rather than amortisation, it can achieve 2 things à a higher ERC going forward (since the balance of ERC is decreased by a lower amortisation number) and a higher revenue number (this also leads to higher net EPS and ROE)

-          4. These are the costs associated with collecting any income from the portfolio. E.g. staff costs, legal costs, bailiffs etc. Benchmarking cost-to-collect ratios is not simple due to different categorisations of costs by competitors

 

Source: Hoist Finance investor presentation.

Why does the opportunity exist?
Market view

-          Arrow has repositioned itself as a largely UK debt purchaser with a single asset class to a leader across 6 countries with its own servicing capabilities covering half the addressable European NPL market consistently outperforming its financial targets

-          The market views the company as an extremely high adj. EBITDA margin business (60%+) with low capex requirements (c.3% of sales) and limited working capital requirements

-          The company is able to consolidate at attractive multiples (Zenith acquired for a headline price of 5.3x LTM EBITDA) as well as extracting meaningful synergies (higher rate of collections) from the use of their proprietary, sophisticated database that they have developed

-          Data-driven analytics (a sophisticated complex process with reams of data-access required) company in a fast growing market (double-digit forecasts), as opposed to just a Business Process Outsourcing (“BPO”)-type business, with attractive fundamentals (market looks at P/E and ROE). Collection on portfolios have consistently exceeded management estimates, with the ongoing promise of quick deleveraging if proceeds are not reinvested (gross collections in next 3 years are c.60% of 84M ERC)

-          The market is expected to continue to benefit from increasing deal flow of NPLs sold by banks in Europe due to increasing regulation (ECB’s Asset Quality Review, Basel 3 and IFRS9)

-          Debt purchasers are able to benefit in good and bad economic periods (recession-resilient/counter-cyclical). In bad periods, the number of NPLs should increase and be available at lower prices (banks under pressure to sell) but default rates are higher and proportion of debtors paying are lower. In good economic times, debtors have higher payment capacity, NPL portfolios will experience lower defaults and increase in supply due to increased consumer lending

-          The accuracy with which debt purchasers predict collections has been incredibly high and consistent (101-103% for Arrow, comparable and always above 100% for competitors), highlighting their data and analytical prowess

-          There are significant barriers to entry such as relationships to debt sellers, databases for NPL bidding, regulatory compliance and economies of scale

-          The ratings agencies have also shared this view, with the company being able to finance its debt at a rate of c.380bps (debt trades above par). 2 main leverage metrics used are net debt / adj. EBITDA and asset-backing (on gross ERC), which are both seen growing at attractive rates

Investment thesis / Our view vs. consensus

-          The company is more akin to a finite-life business. The company is an asset-heavy platform and not a going-concern unless more portfolios are added, otherwise it is in run-off mode. The investment required to “replace[1]" (similar to maintenance capex) what is collected is a lot higher than the market seems to think. Adjusting this, the fundamentals look totally different. The adj. EBITDA that the company shows (which the market and ratings agencies use) ignores the cost of acquiring the debt that is being collected, i.e. not recurring and not indicative of profitability

-          The fundamentals reported have also been deteriorating when looking at the expected gross cash collections money multiples, which is a sign that the company is likely facing more competition (having to bid higher) for similar quality level portfolios. A consistent picture can be seen across the European industry (see appendix).

-          Cost to collect ratios (which we believe is under-stated by the company by excluding staff costs, of which over 50% work in collections) have also been rising, leading lower net cash collections. The gross view on multiples does not take into account the time value of money (exacerbated by the high interest costs) and the gross collections number can be manipulated through accounting assumptions and/or through continuous M&A[2]. M&A expenditure used to acquire portfolios of NPLs should be treated the same way as costs to acquire portfolio (partially expansion and maintenance capex). M&A also helps to cloud organic performance & comparability

-          Management is highly incentivised to grow EPS given compensation arrangements and has the tools to manipulate this and has been selling shares aggressively since their lock-up expired at the end of 2015

-          The marketed performance of debt purchasers to forecast collections on portfolios acquired (bear in mind these are long since defaulted NPLs) implies significant earnings visibility, which is hard to believe. Most portfolios are impossible to stress test, let alone forecast. The only way we know to be profitable in this business is to buy at very low prices to provide for a large margin of safety

-          Looking at the cash flow generation and profile since IPO (from January 1st 2014), it is clear that the company has essentially not generated any cash while growing its portfolio (gross ERC) by raising a substantial amount of debt at record low rates. Should there be any material movement in rates, lowering in profitability of operations (through a decrease in expected collections and/or higher cost-to-collection ratios), the company could quickly start to become a clear value-destructor (via need for additional debt raising). A bonus would be if ratings agencies were to reconsider their methodology of looking at leverage and asset-backing

-          Sell-side reluctant to provide more scrutiny given fees provided by the sector to investment banks and focuses on EPS and ROE (both of which can be manipulated)

-          Regulation may possibly also be a headwind, with the “right to be forgotten” being a key possible short-term development expected in May 2018 (the General Data Protection Regulation replacing the Data Protection Directive 95 / 46 / EC)

Industry / market

-          In the last decade, the industry has benefitted from Europe’s banks being weighed down by almost a trillion euros of NPLs and under pressure from regulators to get those debt off their balance sheets

o    Debt sellers in the UK are selling the last of the backlog built up during the financial crisis, growth in recent years has come from vendors selling debt earlier in the cycle (fresher debt). This fact along with increased competition has led the price paid by Arrow per £ face value of debt to increase from 7p in 2013 to 13p in 2017H1

-          The ongoing consolidation of the debt purchasers market and their increasing influence (dictating prices on purchases of NPLs) has not gone unnoticed by regulators, with the Intrum/Lindhoff merger being forced to make a series of divestments in the Nordic region (both debt collection and purchasing activities) as a condition to close, which cuts the proposed cost savings by nearly a third

o    Furthermore, it has been reported that the EU is considering action to try to boost demand and increase sale prices for secondary market NPLs

o    25th May 2017 ECB started encouraging public-private partnerships to improve data quality and recovery processes to reduce asymmetries between buyers and sellers

-          Prices for financial institutions’ receivables (portfolios) have been increasing, and this has historically been the primary activity of debt purchasers. This is shown by the expected collection multiples falling

o    Pricing pressure, leading to lower gross multiples (before collection costs)
Intrum said in April it saw pricing pressure from increased competition in W. Europe

o    Increasing cost-to-collect ratios reported, leading to lower net multiples (post-collection costs), which may be due to lower quality portfolios (higher cost to collect per value of debt). The ratio has increased from 21.9% in 2013 to 34.2% in 17Q3 LTM (this includes some litigation costs, but excludes staff costs)

o    Higher prices being paid may also reflect the fresher debt with higher collectability being sold, as per Hoist’s experience in the last two years. However, bizarrely, Arrow has experienced higher cost-to-collect ratios as well as diminishing expected collection multiples

o    More types of credit being bought (such as telco, utility and retail receivables) as well as an increased focus on growing the “asset management” side, whereby Arrow services third-party portfolios. This business has increased for Arrow from 1.5% of sales at IPO to 22.3% 17Q3 LTM (£1.4m à £63.3m)
Market assumes this is very high margin, recurring business. However, research has shown that this is a low margin (c.10%), commoditised BPO-type business

-          The European market has large regional differences in market maturity and regulatory landscape (debt collection process, access to credit data…). Deploying large amounts of capital in other countries is not always easy due to a combination of lack of deal flow, high prices and an inability to price competitively

o    Italy is cited as the largest market for NPLs with low prices, however, this may be due to the fact that the quality of the portfolios are low. E.g. UniCredit estimated the price of recent deals would only allow it to recoup 23c on its bad loans, far below the 43c average for Italian banks à it went on to sell these at 13c

o    Cabot ended a JV in Spain after 1 year, Aktiv Kapital withdrew from numerous countries and EoS struggled to establish a leading position in Austria

-          Consolidation possibilities exist, with headline multiples looking highly accretive. Increasing regulation makes it harder for smaller competitors to meet all requirements while maintaining an acceptable level of profitability

o    1) headline / adjusted EBITDA is not a sustainable metric, as mentioned above and; 2) value creation only exists if the buyer of the portfolio is able to extract a higher collections from the portfolio

-          Market relies generally on consumer borrowing, which is highly sensitive to economic conditions

-          The debt purchasers have taken full advantage of record low interest rates to fund the majority of their portfolio purchases via debt and also helped them through the crisis. This situation is expected to reverse (nobody knows when, but we know that rates cannot go below where they are)

-          Arrow boasts about its database and analytics capabilities, however, all other competitors seem to have similar processes, expertise and infrastructure. This is somewhat confirmed by the similar rates of collection multiples that have been achieved (cost-to-collections have generally been rising, but hard to compare due to categorisation)

-          UK-centric debt purchasers will be hit by National Living Wage (increasing collection costs)

Valuation and analysis of deteriorating fundamentals

-          Yield compression: Arrow boasts about its data and analytics (consistently predicting actual collections on long-dated, long since defaulted portfolios to the tune of +/-3% historically), the development of sellers now selling fresher, more-paying debt is a negative development, since these will be with lower expected collection multiples (due to higher price and competition). This is a general trend observed also with competitors[3] (pre-2011 gross cash-on-cash multiples at 2.4-4.0x and the last 2 years at 1.7-2.0x). The unit economics are not profitable, nowhere close to what the market thinks/is pricing. Assuming 1.7-1.9x gross cash-on-cash multiples and then removing collection costs (which are also rising) of between 25-35%[4], you get a net cash-on-cash multiple of 1.11x-1.43x, this is before taking into account other real costs: overhead costs, interest costs (since most portfolio investments are financed by debt), taxes and the time value of money, since collections are over 7-10+ years

-          Adj. EBITDA a measure of profitability?: Arrow reports adj. EBITDA which is not a proxy for cash flow or profitability, as it ignores a key component of expense/investment (opex / maintenance capex) for the business to maintain its status as a going concern, the investment required to keep the portfolio value (ERC) flat, the replacement rate. Adjusting for this, the 17Q3 LTM / FY16 / FY15 EBITDA becomes 42.7 or 86.2[5] / 86.8 / 72.4 instead of 206.2 / 209.2 / 153.1. This highlights the problem that the company has been facing, turning it into a value destroying business (better off for investor that the company runs off the portfolio, as opposed to continuing to invest in it). The going-concern valuation analysis shows that the valuation not even enough to cover the debt, let alone having any residual value for equity holders

-          A sum-of-the-parts analysis: also supports such a hypothesis, taking the latest reported gross ERC profile with a cost-to-collect ratio of 28% and discount rate of 10%, the value of the current portfolio is £944.0m. The value of the asset management services division is estimated at £66.0m (by taking an EBITDA margin of 10.0% and assigning a multiple of 10.0x) à enterprise value = £1,010.0m. The latest reported net debt is £926.2m giving an equity value of £83.8m vs. the current market capitalisation of £698.7m (an implied downside of 88.0%)

-          Under-amortisation of ERC and inflation of revenue: Collections from portfolios are split into 2 categorisations, portfolio amortisation (which reduces the balance sheet value of portfolio investments) and income. The amount of amortisation implies the expected money multiple from a portfolio. E.g. a 50% portfolio amortisation implies a 2.0x expected money multiple[6]. The amount that the company amortises from the collections is entirely at the discretion of the company and it is impossible from the outside-in to assess this due to the large number of different portfolio vintages and limited data disclosure. Arrow expected c.1.8-2.0x gross money multiple on ERC, however, Arrow’s average portfolio amortisation rate has averaged 30.5% since 2013, implying an expected multiple of 3.3x. This means that ERC (used for asset-backing leverage) and revenues have both been overstated over the last 7.75 years (large cumulative effect of over £200m estimated at over 20% of net 84M ERC). Furthermore, by looking at the collections divided by the gross 84M ERC we get

-          Use of cash: since 1st January 2013 (4.75 years), we can see that the business has grown its portfolio (ERC) via large debt issuances and has generated practically no cash due to the combination of a) ongoing investments required to keep the portfolio replaced and; b) investments to acquire and enlarge the portfolio. However, as we have seen above, today’s portfolio is of a much lower quality and likely value destructive, therefore, the return on investments are likely extremely low, or even negative

Cash starting balance

Cash from operations[7]

Cash from investments

Cash from financing

Cash ending balance

£9.6m

-£140.9m

-£248.9m

+£414.6m

£36.2m

                        As per Arrow Global Group plc cash flow statement, from 1st January 2013 until 30th September 2017.

 
Furthermore, from FY10 until today (7.75 year span), operating cash flow has been negative in every period, with financing cash flows (predominantly debt raising) being the driver for cash

-          Changes in data disclosure: Arrow stopped reporting portfolio write-ups since 2015 and old vintage ERC splits. You would expect that the ERC for a vintage to decrease going forward (i.e. you collect something from these portfolios, so even with a slight mark-up, the net should be negative), however, there was a large mark-up in the ERC for 2011 and earlier vintages of £26.6m in 2015, and a slight mark-up of £0.4m for the 2010 vintage ERC from 2013 to 2014 but this is not shown since the company doesn’t disclose those vintages. For example the company does not show pre-2012 vintage ERC splits (stopped in 2015)

-          When looking at credit metrics, the company reports net debt to LTM adj. cash EBITDA of 4.5x and net debt to gross 84M ERC of 63.6%, however, if we adjust for replacement rate of EBITDA and 84M ERC net of collection costs, the ratios become 10.7x (generously using the company’s quoted replacement rate of £120m, instead of the worked out number of c.£164m) and 96.6% (implying little to no equity value, and comfortably exceeding/breaching the covenant set at 75.0%)

-          An aggressively growing portfolio, requiring an increasing amount of maintenance capex to maintain just the replacement rate, with capital being deployed at increasingly lower, or even negative returns

Risks

-          The collection of claims is dependent on general economic conditions, if the benign economic environment continues, the show can keep going on

-          Lack of concrete short-term catalysts

-          Debt purchasers are able to front-load collections to boost near-term earnings (to the extent allowed, and not already done), cash position (by negotiating a lump-sum amount that comes earlier with debtors) and also write-up the portfolio value (manipulating portfolio amortisations/inflating ERC) to keep metrics such as EPS growth and return on equity looking good for the market

-          Given the above, if the company can continue to use its adj. EBITDA metric for leverage purposes and raise debt to enlarge its portfolio. In this scenario, the only catalyst would be when the company is unable to pay interest and other operating costs

-          M&A activity:

o    The company continues to expand its markets and portfolio via M&A, at seemingly accretive headline multiples and the market assigns value to this

o    Given the amount of consolidation in the sector, it is not unthinkable for the company to be a takeover target for larger competitors (such as Intrum) and/or private equity sponsors (the sector has historically been a favourite for private equity, with reputable firms participating)

Views on valuation

The "real" EBITDA of the company (taking into account replacement rate, using a aggressive £120m) is only 41.8% of the adj. EBITDA that the company widely reports. The "real" leverage of the company is 10.7x LTM vs. the adj. 4.5x, with the asset backing (defined as net debt / net 84M ERC) at 96.6% as opposed to the reported 63.6% (which uses the gross expected remaining collections number). This leaves little room for value in the equity. We have not made an attempt to value any further M&A and portfolio purchases, which one could argue could boost the value of the Company slightly (although we have argued that given the current market dynamics and with interest rates only likely to move in one direction, the expected returns should be very low if not bordering on being value destructive). If we further add in to the fact that the ERC is over-valued, along with EBITDA (i.e. revenues recognised are higher than they should be due to the under-amortisation), then there leaves little value to the equity.

Company trades today at 7.9x LTM Adj. EBITDA (without taking into account the replacement rate) and 18.8x LTM "real" EBITDA. We feel that even a stretch valuation for such a Company (commoditised, high competition industry with deteriorating fundamentals) may be 12.0x "real" EBITDA. As such, one could argue that the debt is covered, leaving minimal equity value (1.3x EBITDA of equity).

 

Appendix

 

Summary of findings on KPIs and financing:

Note: Replacement rate used for Arrow is £120m guidance given by management on the latest earnings call, however, it is unclear whether this is a 9 month or full year number. The worked out replacement rate is £163.6m.

Expected and achieved gross money multiples by vintage (year portfolio was acquired)

Note: Dotted lines represent an average on all vintages on and before the year where the dotted lines end. Arrow previous vintages best estimates based on historical disclosure, no adjustments made.
Source: Company filings, estimates.

Competitive landscape and service offering


Source: J.P. Morgan research.

 

Arrow Global Group plc’s latest ERC profile and debt maturities

Source: Company 17H1 investor presentation

The purchased debt landscape

Source: J.P. Morgan research.



[1] Replacement rate = [estimated collections for the next 12 months – (estimated collections for FY10*(FY10/FY9)] / expected gross money multiple on portfolio. This formula takes into account the perpetuity as well as the decay rate of the estimated remaining collections of a portfolio.

[2] Revenue for debt purchasers includes the “income received from portfolio” which can be easily manipulated by the company. The total collections received in each period is split into a) income from the portfolio and b) portfolio amortisation. See Lowell Group’s Accounting Teach-in dated 15th March 2017.

[3] Intrum Justitia, Hoist, Cabot and Lowell Group.

[4] Based on Arrow historic ratio. Higher end adjusted to include a portion of estimated staff collection costs.

[5] The worked out replacement rate is £163.6m vs. company stated number of £120m on earnings call. Unclear if this refers to a 9 month number or a full year.

[6] Amortisation rate = portfolio amortisation / total collections. Expected gross 120M ERC money multiple = 1 / amortisation rate.

[7] Cash from as per company filings from 01/01/13 - 30/09/17. Operating cash flow includes portfolio acquisitions, collections and operating costs.

I do not hold a position with the issuer such as employment, directorship, or consultancy.
I and/or others I advise hold a material investment in the issuer's securities.

 

Catalyst

Potential catalysts

-          The continuation of deteriorating business fundamentals becoming clearer and clearer after every reporting period, with the key metrics being: gross money multiple, availability of debt portfolios on the market, lack of amortisation (lack of cash collections vs. expected multiple) and cost-to-collect ratios

-          Recently failed IPO of Cabot may lead investors to deep dive into the accounting and underlying fundamentals of the business. Cabot cited a “high level of engagement and interest… but the timing has been unfortunate with respect to IPO market conditions.” It has also been reported that it struggled to generate sufficient orders for shares at the price range it wanted. This may have an effect on the refinancing of Lowell’s 2nd lien notes

-          Regulation updates making it harder for debt purchasers. Examples include the “right to be forgotten” which would render part or most of the data that debt purchasers store and claim to use to increase collections from portfolios useless / not usable. This is due to be enforced in May 2018

-          Ratings agencies changing their way of assessing debt purchasers would lead to downgrades and higher rates of interest and difficulties for refinancing (although average maturity is 2024, with first debt due in 2022)

-          Adverse political developments such as Brexit and Cataluña

-          Any light being shone on the situation, i.e. press coverage, buy-side commentary

    sort by    

    Description

     

    A quickly deteriorating business model with flawed economics and suspicious accounting
    Prices here are based on 391p at time of writing

    Introduction of opportunity
    Debt purchasers (“DP”) essentially play a game of arbitrage between the rate of return they get on their portfolio investments and the interest rate they raise capital at. However, returns have been compressed, and interest rates only will move in one direction going forward, with further headwinds (regulation, potential accounting scrutiny, cash burn) that could cause the cost of their debt to increase at further issuances. The valuations of such companies only make sense if you believe that future NPLs can be purchased at more attractive prices, similar to what was achieved in the immediate years after the 08 crisis. Given today’s fundamentals, such companies are too levered, consistently burn through cash and are destroying value while facing an increasingly hostile regulatory and macroeconomic environment. The sector uses complex accounting, misleading metrics/ratios and numerous sector-specific acronyms which have misled regulators, accountants, rating agencies, sell-side research and investors. We believe Arrow Global Group plc (“Arrow” or the “Company”) has little to no equity value.

    Business model

    -          DPs aim to help debtors deal with their debt problems and repair their credit standing with long-term sustainable repayment plans. These are mostly NPLs, already defaulted loans with a mix of paying and non-paying accounts

    -          Banks and other credit institutions are less efficient at handling and collecting debts, it is not their core focus. As such, it makes sense for them to sell these to debt purchasers who are better at extracting value via relationships and data/analytics (accurately predicting future collections). Furthermore, such institutions are not keen to give haircuts as it would damage their core business model, whereas DPs factor this into account when acquiring portfolios

    -          Establish and maintain good relationships with a variety of debt sellers (banks, FIs, utilities, telco, retailers…) to maintain a pool of portfolio investment opportunities

    -          DPs have, to varying degrees also debt collection and asset management businesses (3rd party debt servicing)

    Income statement and accounting of debt purchasers explained

    Source: Arrow Group Global plc investor presentation.

    -          1. Debt purchasers price their portfolio investment/acquisitions based off a gross money multiple

    -          2.  The sum of the expected gross (usually, unless specified as net) collections from the portfolio gives the estimated remaining collection (“ERC”), this is usually measured either on a 7 years basis (84M) or 10 year basis (120M), with some residual claims beyond that period

    -          3. The collections generated from portfolio per period is then classified into 2 categories: a) income; and b) portfolio amortisation. There are also write-ups/revaluations which will also factor into revenue. Therefore, if the company deems more of its collections from a portfolio to be income rather than amortisation, it can achieve 2 things à a higher ERC going forward (since the balance of ERC is decreased by a lower amortisation number) and a higher revenue number (this also leads to higher net EPS and ROE)

    -          4. These are the costs associated with collecting any income from the portfolio. E.g. staff costs, legal costs, bailiffs etc. Benchmarking cost-to-collect ratios is not simple due to different categorisations of costs by competitors

     

    Source: Hoist Finance investor presentation.

    Why does the opportunity exist?
    Market view

    -          Arrow has repositioned itself as a largely UK debt purchaser with a single asset class to a leader across 6 countries with its own servicing capabilities covering half the addressable European NPL market consistently outperforming its financial targets

    -          The market views the company as an extremely high adj. EBITDA margin business (60%+) with low capex requirements (c.3% of sales) and limited working capital requirements

    -          The company is able to consolidate at attractive multiples (Zenith acquired for a headline price of 5.3x LTM EBITDA) as well as extracting meaningful synergies (higher rate of collections) from the use of their proprietary, sophisticated database that they have developed

    -          Data-driven analytics (a sophisticated complex process with reams of data-access required) company in a fast growing market (double-digit forecasts), as opposed to just a Business Process Outsourcing (“BPO”)-type business, with attractive fundamentals (market looks at P/E and ROE). Collection on portfolios have consistently exceeded management estimates, with the ongoing promise of quick deleveraging if proceeds are not reinvested (gross collections in next 3 years are c.60% of 84M ERC)

    -          The market is expected to continue to benefit from increasing deal flow of NPLs sold by banks in Europe due to increasing regulation (ECB’s Asset Quality Review, Basel 3 and IFRS9)

    -          Debt purchasers are able to benefit in good and bad economic periods (recession-resilient/counter-cyclical). In bad periods, the number of NPLs should increase and be available at lower prices (banks under pressure to sell) but default rates are higher and proportion of debtors paying are lower. In good economic times, debtors have higher payment capacity, NPL portfolios will experience lower defaults and increase in supply due to increased consumer lending

    -          The accuracy with which debt purchasers predict collections has been incredibly high and consistent (101-103% for Arrow, comparable and always above 100% for competitors), highlighting their data and analytical prowess

    -          There are significant barriers to entry such as relationships to debt sellers, databases for NPL bidding, regulatory compliance and economies of scale

    -          The ratings agencies have also shared this view, with the company being able to finance its debt at a rate of c.380bps (debt trades above par). 2 main leverage metrics used are net debt / adj. EBITDA and asset-backing (on gross ERC), which are both seen growing at attractive rates

    Investment thesis / Our view vs. consensus

    -          The company is more akin to a finite-life business. The company is an asset-heavy platform and not a going-concern unless more portfolios are added, otherwise it is in run-off mode. The investment required to “replace[1]" (similar to maintenance capex) what is collected is a lot higher than the market seems to think. Adjusting this, the fundamentals look totally different. The adj. EBITDA that the company shows (which the market and ratings agencies use) ignores the cost of acquiring the debt that is being collected, i.e. not recurring and not indicative of profitability

    -          The fundamentals reported have also been deteriorating when looking at the expected gross cash collections money multiples, which is a sign that the company is likely facing more competition (having to bid higher) for similar quality level portfolios. A consistent picture can be seen across the European industry (see appendix).

    -          Cost to collect ratios (which we believe is under-stated by the company by excluding staff costs, of which over 50% work in collections) have also been rising, leading lower net cash collections. The gross view on multiples does not take into account the time value of money (exacerbated by the high interest costs) and the gross collections number can be manipulated through accounting assumptions and/or through continuous M&A[2]. M&A expenditure used to acquire portfolios of NPLs should be treated the same way as costs to acquire portfolio (partially expansion and maintenance capex). M&A also helps to cloud organic performance & comparability

    -          Management is highly incentivised to grow EPS given compensation arrangements and has the tools to manipulate this and has been selling shares aggressively since their lock-up expired at the end of 2015

    -          The marketed performance of debt purchasers to forecast collections on portfolios acquired (bear in mind these are long since defaulted NPLs) implies significant earnings visibility, which is hard to believe. Most portfolios are impossible to stress test, let alone forecast. The only way we know to be profitable in this business is to buy at very low prices to provide for a large margin of safety

    -          Looking at the cash flow generation and profile since IPO (from January 1st 2014), it is clear that the company has essentially not generated any cash while growing its portfolio (gross ERC) by raising a substantial amount of debt at record low rates. Should there be any material movement in rates, lowering in profitability of operations (through a decrease in expected collections and/or higher cost-to-collection ratios), the company could quickly start to become a clear value-destructor (via need for additional debt raising). A bonus would be if ratings agencies were to reconsider their methodology of looking at leverage and asset-backing

    -          Sell-side reluctant to provide more scrutiny given fees provided by the sector to investment banks and focuses on EPS and ROE (both of which can be manipulated)

    -          Regulation may possibly also be a headwind, with the “right to be forgotten” being a key possible short-term development expected in May 2018 (the General Data Protection Regulation replacing the Data Protection Directive 95 / 46 / EC)

    Industry / market

    -          In the last decade, the industry has benefitted from Europe’s banks being weighed down by almost a trillion euros of NPLs and under pressure from regulators to get those debt off their balance sheets

    o    Debt sellers in the UK are selling the last of the backlog built up during the financial crisis, growth in recent years has come from vendors selling debt earlier in the cycle (fresher debt). This fact along with increased competition has led the price paid by Arrow per £ face value of debt to increase from 7p in 2013 to 13p in 2017H1

    -          The ongoing consolidation of the debt purchasers market and their increasing influence (dictating prices on purchases of NPLs) has not gone unnoticed by regulators, with the Intrum/Lindhoff merger being forced to make a series of divestments in the Nordic region (both debt collection and purchasing activities) as a condition to close, which cuts the proposed cost savings by nearly a third

    o    Furthermore, it has been reported that the EU is considering action to try to boost demand and increase sale prices for secondary market NPLs

    o    25th May 2017 ECB started encouraging public-private partnerships to improve data quality and recovery processes to reduce asymmetries between buyers and sellers

    -          Prices for financial institutions’ receivables (portfolios) have been increasing, and this has historically been the primary activity of debt purchasers. This is shown by the expected collection multiples falling

    o    Pricing pressure, leading to lower gross multiples (before collection costs)
    Intrum said in April it saw pricing pressure from increased competition in W. Europe

    o    Increasing cost-to-collect ratios reported, leading to lower net multiples (post-collection costs), which may be due to lower quality portfolios (higher cost to collect per value of debt). The ratio has increased from 21.9% in 2013 to 34.2% in 17Q3 LTM (this includes some litigation costs, but excludes staff costs)

    o    Higher prices being paid may also reflect the fresher debt with higher collectability being sold, as per Hoist’s experience in the last two years. However, bizarrely, Arrow has experienced higher cost-to-collect ratios as well as diminishing expected collection multiples

    o    More types of credit being bought (such as telco, utility and retail receivables) as well as an increased focus on growing the “asset management” side, whereby Arrow services third-party portfolios. This business has increased for Arrow from 1.5% of sales at IPO to 22.3% 17Q3 LTM (£1.4m à £63.3m)
    Market assumes this is very high margin, recurring business. However, research has shown that this is a low margin (c.10%), commoditised BPO-type business

    -          The European market has large regional differences in market maturity and regulatory landscape (debt collection process, access to credit data…). Deploying large amounts of capital in other countries is not always easy due to a combination of lack of deal flow, high prices and an inability to price competitively

    o    Italy is cited as the largest market for NPLs with low prices, however, this may be due to the fact that the quality of the portfolios are low. E.g. UniCredit estimated the price of recent deals would only allow it to recoup 23c on its bad loans, far below the 43c average for Italian banks à it went on to sell these at 13c

    o    Cabot ended a JV in Spain after 1 year, Aktiv Kapital withdrew from numerous countries and EoS struggled to establish a leading position in Austria

    -          Consolidation possibilities exist, with headline multiples looking highly accretive. Increasing regulation makes it harder for smaller competitors to meet all requirements while maintaining an acceptable level of profitability

    o    1) headline / adjusted EBITDA is not a sustainable metric, as mentioned above and; 2) value creation only exists if the buyer of the portfolio is able to extract a higher collections from the portfolio

    -          Market relies generally on consumer borrowing, which is highly sensitive to economic conditions

    -          The debt purchasers have taken full advantage of record low interest rates to fund the majority of their portfolio purchases via debt and also helped them through the crisis. This situation is expected to reverse (nobody knows when, but we know that rates cannot go below where they are)

    -          Arrow boasts about its database and analytics capabilities, however, all other competitors seem to have similar processes, expertise and infrastructure. This is somewhat confirmed by the similar rates of collection multiples that have been achieved (cost-to-collections have generally been rising, but hard to compare due to categorisation)

    -          UK-centric debt purchasers will be hit by National Living Wage (increasing collection costs)

    Valuation and analysis of deteriorating fundamentals

    -          Yield compression: Arrow boasts about its data and analytics (consistently predicting actual collections on long-dated, long since defaulted portfolios to the tune of +/-3% historically), the development of sellers now selling fresher, more-paying debt is a negative development, since these will be with lower expected collection multiples (due to higher price and competition). This is a general trend observed also with competitors[3] (pre-2011 gross cash-on-cash multiples at 2.4-4.0x and the last 2 years at 1.7-2.0x). The unit economics are not profitable, nowhere close to what the market thinks/is pricing. Assuming 1.7-1.9x gross cash-on-cash multiples and then removing collection costs (which are also rising) of between 25-35%[4], you get a net cash-on-cash multiple of 1.11x-1.43x, this is before taking into account other real costs: overhead costs, interest costs (since most portfolio investments are financed by debt), taxes and the time value of money, since collections are over 7-10+ years

    -          Adj. EBITDA a measure of profitability?: Arrow reports adj. EBITDA which is not a proxy for cash flow or profitability, as it ignores a key component of expense/investment (opex / maintenance capex) for the business to maintain its status as a going concern, the investment required to keep the portfolio value (ERC) flat, the replacement rate. Adjusting for this, the 17Q3 LTM / FY16 / FY15 EBITDA becomes 42.7 or 86.2[5] / 86.8 / 72.4 instead of 206.2 / 209.2 / 153.1. This highlights the problem that the company has been facing, turning it into a value destroying business (better off for investor that the company runs off the portfolio, as opposed to continuing to invest in it). The going-concern valuation analysis shows that the valuation not even enough to cover the debt, let alone having any residual value for equity holders

    -          A sum-of-the-parts analysis: also supports such a hypothesis, taking the latest reported gross ERC profile with a cost-to-collect ratio of 28% and discount rate of 10%, the value of the current portfolio is £944.0m. The value of the asset management services division is estimated at £66.0m (by taking an EBITDA margin of 10.0% and assigning a multiple of 10.0x) à enterprise value = £1,010.0m. The latest reported net debt is £926.2m giving an equity value of £83.8m vs. the current market capitalisation of £698.7m (an implied downside of 88.0%)

    -          Under-amortisation of ERC and inflation of revenue: Collections from portfolios are split into 2 categorisations, portfolio amortisation (which reduces the balance sheet value of portfolio investments) and income. The amount of amortisation implies the expected money multiple from a portfolio. E.g. a 50% portfolio amortisation implies a 2.0x expected money multiple[6]. The amount that the company amortises from the collections is entirely at the discretion of the company and it is impossible from the outside-in to assess this due to the large number of different portfolio vintages and limited data disclosure. Arrow expected c.1.8-2.0x gross money multiple on ERC, however, Arrow’s average portfolio amortisation rate has averaged 30.5% since 2013, implying an expected multiple of 3.3x. This means that ERC (used for asset-backing leverage) and revenues have both been overstated over the last 7.75 years (large cumulative effect of over £200m estimated at over 20% of net 84M ERC). Furthermore, by looking at the collections divided by the gross 84M ERC we get

    -          Use of cash: since 1st January 2013 (4.75 years), we can see that the business has grown its portfolio (ERC) via large debt issuances and has generated practically no cash due to the combination of a) ongoing investments required to keep the portfolio replaced and; b) investments to acquire and enlarge the portfolio. However, as we have seen above, today’s portfolio is of a much lower quality and likely value destructive, therefore, the return on investments are likely extremely low, or even negative

    Cash starting balance

    Cash from operations[7]

    Cash from investments

    Cash from financing

    Cash ending balance

    £9.6m

    -£140.9m

    -£248.9m

    +£414.6m

    £36.2m

                            As per Arrow Global Group plc cash flow statement, from 1st January 2013 until 30th September 2017.

     
    Furthermore, from FY10 until today (7.75 year span), operating cash flow has been negative in every period, with financing cash flows (predominantly debt raising) being the driver for cash

    -          Changes in data disclosure: Arrow stopped reporting portfolio write-ups since 2015 and old vintage ERC splits. You would expect that the ERC for a vintage to decrease going forward (i.e. you collect something from these portfolios, so even with a slight mark-up, the net should be negative), however, there was a large mark-up in the ERC for 2011 and earlier vintages of £26.6m in 2015, and a slight mark-up of £0.4m for the 2010 vintage ERC from 2013 to 2014 but this is not shown since the company doesn’t disclose those vintages. For example the company does not show pre-2012 vintage ERC splits (stopped in 2015)

    -          When looking at credit metrics, the company reports net debt to LTM adj. cash EBITDA of 4.5x and net debt to gross 84M ERC of 63.6%, however, if we adjust for replacement rate of EBITDA and 84M ERC net of collection costs, the ratios become 10.7x (generously using the company’s quoted replacement rate of £120m, instead of the worked out number of c.£164m) and 96.6% (implying little to no equity value, and comfortably exceeding/breaching the covenant set at 75.0%)

    -          An aggressively growing portfolio, requiring an increasing amount of maintenance capex to maintain just the replacement rate, with capital being deployed at increasingly lower, or even negative returns

    Risks

    -          The collection of claims is dependent on general economic conditions, if the benign economic environment continues, the show can keep going on

    -          Lack of concrete short-term catalysts

    -          Debt purchasers are able to front-load collections to boost near-term earnings (to the extent allowed, and not already done), cash position (by negotiating a lump-sum amount that comes earlier with debtors) and also write-up the portfolio value (manipulating portfolio amortisations/inflating ERC) to keep metrics such as EPS growth and return on equity looking good for the market

    -          Given the above, if the company can continue to use its adj. EBITDA metric for leverage purposes and raise debt to enlarge its portfolio. In this scenario, the only catalyst would be when the company is unable to pay interest and other operating costs

    -          M&A activity:

    o    The company continues to expand its markets and portfolio via M&A, at seemingly accretive headline multiples and the market assigns value to this

    o    Given the amount of consolidation in the sector, it is not unthinkable for the company to be a takeover target for larger competitors (such as Intrum) and/or private equity sponsors (the sector has historically been a favourite for private equity, with reputable firms participating)

    Views on valuation

    The "real" EBITDA of the company (taking into account replacement rate, using a aggressive £120m) is only 41.8% of the adj. EBITDA that the company widely reports. The "real" leverage of the company is 10.7x LTM vs. the adj. 4.5x, with the asset backing (defined as net debt / net 84M ERC) at 96.6% as opposed to the reported 63.6% (which uses the gross expected remaining collections number). This leaves little room for value in the equity. We have not made an attempt to value any further M&A and portfolio purchases, which one could argue could boost the value of the Company slightly (although we have argued that given the current market dynamics and with interest rates only likely to move in one direction, the expected returns should be very low if not bordering on being value destructive). If we further add in to the fact that the ERC is over-valued, along with EBITDA (i.e. revenues recognised are higher than they should be due to the under-amortisation), then there leaves little value to the equity.

    Company trades today at 7.9x LTM Adj. EBITDA (without taking into account the replacement rate) and 18.8x LTM "real" EBITDA. We feel that even a stretch valuation for such a Company (commoditised, high competition industry with deteriorating fundamentals) may be 12.0x "real" EBITDA. As such, one could argue that the debt is covered, leaving minimal equity value (1.3x EBITDA of equity).

     

    Appendix

     

    Summary of findings on KPIs and financing:

    Note: Replacement rate used for Arrow is £120m guidance given by management on the latest earnings call, however, it is unclear whether this is a 9 month or full year number. The worked out replacement rate is £163.6m.

    Expected and achieved gross money multiples by vintage (year portfolio was acquired)