Download Difference in Book Value & Market Price of Bad Loans in Italian Banks and more Summaries Accounting in PDF only on Docsity!
Notes on
Financial
Stability and
Supervision
No. 3
April 2016
Contents Introduction and main conclusions ................ 2 1 NPLs: basic concepts and accounting rules ................. 3 2 NPLs in the Italian banking system: the current situation ... 4 3 Main reasons for the difference between book value and market price of bad loans ................................ 5 4 Effect of recovery times on the price and stock of bad loans ................................ 8 5 Conclusions ................................ 9
What’s the value
of NPLs?*
L. G. Ciavoliello – F. Ciocchetta – F. M. Conti
I. Guida – A. Rendina – G. Santini*
Summary
The main reasons for the difference between the book value of bad loans and the price that specialized market operators are willing to pay lie in the different criteria used in the financial statement and by investors to calculate their value. This paper demonstrates that these different criteria can account for the entire gap between the book value of bad loans and the price offered by an investor and that the gap is proportionate to the length of recovery procedures (judicial and extra-judicial). In other words, recovery times play a key role in the valuation of these assets.
The papers published in the
Notes on Financial Stability
and Supervision series
reflect the views of the authors
and do not involve
the responsibility
of the Bank of Italy
- Directorate General for Financial Supervision and Regulation. We thank Paolo Angelini for reading several draft versions of the paper, always making useful suggestions. Our thanks also go to Alessio De Vincenzo, Giorgio Gobbi, Antonio Renzi and Enzo Serata.
ISSN 2284-2853 (online)
Introduction and main conclusions
One of the reasons for the failure to develop a secondary market for non-performing loans (NPLs) in Italy is that there continues to be a substantial difference between the book value of these assets and the prices offered by investors. This paper investigates the main drivers of this difference, of which there appear to be two:
- Investors in NPLs demand a very high rate of return, partly because they generally have less financial leverage than banks. This return is used to discount the expected cash flows from NPLs (banks adopting the IAS/IFRS international accounting principles instead use the original effective interest rate on the assets, which is usually much lower) and results in a lower NPL price.
- Banks, as required by international accounting principles, include the indirect costs of managing NPLs in their financial statement of the year in which they are incurred, whereas potential acquirers deduct them immediately from the value, thus reducing the purchase price.
This paper shows that these two factors alone can account for the entire difference between the book value of bad loans and the price an investor is willing to pay, and that this difference is proportionate to the length of the recovery procedure (judicial or extra-judicial). In other words, recovery times play a key role in the valuation of these assets.
These results have major implications for solving the problem of the sizeable stock of NPLs. First, a shortening of recovery times would almost immediately increase the value of NPLs, with positive consequences for banks’ ability to allocate sufficient resources to financing the economy and for financial stability. Simulations conducted by the Bank of Italy show that a two-year reduction in recovery times would entail a market price increase of approximately 10 percentage points and, other things being equal, a significant reduction in long-term stocks of NPLs.
As was recently highlighted by the senior management of the ECB and the Bank of Italy, the disposal of NPLs will take place gradually. The supervisory authorities assess the situation of each bank – the effectiveness of internal procedures for the management and recovery of NPLs, the coverage ratios, the ratio of NPLs to total loans – in order to identify the most appropriate supervisory measures, also taking account of the external context in which the banks operate. They do not push all banks, indiscriminately, to sell these assets on the market at the earliest opportunity.
Banks need to adopt more efficient internal procedures for NPL management, carefully considering the advantages of outsourcing to specialized operators and scheduling disposals as part of their business plans.
is given by the sum of value adjustments (and possible recoveries) recorded over time. The coverage ratio is given by the ratio between the amount of write-downs and the gross amount of impaired positions. 4
2 NPLs in the Italian banking system: the current situation
The large stock of NPLs of Italian banks is mainly due to the exceptional recession that has buffeted the Italian economy in recent years, as well as to long credit recovery times. The very limited development of a secondary NPL market has also contributed.^5
Non-performing loans, gross of provisions, amounted to €360 billion in December 2015, that is 18.1 per cent of total loans to customers. The amount of bad loans (the ‘worst’ category of NPLs) 6 was €210 billion (10.6 per cent of total loans). For balance sheet purposes, the amounts net of provisions have to be considered: €197 billion and €87 billion respectively (Table 1).
Table 1 – NPLs: amounts, coverage ratios and guarantees (billions of euros and percentage points; December 2015)
Gross exposure Provisions Net exposure Coverage ratio Collateral(1) guarantees (1)Personal
Total NPLs 360 163 197 45.4% 160 52 of which: bad loans^210 123 87 58.7%^85 Source: Supervisory reports, on a consolidated basis for banking groups, individual for the rest of the system. (1) Amounts on an individual basis. The amount of guaranteed credits is shown, not the amount of the guarantees: where the amount of the guarantee exceeds that of the credit, the largest amount shown is the amount of the credit itself.
Coverage ratios have increased progressively since 2012, partly thanks to targeted supervisory action by the Bank of Italy.^7 The average NPL coverage ratio is 45 per cent (59 per cent for bad loans only), in line with the European average 8 (37 per cent and 55 per cent respectively in June 2012).
Collateral for non-performing loans amounts to €160 billion. This figure does not necessarily correspond to the collateral’s fair value, but to the amount of credit backed
4 In the notes to financial statements value adjustments can be represented in two ways, both indicating the same loan value in the balance sheet, that is, net of adjustments. The first is to write down part of the exposure no longer recoverable; the second, if there is no longer any reasonable expectation of recovery, is to write off the expected loss, subsequently reducing the original gross value of the loan. Write-offs must also be considered when calculating coverage ratios, as otherwise the indicator would be underestimated. For details see the box ‘Coverage ratios and write-offs’, Financial Stability Report , No. 4 (2012). 5 Sales of bad loans were only for small amounts in 2012-14 (about €11 billion, corresponding, on a yearly average basis, to 2 per cent of the average stock). The amount increased in 2015, but was still small (about €9 billion). 6 NPLs are divided into categories according to their recoverability. If the debtor’s difficulties are expected to be only transitory, banks can opt to make lower provisions. 7 See ‘The recent asset quality review on non-performing loans conducted by the Bank of Italy: Main features and results’, Bank of Italy, 2013. 8 Unlike other countries, even in Europe, the share of foreclosed assets (properties seized by banks due to debtor default) is almost negligible in Italy. This has to be taken into account for a true international comparison. Foreclosed assets are not in fact NPLs from a technical point of view, but from an economic one they do represent a risk for banks, comparable to that of NPLs secured by real estate collateral. In both cases banks are exposed to real estate market trends.
The large
stock of NPLs
is due to the
recession and
long credit
recovery times
The average
share
of secured
lending is high
by collateral.^9 The average ratio of credits backed by collateral or personal guarantees in total non-performing loans is 67 per cent. Considering only residential mortgages to households, collateralized credits (nearly all with real estate) amount to 94 per cent.
3 Main reasons for the difference between book value and market price of bad loans
The NBV of bad loans is significantly higher than the price that investors in this market (generally international hedge funds) are willing to pay. The figures on NBV are conclusive: currently, system-wide average NBV is 41 per cent of GBV (the difference between the coverage ratio shown in Table 1 and 100). There are no figures indicative of market prices because the market is very thin and the loans sold vary widely as to type, guarantees and amount of the write-down. There have been cases in which the sale value has topped 45 per cent of GBV, where the loan was secured by high value guarantees (such as prime residential property), and others in which it has barely reached 3 per cent (unsecured positions). For the time being, therefore, it is impossible to quote an average value that is representative of market prices. To give an example, in the case of the bad loans of the four banks put into resolution last November (GBV of €8.5 billion), independent experts’ latest estimates set the disposal value at 22.3 per cent of GBV.
Although investors do not have access to valuation methods, it is possible to analyse some of the factors that can generate gaps between market prices and banks’ book values that are similar to those observed.
Let us take an exposure classified as a bad loan with a GBV of €100, partly secured by a real guarantee. We assume that the bank’s estimate of expected cash flows is the same as that of investors in the market, 10 with only one inflow whose expected value is 47 per cent of the gross value of the loan (already net of the direct costs of selling the guarantee), which will be collected in full at the end of the recovery procedure. 11 We also assume that the expected residual recovery time is four years. This is consistent with the result of the survey of recovery times for credit to firms that the Bank of Italy carried out in 2015. 12
Let us look first at how these assumptions translate into the valuation of a position from the bank’s point of view. To do so, a further assumption must be made regarding the original effective interest rate on the loan which the banks, in accordance with
9 For example, if a credit is backed by a guarantee with a higher fair value than the credit itself, the amount reported is that of the credit. Furthermore, guarantee values are based on individual supervisory reports of Italian banks, whereas NPL values are based on consolidated reports (which include NPLs relating to foreign intermediaries and to Italian financial institutions belonging to a banking group). 10 This assumption may not be true and the low price offered by market operators may be the result of a lower estimate of future cash flows than that of the banks. In other words, the coverage ratio may be too low. In such a situation, the assumption merely pinpoints the causes of the ‘spread’ that are not linked to the coverage ratio. 11 The example does not take into consideration partial reimbursements over time. 12 Luisa Carpinelli, Giuseppe Cascarino, Silvia Giacomelli and Valerio Vacca, ‘The management of non-performing loans: a survey among the main Italian banks’, Questioni di Economia e Finanza (Occasional Papers), No. 311, February 2016. According to this study, almost 80 per cent of loans involved in liquidations have been the object of recovery procedures for less than 5 years and the average duration to date of the liquidation proceedings, weighted for the loan amount under a number of simplifying assumptions, was 3.5 years in 2014. In bankruptcies, the average duration was 3.8 years, against 2. years for compositions with creditors and 3.3 years for foreclosures.
The book
value of bad
loans is much
higher than
the price
investors are
willing to pay
… owing to
the different
valuation
methods
Banks follow
IAS 39 rules …
also takes account of performance fees levied by fund managers, which can be as high as 20 per cent of net profits. Finally, there may be genuine differences in the valuation of future cash flows due to information asymmetries in the credit market.
The following two factors play a key role in determining prices.
(i) Indirect management cost effect
To our knowledge there are no reliable publicly available statistics on the indirect costs of managing bad loans. Anecdotal evidence suggests they may account for as much as 6 per cent of nominal expected cash flows.
Table 2, column 2, shows the valuation of a bank which includes these costs, contravening accounting principles, and does not change the other factors listed in column 1. In this case, the present value of the bad loan is 37 per cent of GBV, some 3 percentage points below the example in column 1. Provisioning should therefore be increased by the same amount.
(ii) Rate of return effect
Based on available evidence, again anecdotal, the simulation assumes that investors’ IRR to acquire bad loans is between 15 and 25 per cent.
The results, which appear in columns 3 and 4 of Table 2, show that the effect on the valuation of bad loans is substantial, ranging from 13 to 21 percentage points of GBV according to the IRR.^13
(iii) Overall effect
As a whole, taking both factors into account (indirect costs and IRR), the different approach followed by investors would warrant a price gap with respect to NBV of between 16 and over 24 percentage points of GBV. This represents a purchase price for the investor of between 24.1 and 16.4 per cent of GBV.
These tentative findings suggest that the main reasons for the gap in the market price of bad loans relate to the different valuation criteria used in the financial statement and by investors to compute the value, rather than to insufficient coverage ratios.
4 Effect of recovery times on the price and stock of bad loans
The valuation of a bad loan position can differ greatly based on the recovery time, both in terms of accounting value and, especially, of market value. Various factors affect recovery times, including the efficacy of a bank’s internal procedures and the efficiency of a country’s legal and judicial system. As a result of the latter, the valuation can
13 In the interval considered (15-25 per cent), the effect of the IRR on the price offered is approximately linear. For example, with an IRR of 20 per cent, the valuation is 19.8 per cent of the GBV, compared with the 16.4 and 24.1 per cent reported in Table 2 for IRRs of 25 and 15 per cent respectively.
… and
discount
future cash
flows with
the expected
effective rate
of return
The overall
effect ranges
from 16 to 24
percentage
points of GBV
Recovery
times affect
the valuation
of bad loans …
change significantly even within the same country as different courts process recovery procedures at different speeds.
Table 3 contains a sensitivity analysis of the value of the bad loan with respect to the cash flow recovery time. It shows the price that investors would be willing to pay to buy the bad loan as a function of various recovery times, assuming an IRR of 20 per cent. Shortening the recovery time by even a year, from 4 to 3 years, increases the price by 4.6 per cent of GBV.
Tav. 3 - Sensitivity of bad loan prices to recovery times (1) (per cent of GBV) Recovery time (years) Price* 1 36. 2 29. 3 24. 4 19. 5 16. 6 12. (1) Assuming 20 per cent IRR.
Not only do recovery times affect the valuation of bad loans and, more generally, NPLs, but balance sheet values as well. The longer the recovery time, the higher the ratio of bad loans to total loans. Recent estimates show that two banking systems with a loan growth rate of 5 per cent and a rate of new bad loans of 2 per cent but with different bad loan recovery times (2 years and 5 years), in equilibrium would have respective bad loan to total loan ratios of 3.5 per cent and 7.4 per cent. 14
5 Conclusions
Recently, the press and specialized operators have spread the rumour that the Single Supervisory Mechanism (SSM) intends to force banks to rapidly offload NPLs on the market. This perception may be one of the causes of the recent sharp fall in bank stocks in the euro area and Italy. The idea that the SSM intends to force banks to indiscriminately and rapidly offload their NPLs is incorrect, as underlined on various occasions by senior members of the ECB, the SSM and the Bank of Italy. 15 With regard to NPLs, the Supervisory Authority carefully evaluates each case, keeping in mind the numerous internal variables of each bank (such as the efficacy of recovery procedures, the adequacy of the coverage ratio, and the share of NPLs in total loans) as well as the external context within which the bank operates.
14 See the box ‘The relationship between length of credit recovery procedures and volume of bad debts on banks’ balance sheets’, Financial Stability Report , No. 5, 2013. 15 M. Draghi, ‘Introductory statement to the press conference’ – Governing Council decisions, 21 January 2016; D. Nouy, ‘Introductory statement at the Presentation of the ECB Annual Report on supervisory activities 2015’, ECON Committee of the European Parliament, 22 March 2016; I. Visco, ‘Fact-finding inquiry on the Italian banking and financial system and the protection of savings, also regarding supervision, crisis resolution and European deposit insurance’, 19 April 2016.
… and the
amount of
bad loans in
equilibrium
We must
accept that the
reduction of
the high stock
of NPLs will be
gradual …
