Retail loans in the Polish Banking System Luigi Ruggerone - - PowerPoint PPT Presentation

1

Luigi Ruggerone – Intesasanpaolo Risk Management Dept.

Retail loans in the Polish Banking System

2

• 3. Executive Summary
• 4. House price evolution in Poland
• 5. Main risk drivers in FX denominated housing loans
• 6. Impact of Zloty depreciation: assumptions
• 7. Analysis of the impact of Zloty depreciation on the Polish

banking system

• 8. Estimated functional relationship between FX and NPLR
• 9. Estimated impact of a depreciation scenario on NPLR

10.Possible impact of Zloty depreciation on Ec. Capital 11.Summary of the results of the functional relationship

• 12. Impact of macroeconomic scenario

13.Vintage analysis on systemic figures

• 14. Macroeconomic stress test
• 15. Stress test on different vintages (1/2)
• 16. Stress test on different vintages (2/2)
• 17. Impact of unemployment on impaired rate
• 18. Conclusions

Summary

3

Executive Summary

 The main risk driver of the polish banking system is the retail credit portfolio, in particular the mortgage loans portfolio denominated in CHF;  Despite the freshness of the portfolio and the limited time history, our analysis highlights a non linear relationship between the FX and the NPLR and allows us to draw some conclusions on additional provisioning that may become necessary in case of depreciation of the local currency. A 30% depreciation of the PLN determines an increase of more than 6% in the non performing loan rate;  We also estimate the impact of an alternative scenario characterized, over a longer time span (until the end of 2014), by an appreciation of the PLN (15%) and higher CHF rates (to the level reached in 2007) and find out that the housing loans extended in 2007 until Q2 2008 are the riskiest. The reason being a residual capital well above the original amount extended and, above all, a sharp increase in the installments;  We also run, with the help of Moody’s Credit Cycle model, an exercise to estimate the impact of unemployment on retail credit delinquency. A 350 bps increase in unemployment over the next 2 years, determines, on average, a 100 bps increase in the delinquency rate;

4

House price evolution in Poland

 We report below the house price evolution according to AMRON-SARFiN Report and we concentrate on the major cities;

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 1Q 2006 2Q 2006 3Q 2006 4Q 2006 1Q 2007 2Q 2007 3Q 2007 4Q 2007 1Q 2008 2Q 2008 3Q 2008 4Q 2008 1Q 2009 2Q 2009 3Q 2009 4Q 2009 1Q 2010 2Q 2010 3Q 2010 4Q 2010 1Q 2011

House price evolution

price per square meter

Warsaw Katowicw Wroclaw Gdansk Krakow Poznan

 From 2007 on, the value of the houses does not change significantly, so it is reasonable to believe that the re-assessment of house values impacts mainly on 2006 vintages;  For vintages from 2Q 2007 up to 2Q 2008 the re-assessment could negatively impact on the LTV level;

5

Main risk drivers in FX denominated housing loans

• Economic downturn, especially in terms of unemployment rate increase, can trigger an increase in

the non performing loan ratio; this is also confirmed in the Financial Stability report, where it is stated that “The deterioration in the condition of some household related, inter alia, to unemployment growth during the recent economic slowdown, may continue to exert a negative effect on loan quality”.

• The majority of the loans has been granted to customers during a credit boom period, with low

interest rates, that probably determined excessively benign assessment of the creditworthiness of clients;

• The impact of the rules set by the National Bank, within Recommendation S on good practices with

regard to management of credit exposures that finance property and are mortgage-secured, could be relevant:

• On the one hand it is stated that the repayment burden on retail borrower’s net income should

not exceed 42%

• On the other hand, starting from June 2012, the risk weight to be applied for foreign currency

denominated retail exposure will increase to 100%;

• This last change, considering the standard methodology for calculating Capital Requirement,

will determine a consistent increase of capital requirement for FX loans, and a consequent reduction of the solvency ratio; considering only the FX denominated portfolio, capital absorption will consequently increase by 33%.

6

Impact of Zloty depreciation: assumptions

• The current and the forecasted value of Switzerland inflation rate is reported below:
• Our analysis is focused on the impact of zloty depreciation vs Swiss franc as the main risk factor;
• The short time series of exposures and Non Performing Loans require a very cautious approach;
• Even if an increase of CHF interest rates can impact banking system portfolios, we do not foresee this

as the major risk over the short term, considering the statement of the Swiss National Bank, on September the 6 th, and the current and forecasted value of inflation in Switzerland: Reuters: “Using some of the strongest language from a central bank in the modern era, the SNB said it would no longer tolerate an exchange rate CHF/€ below 1.2 and would defend the target by buying other currencies in unlimited quantities”.

• Any interest rate increase, in the current economic environment, will affect the commitment of

maintaining the FX at the current target level.

7

Analysis of the impact of Zloty depreciation on the Polish banking system

 This analysis assesses the impact of zloty depreciation vs Swiss franc on the Non Performing Loan Rate;  Due to the lack of data, but above all due to the conversion into PLN of problematic housing loans, we both analyze CHF denominated housing loans and total housing loans (independently of the currency);  Based on the available data, we estimate for the Polish banking system a functional relationship between the level of the CHF/PLN exchange rate and the level of the Non-Performing Loan Ratio (NPLR from now on);  The two estimated relations between NPLR and FX rate are non linear and monotonic; i.e. larger depreciations have increasing effects on NPLR;  In our analysis we use monthly average FX rate, calculated starting from the figures provided by the Polish National Bank (NBP) on a one year time horizon;  We assume that the effect of a change in the FX rate will impact on NPLR with a delay of three months; so the relation between NPLR and FX has been estimated considering a time lag of 3 months;  Furthermore we estimate the expected loss under three different FX stressed scenarios. In particular three different thresholds of depreciation of PLN vs. CHF have been applied: 10% - 20%

• 30%.

8

Estimated functional relationship between FX and NPLR

 As already anticipated the functional relationship is monotonic and non linear. The NPLR grows faster than the PLN depreciates vs. CHF;  Since the dynamics of total NPLR and NPLR of the CHF denominated housing loans are very similar over time, the regression function has a similar trend; this confirms the hypothesis that the housing loans denominated in CHF after the restructuring fall within Total Housing Loan Portfolio;  As already stated in the previous page, the issue of restructured housing loans should be carefully analyzed.

CHF NPLR and CHF/PLN rate

0.80% 1.00% 1.20% 1.40% 1.60% 1.80% 2.00% 2.5 2.6 2.7 2.8 2.9 3 3.1 Fx NPL vs CHF NPLR

Total NPLR and CHF/PLN rate

1.20% 1.40% 1.60% 1.80% 2.00% 2.20% 2.40% 2.5 2.6 2.7 2.8 2.9 3 3.1 Fx NPL vs CHF NPLR

9

Estimated impact of a depreciation scenario on NPLR

 Using the non linear regression function presented above, we shocked the monthly average FX rate calculated on a one year time horizon by 10%, 20%, and 30% respectively. We provide below the results of this stress test:  The impact, in term of additional provisioning, will be determined through the Cover ratio prevailing within each Polish bank;  We have estimated also the impact on NPLR of interest rate level; however due to the short time series available, currently interest rates have a very limited explanatory power, so the only variable considered here has been the FX rate level;

Systemic NPLR June 11 Strees on FX Stressed FX* Tot NPLR

Δ NPLR

2.07% 10% 3.58 4.22% 2.15% 2.07% 20% 3.90 6.08% 4.01% 2.07% 30% 4.23 8.41% 6.34%

* The fx means in Apr 11 and Oct 11 were 3.009 and 3.25 respectively

Systemic NPLR June 11 Strees on FX Stressed FX* CHF NPLR Δ CHF NPLR 1.53% 10% 3.58 3.51% 1.98% 1.53% 20% 3.90 5.20% 3.66% 1.53% 30% 4.23 7.29% 5.76%

* The fx means in Apr 11 and Oct 11 were 3.009 and 3.25 respectively

Total housing loans CHF housing loans

10

Possible impact of Zloty depreciation on Ec. Capital

 Apart from an estimation of the level on NPLRC, our estimated function allows to calculate the economic capital (i.e. the unexpected loss) with different confidence intervals;  This analysis could be very useful in case of a comprehensive ICAAP exercise. The economic capital is the difference between the unexpected loss (red or green lines in the graph above, depending upon the confidence interval) and the expected loss (black estimated function) for each given FX rate;

Impact of FX level on Ec. Cap.

0.40% 0.90% 1.40% 1.90% 2.40% 2.70 2.80 2.90 3.00 3.10 atual nplrc values Estimated function 95% 99%

NPLRC CHF/PLN

11

Summary of the results of the functional relationship

 The appreciation of CHF from 2009 up to now, determined a consistent increase of the residual capital to be repaid, and consequently the LTV climbed up; this increase affects considerably also the level of installment, once CHF IR will increase again;  The increase of installments has, so far, been limited by the decrease of CHF interest rates; anyway current installments are, on average, higher than original ones;  Applying a consensus scenario (appreciation of the polish zloty and increase of CHF interest rates) installments will, on average, increase, while LTV will decrease considerably;

12

Impact of macroeconomic scenario

 In the following slides, with the help of Moody’s Credit Cycle model, we report the results of an exercise aimed at estimating the impact of unemployment on retail credit delinquency.  We assessed the impact of a 350 bps increase in unemployment over the next 2 years, for estimating the changes in the delinquency rate;  This kind of exercise is somehow suggested also by the Polish regulator itself and by the empirical

• bservation of the impact on repayment capacity from macroeconomic dynamics;

13

Vintage analysis on systemic figures

 According to systemic figures, we firstly analyzed the 90+ day delinquency rate for retail mortgages extended from 2005 up to 2008;  It should be noticed that this analysis refers to the whole mortgage portfolio, as data series do not contain breakdown by currency of origination; Mortgage Market, 90+ day

Annual Vintages 2005-2008

0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Quarters since origination

90+ day, % of outstanding

2005 Vintage 2006 Vintage 2007 Vintage 2008 Vintage

 The graph shows that younger vintages are the riskiest. This is confirmed also by the Polish Financial stability report.

14

Macroeconomic stress test

 The following analysis is based on the assumption that an increase of the unemployment rate

• ccurs jointly with a global scenario, similar to that in 2009 and 2010, with a set of pessimistic

hypothesis for some core European countries;  The stress scenario, implying an unemployment that hovers around 12% in the next 5 years, is reported below:  The maximum increase of unemployment, with respect to the current level, is in the tune of 3,5% throughout 2014;

Unemployment Rate

Stress Scenario

4 6 8 10 12 14 16 18 20 22 Jan-00 Sep-00 May-01 Jan-02 Sep-02 May-03 Jan-04 Sep-04 May-05 Jan-06 Sep-06 May-07 Jan-08 Sep-08 May-09 Jan-10 Sep-10 May-11 Jan-12 Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 Unemployment Rate, % stress historical

15

Stress analysis on different vintages 1/2

 The stress test per vintage confirms the results of the previous analysis, where the latest vintages appear as the most vulnerable;

Mtg 90+ day Delinquency Rate

Vintages 2005 2006

0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 2005_Q2 2005_Q4 2006_Q2 2006_Q4 2007_Q2 2007_Q4 2008_Q2 2008_Q4 2009_Q2 2009_Q4 2010_Q2 2010_Q4 2011_Q2 2011_Q4 2012_Q2 2012_Q4 2013_Q2 2013_Q4 2014_Q2 2014_Q4

2005 2005 stress 2006 2006 stress

16

Stress analysis on different vintages 2/2

 At a banking system level 2008 vintage is the riskiest.

Mtg 90+ day Delinquency Rate

Vintages 2007 2008

0.00% 1.00% 2.00% 3.00% 4.00% 2007_Q2 2007_Q4 2008_Q2 2008_Q4 2009_Q2 2009_Q4 2010_Q2 2010_Q4 2011_Q2 2011_Q4 2012_Q2 2012_Q4 2013_Q2 2013_Q4 2014_Q2 2014_Q4

2007 2007 stress 2008 2008 stress

17

Impact of unemployment on Impaired rates

Mortgage Impaired Rates and Unemployment

0.00% 1.00% 2.00% 3.00% 4.00% 5.00%

Mar_2009 Sep_2009 Mar_2010 Sep_2010 Mar_2011 Sep_2011 Mar_2012 Sep_2012 Mar_2013 Sep_2013 Mar_2014 Sep_2014 Mar_2015 Sep_2015 Mar_2016 Sep_2016 Impaired Rate, %, Mortgages 2 4 6 8 10 12 14 16 18 Unemployment Rate, % Total Mortgages stress Total Mortgages Unemployment_stress Unemployment

18

Conclusions

 The risk profile of polish banking system is focused on the retail segment and a loan portfolio characterized by a high exposure to the residential real estate sector;  Large part of these residential loans have been extended throughout a credit boom in foreign currency, namely, CHF, and this introduces an extra risk factor;  However our analysis on systemic figures suggests that, even under very adverse scenarios (sharp depreciation of the local currency and considerable increase in the unemployment rate), the deterioration of the credit portfolio would not be dramatic and would remain fairly manageable. This conclusion, obviously, has to be taken with a pinch of salt, given the “youth” of the portfolio for the entire Polish banking system;  In particular, under a mid term “consensus” scenario (stronger PLN and higher CHF interest rates) we find that the major concerns are raised by the housing loan portfolio vintages extended in 2007 and in the first half of 2008, due to the sharp increase of the installments and the high residual capital to be given back by the clients. For these vintages, DTI increases up to 40% w.r.t. the DTI at origination;  Even under a negative macro scenario (unemployment rate up to 16%), the deterioration of the mortgage portfolio is not dramatic;  In sum, our overall judgment is quite benign. The overall Polish environment is better than, say, the Hungarian one: in Poland the National Bank intervened promptly to refrain banks from extending CHF loans, introduced the Recommendation S which puts tight constraints on CHF lending and increased the risk weight of these loan products to 100% as of June 2012. Finally, the Government, recently confirmed, has been trying to find smooth solutions without burdening excessively the banking sector.