ECONOMIC POLICY NOTE 28/10/2016

Non-performing loans in the euro periphery

were not built in a day

AGNIESZKA GEHRINGER

Non-performing loans (NPL) in the Southern European countries have continued to grow relative

to total loans since the great financial crisis.

The bulk of the NPL stock is in the corporate sector, and more specifically, in construction and

real estate activities. But difficulties affect also other sectors.

Sluggish economic growth, a low rate of innovation, a weak business environment, and the

common monetary policy of the ECB are responsible for the weakness of banks.

According to the ECB Banking Supervision’s as-

sessment of the bank stress test from July 2016,

there is no reason to worry about the European

banking system. In the press release announcing

the results of the test, the ECB stated: “The

results of EU-wide bank stress tests show that

euro area banks improved their resilience and

overall supervisory capital expectations will

remain broadly stable compared to 2015…”.1

1 Danièle Nouy, Chair of the ECB’s Supervisory Board went further to assure that “[t]he results reflect the significant amount of capital raised and the additional balance sheet repairs by the banks over the past two years. (…)The banking sector today is more resilient and can much better absorb economic shocks than two years ago.” See the press release of the ECB’s Banking Super-vision from July 29, 2016, available at: https://www.bankingsupervision. europa.eu/press/ pr/date/ 2016/html/sr160729.en.html.

Markets seem to take a different view. Both

asset and bond prices for banks have fallen be-

hind their counterparts in other industries. The

EURO Stoxx financials has lost almost 60 points

since 2007 compared to -22 points for the EURO

Stoxx broad index. Although the bond price

index IBoxx Banks gained in value in absolute

terms – due to falling interest rates – it stood 14

points below the value for the IBoxx overall

index in September 2016 (Figure 1).

A similar opinion has been recently expressed by the ECB’s chief economist, Peter Praet, in his speech at VII Financial Forum in Madrid on October 4, 2016. In the opening sentence, he stated: “The euro area banking system today is stronger than when it entered the financial crisis.”

2

These developments reflect investors’ worries

about the banking sector. Although problems

have affected all European banks, they have

been most serious in Southern European coun-

tries, particularly, Cyprus, Greece, Italy, Portugal

and Spain.

However, this is not reflected in the Tier 1 capi-

tal ratios (so called CET1 capital levels), which

are the main indicator used in the bank stress

test of the ECB and of the European Banking

Authority. In all five countries, these ratios in-

creased rapidly to levels between 12 and 16% -

well above 5.5% broadly considered as the

minimum acceptable level by the European

banking supervisory authorities (Figure 2).

Much of the evidence for banks losing ground

lies in the continuous rise of non-performing

loans (NPLs) on their books.2 Measured as a

ratio of total bank loans, NPLs increased re-

markably in all Southern European countries

(Figure 3). Only in the case of Spain the ratio has

started falling since the end of 2013 thanks to

improving GDP growth. But given the political

uncertainty and unresolved structural problems,

it is unsure how persistent the recent reduction

of NPL ratios is going to be.

2 Since January 2015, a new harmonized definition of non-

performing loans by the European Banking Authority is in force. They are defined as past due more than 90 days and/or unlikely to pay.

Figure 1. Bond (IBoxx overall and IBoxx banks) and shares (EURO Stoxx broad and EURO Stoxx financials) price indexes,

2007=100.

Source: Haver Analytics

75

85

95

105

115

125

20

40

60

80

100

120

Jan

-07

May

-07

Sep

-07

Jan

-08

May

-08

Sep

-08

Jan

-09

May

-09

Sep

-09

Jan

-10

May

-10

Sep

-10

Jan

-11

May

-11

Sep

-11

Jan

-12

May

-12

Sep

-12

Jan

-13

May

-13

Sep

-13

Jan

-14

May

-14

Sep

-14

Jan

-15

May

-15

Sep

-15

Jan

-16

May

-16

Sep

-16

EURO Stoxx financials (left scale) EURO Stoxx broad (left scale)

IBoxx EUR overall (right scale) IBoxx EUR banks (right scale)

3

Figure 2. CET1 capital levels of banks in the Euro periphery.

Source: Haver Analytics

Figure 3. Non-performing loans as a ratio of total loans.

Note: Due to better data availability, for Portugal, the ratios represent NPLs to the private non-financial sector (households

and non-financial corporations), whereas for Italy, we use bad debts (which are the main subcategory of non-performing

loans) to total loans ratios for financing economic activities (i.e. total across all branches of economic activities). Based on

annual data, the NPL ratio for Italy stood at 16 % in 2015.

Source: Banco de Portugal, Thomson Reuters

0

2

4

6

8

10

12

14

16

18

2007 2008 2009 2010 2011 2012 2013 2014 2015

%

Portugal Italy Cyprus Greece Spain security threshold

0

10

20

30

40

50

60

0

2

4

6

8

10

12

14

16

18

20

Q2

20

07

Q4

20

07

Q2

20

08

Q4

20

08

Q2

20

09

Q4

20

09

Q2

20

10

Q4

20

10

Q2

20

11

Q4

20

11

Q2

20

12

Q4

20

12

Q2

20

13

Q4

20

13

Q2

20

14

Q4

20

14

Q2

20

15

Q4

20

15

Q2

20

16

% %

Italy Spain Portugal Greece (r.h.s) Cyprus (r.h.s)

4

A sectoral view on NPL ratios

Using data for Cyprus, Italy, Portugal and Spain,3

a more detailed picture on the sectoral compo-

sition (in terms of economic sectors and of in-

dustries) of NPL can be drawn.

In all countries, a great proportion of NPLs has

its origins in the private sector, with non-

financial corporations dominating the numbers

in absolute terms (Figure 4).

3 Comparable data for Greece are not available.

At the industry level, construction and real es-

tate appear among those industries with the

highest NPL ratios. Especially in Spain, this is

clearly the consequence of subdued housing

investment during the deepening of the great

financial crisis. However, also other, more tradi-

tional industries, like textiles, wood manufactur-

ing and motor vehicle manufacturing in Italy,

agriculture in Cyprus and hotels and restaurants

in Spain are heavily affected (Figure 5). This

suggests that the problem is of truly systemic

nature.

Figure 4. Non-performing loans over total loans by sector.

Note: In all figures, the ratios are expressed as shares in total loans. For Cyprus, comparable data prior to December 2014 are not available. In Italy, we show data for bad debts, as data for non-performing loans by sector are not available.

Source: Central Bank of Cyprus, Banca d’Italia, Banco de España, Banco de Portugal

48

50

52

54

56

58

60

Dec

14

Feb

15

Ap

r 1

5

Jun

15

Au

g 1

5

Oct

15

Dec

15

Feb

16

Ap

r 1

6

Jun

e 1

6

%

Cyprus non-financial corporations

households

0

5

10

15

20

Jan

07

Sep

07

Mai

08

Jan

09

Sep

09

Mai

10

Jan

11

Sep

11

Mai

12

Jan

13

Sep

13

Mai

14

Jan

15

Sep

15

Mai

16

%

Italy

non-financial corporations

producer households (up to 5 employees)

consumer households

0

4

8

12

16

20

07

:Q1

07

:Q4

08

:Q3

09

:Q2

10

:Q1

10

:Q4

11

:Q3

12

:Q2

13

:Q1

13

:Q4

14

:Q3

15

:Q2

16

:Q1

%

Portugal

non-financial corporations households

0

5

10

15

20

25

07

:Q1

07

:Q4

08

:Q3

09

:Q2

10

:Q1

10

:Q4

11

:Q3

12

:Q2

13

:Q1

13

:Q4

14

:Q3

15

:Q2

16

:Q1

%

Spain

non-financial corporations households

5

Figure 5. Non-performing loans over total loans across industries.

Note: Data older than Q2 2013 are not available.

Source: Central Bank of Cyprus

Note: We show data for bad debts, as data for non-performing loans by industry are not available.

Source: Banca d’Italia

0

10

20

30

40

50

60

70

80

agri

cult

ure

con

stru

ctio

n

real

est

ate

pro

fess

., s

cien

t. &

tec

h.

acti

viti

es

adm

in. &

su

pp

ort

ser

vice

arts

, en

tert

. & r

ecre

atio

n

oth

er s

ervi

ces

tota

l

%

Cyprus

Q2 2013

Q2 2016

0

5

10

15

20

25

30

35

May

10

Jul 1

1

Sep

10

No

v 1

0

Jan

11

Mrz

11

May

11

Jul 1

1

Sep

11

No

v 1

1

Jan

12

Mrz

12

May

12

Jul 1

2

Sep

12

No

v 1

2

Jan

13

Mrz

13

May

13

Jul 1

3

Sep

13

No

v 1

3

Jan

14

Mrz

14

May

14

Jul 1

4

Sep

14

No

v 1

4

Jan

15

Mrz

15

May

15

Jul 1

5

Sep

15

No

v 1

5

Jan

16

Mrz

16

May

16

Jul 1

6

%

Italy

motor vehicle manufacturing textile companies

wood products manufacturing total

construction real estate

6

Figure 5 cont.

Source: Banco de Portugal

Source: Banco de España

0

5

10

15

20

25

30

35

02

:Q4

03

:Q2

03

:Q4

04

:Q2

04

:Q4

05

:Q2

05

:Q4

06

:Q2

06

:Q4

07

:Q2

07

:Q4

08

:Q2

08

:Q4

09

:Q2

09

:Q4

10

:Q2

10

:Q4

11

:Q2

11

:Q4

12

:Q2

12

:Q4

13

:Q2

13

:Q4

14

:Q2

14

:Q4

15

:Q2

15

:Q4

16

:Q2

%

Portugal

wood & cork total furniture construction real estate

0

5

10

15

20

25

30

35

40

45

98

:Q4

99

:Q2

99

:Q4

00

:Q2

00

:Q4

01

:Q2

01

:Q4

02

:Q2

02

:Q4

03

:Q2

03

:Q4

04

:Q2

04

:Q4

05

:Q2

05

:Q4

06

:Q2

06

:Q4

07

:Q2

07

:Q4

08

:Q2

08

:Q4

09

:Q2

09

:Q4

10

:Q2

10

:Q4

11

:Q2

11

:Q4

12

:Q2

12

:Q4

13

:Q2

13

:Q4

14

:Q2

14

:Q4

15

:Q2

15

:Q4

16

:Q2

%

Spain

total construction hotels & restaurants real estate

7

Economic underperformance of Southern Europe

Problems with servicing debt and the substan-

tial accumulation of non-performing loans by

banks in Southern Europe do not reflect a single

factor but a mix of different circumstances.

Most importantly, rigid labor markets, weak

business environments and, as a consequence

of this, a weak economy are among the most

severe problems.

Labor market rigidities, notably barriers to hir-

ing and firing, are at the root of the problems,

leading to persistently high unemployment

rates (Figure 6). Unemployment rates are al-

most three percentage points higher in Portu-

gal, Italy and Cyprus than in the rest of the euro

area. They exceed this benchmark by 13 per-

centage points in Spain and 16 percentage

points in Greece. If labor market rigidities were

lower, unemployment would be lower as well.

However, structural problems reside also else-

where. According to different measures of the

strength of the business environment, all five

economies are weaker than the average of the

rest of the euro area (Figure 7). Although almost

all indicators considered in Figure 7 – corrup-

tion, competitiveness and ease of doing busi-

ness – improved in recent years, the improve-

ments were not strong enough to eliminate or

substantially reduce non-performing loans ac-

cumulated in earlier years.

The evidence for this is in Figure 8 which shows

a negative relationship between the real GDP

growth and the NPL ratio: a one percentage

point decrease in the GDP growth rate was as-

sociated with a 1.7 percentage points increase

in the NPL ratio on average for Southern Europe

in the period 2008-2015.

Figure 6. Unemployment rates.

Note: Rest of the euro area include: Austria, Belgium, Estonia, Finland, France, Germany, Ireland, Latvia, Lithuania, Luxem-burg, Malta, Netherlands, Slovakia and Slovenia.

Source: Haver Analytics

0

5

10

15

20

25

30

07

:Q1

07

:Q3

08

:Q1

08

:Q3

09

:Q1

09

:Q3

10

:Q1

10

:Q3

11

:Q1

11

:Q3

12

:Q1

12

:Q3

13

:Q1

13

:Q3

14

:Q1

14

:Q3

15

:Q1

15

:Q3

16

:Q1

%

Rest of the Euro area Greece Cyprus Italy Portugal Spain

8

Figure 7. Business environment, corruption and competitiveness.

Note: * For the Global Competitiveness Index the last available observation is for 2016. Rest of the euro area include: Aus-

tria, Belgium, Estonia, Finland, France, Germany, Ireland, Latvia, Lithuania, Luxemburg, Malta, Netherlands, Slovakia and

Slovenia. The Corruption Index ranks countries by their perceived levels of corruption, as determined by expert assessments

and opinion surveys. The Corruption Rank indicates the country’s position relative to the other countries from 0 (the lowest

corruption) and 100. The Global Competitiveness Index assesses the set of institutions, policies and factors influencing com-

petitiveness conditions. It ranges from 0 (the highest competitiveness score) to 100. The ease of doing business is calculated

based on several indicators influencing the country’s business environment. The country ranking ranges between 1 (the

most supportive business environment) and 190.

Source: World Bank, World Economic Forum, Transparency International

Figure 8. Scatter plot of real GDP growth in relation to the NPL ratio for Cyprus, Greece, Portugal, Italy and Spain, 2008-

2015.

Note: GDP growth rates are two years lagged.

Source: Own elaboration based on Haver Analytics

60

47

45

33

23

30

81

65

43

33

38

29

58

32

61

36

28

23

0 20 40 60 80 100

Greece

Cyprus

Italy

Spain

Portugal

Rest of the euro area

2015* Corruption Rank

Global Competitiveness Index

Ease of Doing Business

86

44

68

46

33

29

79

50

46

34

43

30

70

32

61

31

31

26

0 20 40 60 80 100

Greece

Cyprus

Italy

Spain

Portugal

Rest of the euro area

ø 2008-2015* Corruption Rank

Global Competitiveness Index

Ease of Doing Business

y = -1.69x + 10.83

R² = 0,3189

0

5

10

15

20

25

30

35

40

-12 -10 -8 -6 -4 -2 0 2 4 6 8 10

NP

L ra

tio

real GDP growth

9

Although much of the economic underperfor-

mance comes from country-specific factors,

there are reasons to believe that the particular

institutional setting of the European Monetary

Union (EMU) and the common monetary policy

of the ECB played a crucial role in exacerbating

the dynamics of NPL accumulation. Figure 9

shows how the commitment to participate in

the EMU in the years prior to 1999 and – after

the establishment of the euro – the common

monetary policy drove interest rates down

across the Southern European economies and

how it contributed to the accelerated credit

growth thereafter.

An econometric assessment

To deepen the above discussion, we perform an

econometric analysis on panel data for Italy and

Spain, for the first quarter of 1999 to the second

quarter of 2016.4 Applying the dynamic OLS

estimation technique, the following empirical

model is estimated:

4 This is the longest period possible to construct with the

available data. Our data come from Haver Analytics.

𝑦𝑖𝑡 = 𝛼 + βx𝑖𝑡 + 𝜇𝑖 + 𝜔𝑖𝑡 (1)

where yit is the dependent variable, given by the

ratio of non-performing loans in country i at

time t, 𝛼 is a constant term, x𝑖𝑡 is a vector of

country-specific explanatory variables, β the

vector of coefficients to be estimated, 𝜇𝑖 is a

country fixed effect and 𝜔𝑖𝑡 a composite error

term that has the following form:

𝜔𝑖𝑡 = ∑ b1𝑝∆x𝑖𝑡−𝑝+𝑝−𝑝 + 𝜀𝑖𝑡. (2)

The first term in equation 2 is composed of p

leads and lags of first-differenced explanatory

variables, while the second term is the usual

random error.5

5 Dynamic OLS method was introduced by James Stock and

Mark Watson [“A simple estimator of cointegrating vectors in higher order of integrated systems.” Econometrica 61(4), 783-820, July 1993.] to account for possible endoge-neity deriving from the influence of short-term dynamic factors of explanatory variables. They explicitly correct for this in the composite error term, as shown in equation 1. The application of the method requires the variables to be non-stationary and cointegrated – the condition that we tested for and confirmed in our sample.

Figure 9. Long term interest rates (left panel) and credit to the private sector as percentage of GPD (right panel) in South-

ern Europe and Germany.

Source: Haver Analytics

-10

0

10

20

30

-5

0

5

10

15

Jan

-93

Ap

r-9

4Ju

l-9

5

Oct

-96

Jan

-98

Ap

r-9

9Ju

l-0

0

Oct

-01

Jan

-03

Ap

r-0

4

Jul-

05

Oct

-06

Jan

-08

Ap

r-0

9

Jul-

10

Oct

-11

Jan

-13

Ap

r-1

4Ju

l-1

5

%

10-year gov. bond yield

Germany SpainPortugal ItalyGreece (r.h.s)

20

60

100

140

180

Q1

-97

Q1

-98

Q1

-99

Q1

-00

Q1

-01

Q1

-02

Q1

-03

Q1

-04

Q1

-05

Q1

-06

Q1

-07

Q1

-08

Q1

-09

Q1

-10

Q1

-11

Q1

-12

Q1

-13

Q1

-14

Q1

-15

Q1

-16

% of GDP

credit to the private sector as % of GDP

Spain Italy Greece Germany

10

Among the explanatory variables, we control for

the level of long-term interest rates, as repre-

sented by yields on 10-year government bonds.

This variable should reflect the credit conditions

in the respective economies. Moreover, given

that these conditions are influenced by interest

rates set by the ECB, we can relate the stock of

NPLs to the common monetary policy in the

euro area.6

We also include the unemployment rate, as

unemployment makes debt servicing harder for

households. Moreover, increasing unemploy-

ment may signal diminishing production, which

consequently leads to decreasing revenues and

increasing corporate debt burdens. We also

control for productivity: the higher productivity

the easier it is to generate economic value and

repay debt. Finally, we include two variables

reflecting developments in capital markets, a

stock market price index and a housing price

index. The former accounts for general financial

market conditions, but additionally it is an indi-

cator of bank asset quality: declining stock pric-

es can exercise a negative impact on bank asset

quality. The latter reflects the value of assets

frequently used as collateral for credit: rising

housing prices mean an increasing value of col-

lateral and thus a decreasing stock of non-

performing loans. Finally, to account for possi-

ble cyclical influences, in one specification, we

include the real GDP growth rate.7

6 For evidence regarding the influence of monetary policy on long-term interest rates, see Agnieszka Gehringer and Thomas Mayer (2015), Understanding low interest rates, Flossbach von Storch Research Institute, Economic Policy Note Nr. 23/10/2015. 7 Most of these control variables are considered as a standard in previous empirical analyses, for instance, Roland Beck, Petr Jakubil and Anamaria Piloiu (2013), Non-performin loans: What matters in addition to the economic cycle? ECB Working Paper Nr. 1515, or Dimitrios P. Louzis, Angelos T. Vouldis and Vasilios L. Metaxas (2010), Macroeconomic and bank-specific determinants of non-performing loans in Greece: A comparative study of mortgage, business and consumer loan portfolios. Bank of Greece Working Paper Nr. 118.

It is reasonable to assume that the influence of

economic factors on the stock of NPLs is not

immediate but requires time, before a loan be-

comes eventually non-performing. For this rea-

son, we consider in separate specifications ex-

planatory variables that are two-, three- and

four-year lagged. This permits us also to better

assess the timing of influence of such factors on

NPLs.

The table below summarizes the results of the

regressions. They show a strong and significant

role played by the credit conditions and –

through the credit channel – by monetary policy

of the ECB. A one percentage point decline in

long-term interest rates led to 1 – 1.8 percent-

age points increase in the NPL ratios, depending

on the time dimension. The strongest influence

is seen after two years and it diminishes as time

goes by. A similarly strong effect, but in the

opposite direction, is found for the unemploy-

ment rate, meaning that increasing unemploy-

ment leads to the accumulation of the NPL

stock. Both rising productivity and housing pric-

es contribute to the reduction of NPLs on the

banks’ books. Finally, for the stock market price

index we do not find any significant effect.

11

A solid banking system?

European banks are in a state of emergency.

The view of the European banking supervisors

and monetary policy makers that the euro area

banking system is stronger today than at the

beginning of the financial crisis gives a false

sense of security.

From the analysis of Southern European coun-

tries, two main conclusions emerge. First, NPL

stocks are not only high but also broad-based,

involving different economic sectors. Second,

the accumulation of NPLs was a long-lasting

process, reflecting poor economic growth and

supply-side rigidities. Based on estimations for

Italy and Spain, we find evidence that both low

productivity and high unemployment were at

the roots of the NPL accumulation. But the

common monetary policy also played a crucial

role. By depressing interest rates, it boosted the

demand for credit, which often resulted in un-

sound borrowing. This paved the way for the

accumulation of non-performing loans.

Table Dynamic OLS estimates of equation 1.

(1) (2) (3) (4)

10-year gov. bond -1.810***

(0.233)

-1.476***

(0.256)

-1.290***

(0.260)

-1.052**

(0.335)

Unemployment rate

1.491***

(0.128)

1.252***

(0.119)

1.432***

(0.134)

1.357***

(0.146)

Productivity -0.443**

(0.171)

-0.526***

(0.153)

-0.375**

(0.164)

-0.323*

(0.171)

Stock market price index -0.016

(0.039)

-0.011

(0.035)

-0.010

(0.033)

0.028

(0.041)

Housing price index -0.128**

(0.057)

-0.277***

(0.056)

-0.212***

(0.057)

-0.277***

(0.060)

GDP growth rate -- 0.691**

(0.257)

-- --

N. obs. 88 88 84 80

R-squared adj. 0.943 0.956 0.949 0.944

UR test -3.698***

[0.000]

-3.767***

[0.000]

-4.201***

[0.000]

-4.976***

[0.000]

Note: Dependent variable is the ratio of non-performing loans (or their sub-category of bad debts for Italy) to total loans. *,

**, and *** denote statistical significance at the 10%, 5%, and 1% level. Robust standard errors are in parenthesis. The last

row reports the test statistic and, in squared parenthesis, its p-values of the Im-Pesaran-Shin panel unit root test of the

residual term from the respective estimations. The null hypothesis of the test assumes the presence of a unit root in all

panels. All specifications include country fixed effects. Column (1) reports results from the estimations with explanatory

variables two years lagged, column (2) the same as column (1), but including the GDP growth rate to account for the busi-

ness cycle, column (3) with three years lagged explanatory variables and column (4) with four years explanatory variables.

12

LEGAL NOTICE

The information contained and opinions expressed in this document reflect the views of the author at the time of publica-

tion and are subject to change without prior notice. Forward-looking statements reflect the judgement and future expecta-

tions of the author. The opinions and expectations found in this document may differ from estimations found in other

documents of Flossbach von Storch AG. The above information is provided for informational purposes only and without any

obligation, whether contractual or otherwise. This document does not constitute an offer to sell, purchase or subscribe to

securities or other assets. The information and estimates contained herein do not constitute investment advice or any

other form of recommendation. All information has been compiled with care. However, no guarantee is given as to the

accuracy and completeness of information and no liability is accepted. Past performance is not a reliable indicator of fu-

ture performance. All authorial rights and other rights, titles and claims (including copyrights, brands, patents, intellectual

property rights and other rights) to, for and from all the information in this publication are subject, without restriction, to

the applicable provisions and property rights of the registered owners. You do not acquire any rights to the contents. Copy-

right for contents created and published by Flossbach von Storch AG remains solely with Flossbach von Storch AG. Such

content may not be reproduced or used in full or in part without the written approval of Flossbach von Storch AG.

Reprinting or making the content publicly available – in particular by including it in third-party websites – together with

reproduction on data storage devices of any kind requires the prior written consent of Flossbach von Storch AG.

© 2016 Flossbach von Storch. All rights reserved.

SITE INFORMATION

Publisher: Flossbach von Storch AG, Research Institute, Ottoplatz 1, 50679 Cologne, Germany; Phone +49 221 33 88-291,

research@fvsag.com, Directors: Dr. Bert Flossbach, Kurt von Storch, Dirk von Velsen; Registration: No. 30 768 in the Com-

mercial and Companies Register held at Cologne District Court; VAT-No. DE200075205; Supervisory authority: German

Federal Financial Services Supervisory Authority, Marie-Curie-Straße 24 – 28, 60439 Frankfurt / Graurheindorfer Straße 108,

53117 Bonn, www.bafin.de; Author: Priv. Doz. Angieszka Gehringer, PhD; Editorial deadline: 27. October 2016