Eastern European Economics

Fo
rP

MACROECONOMIC AND BANK-SPECIFIC DETERMINANTS OF

NON-PERFORMING LOANS: THE CASE OF BALTIC STATES
ee
rR

Journal: Eastern European Economics

Manuscript ID MEEE-2018-0011
ev

Manuscript Type: Original Article

Non-performing loans, Macroeconomic determinants, Bank-specific

Keywords:
determinants, Baltic States, System Generalized Method of Moments
iew

This study examines macroeconomic and bank-specific determinants of

non-performing loans for a panel of 21 banks from the Baltics, for the
period 2005 – 2016. We apply three alternative estimation techniques:
fixed- effects model, difference and system Generalized Method of
Moments to find the key determinants in the regression model. Results
Abstract:
On

provide evidence that the most important macroeconomic factors

influencing NPLs are growth of GDP,public debt, inflation and
unemployment. As for the bank-specific determinants, we found that
equity to total assets ratio, return on assets, return on equity and growth
of gross loans have an impact on the amount of NPLs.
ly

URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]

Page 1 of 28 Eastern European Economics

1
2
3 Table 1 Correlation matrix
4
5 NPLs CGD UN GDPG INF ROA ROE ETA GGL
6 NPLs 1.0000
7 PD 0.3885 1.0000
8
9 UN 0.3607 0.5175 1.0000
10 GDPG -0.0132 0.0208 -0.3607 1.0000
11
INF -0.3340 -0.4607 -0.4184 0.0044 1.0000
12
13 ROA -0.6456 -0.2552 -0.3011 -0.0107 0.1947 1.0000
14 ROE -0.2860 -0.1614 -0.3616 0.0412 0.1507 0.7914 1.0000
15
16 ETA -0.4758 -0.1769 -0.0168 -0.0268 0.0156 0.2842 -0.1283 1.0000
Fo
17 GGL -0.4508 -0.4037 -0.5620 0.1138 0.5602 0.3543 0.3411 0.0309 1.0000
18 Source: Author’s calculations
19
rP
20
21
22
23
ee

24
Table 2 Panel unit root tests
25
26 Test ADF-Fisher Chi square PP-Fisher Chi square Breitung
rR

27 variables
28 Level First Level First Level First
29 Difference Difference Difference
ev

NPLs 0.4612 -1.478***

30 99.3397*** 117.431***
31 PD 205.422*** 164.631*** 6.5870***
32 UN 100.287*** 98.0210*** 3.8414***
iew

33 GDPG 86.2740*** 28.7410* 5.4369***

34 3.0501 4.334***
INF 42.7059* 14.5055*
35
ROA 1.5413**
36 68.2197* 101.425***
37 ROE 44.9057* 98.5600*** 1.3679 3.521***
On

38 ETA 62.3250* 67.6904*** 2.0207*

39 GGL 68.9775** 71.1816*** 3.6052*
40 Source:Author’s calculations
41 ***, **, * denote statistical significance at the 1, 5, 10 percent level respectively
ly

42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 2 of 28

1
2
3
4
Table 3 Estimation Results
5
6 Fixed Effects (FE) regressions Difference GMM System GMM
Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
7
NPL(-1) 0.51*** 0.78*** 0.57*** 0.62*** 0.87*** 0.69***
8 (0.57) (0.78) (0.43) (1.23) (1.17) (0.78)
9 Const -20.07*** -9.06*** 11.38** 10.94*** -6.80 6.57** 8.22*** -8.00** 1.81
10 (1.08) (1.44) (0.35) (8.23) (1.14) (0.86) (0.72) (0.45) (1.28)
11 0.67 -0.25*** 0.15 0.12* 0.12 -0.03
12 PD (0.32) (0.03) (0.26) (1.06) (0.26) (0.94)
13 UN
0.51** -0.88 0.61*** 0.20 0.67 *** 0.40***
14 (0.85) (0.72) (0.43) (0.27) (0.18) (0.11)
15 -0.14 -0.16* -0.17 -0.63** -0.23*** -1.58***
GDPG (0.33) (0.08) (0.03) (0.60) (0.43) (1.58)
16
Fo
0.15 0.24 0.22 0.21*** 0.32 0.26 *
17 INF (1.62) (0.24) (1.37) (0.04) (0.75) (0.32)
18 -0.88*** -0.84* -0.68*** -0.62 -0.97*** -0.84***
19 ROA (0.14) (0.97) (0.44) (0.06) (0.38) (0.02)
rP
20 -0.45*** -0.45*** -0.43*** -0.40 -0.40 -0.37***
21 ROE (0.42) (0.61) (1.52) (0.82) (0.72) (0.18)
22 -0.43*** -0.42*** -0.21* -0.19* -0.09 -0.91
ETA
23
ee
(0.20) (1.01) (0.49) (0.67) (1.13) (0.19)
24 GGL
0.20*** 0.15*** 0.85*** 0.79*** 0.68*** 0.45*
(0.59) (0.73) (0.28) (0.04) (1.52) (0.89)
25
DUM 2008 2.41* 0.59 1.74 2.16 2.32* 0.90 1.31* 1.50 1.48
26
rR

(0.92) (0.11) (0.85) (0.80) (0.58) (0.21) (0.88) (0.52) (0.31)

27 DUM 2009 1.11*** 1.17 2.45** 1.65*** 1.55* 1.96* 0.68 2.29* 1.28**
28 (0.95) (0.98) (0.59) (0.25) (0.45) (0.47) (0.22) (1.04) (0.57)
29
ev

Number of
30 Banks 21 21 21
31 Hansen test
32 (p-value) 0.478 0.515 0.457 0.539 0.619 0.721
iew

33 Test for AR(1)

34 errors 0.068 0.005 0.032 0.047 0.018 0.062
35 Test for AR(2)
36 errors 0.227 0.216 0.679 0.565 0.578 0.684
37
On

38 Source: Author’s calculation

39 1. Arellano-Bond test that average autocovariance in residuals of order 1 is 0 (HB0B: No autocorrelation).
40 2. Arellano-Bond test that average autocovariance in residuals of order 2 is 0 (HB0B: No autocorrelation).
41 Standard errors are in parenthesis
ly

42 *,** and *** show that the null hypothesis can be rejected at 10%, 5% and 1% significance levels respectively
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 3 of 28 Eastern European Economics

1
2
3
4
5 MACROECONOMIC AND BANK-SPECIFIC DETERMINANTS OF NON-
6
7 PERFORMING LOANS: THE CASE OF BALTIC STATES
8
9
10
11
12 Abstract
13
14
15
16 This study examines macroeconomic and bank-specific determinants of non-performing
Fo
17
18 loans for a panel of 21 banks from the Baltics, for the period 2005 – 2016. We apply three
19
rP
20 alternative estimation techniques: fixed- effects model, difference and system Generalized
21
22
Method of Moments to find the key determinants in the regression model. Results provide
23
ee

24
25
evidence that the most important macroeconomic factors influencing NPLs are growth of
26
rR

27 GDP,public debt, inflation and unemployment. As for the bank-specific determinants, we

28
29 found that equity to total assets ratio, return on assets, return on equity and growth of gross
ev

30
31 loans have an impact on the amount of NPLs.
32
iew

33
34
35 Key words: Non-performing loans; Macroeconomic determinants; Bank-specific
36
37
determinants; Baltic States; System Generalized Method of Moments
On

38
39
40
41
ly

42 JEL classification F61,F62,EG01,G21,

43
44
45
46
47
48
49
50
51
52
53 Introduction
54
55
56
57
58
1
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 4 of 28

1
2
3 Information on the banks’ loan quality is an important issue that has aroused the interest of
4
5 the public as a user of banking services, the public as a potential investor in the banks’ equity,
6
7 the banks’ management, the financial markets, the banking supervisors and regulators in
8
9 terms of controlling the stability of the financial system and of the academic circles. This
10
11
12
interest has intensified significantly in the last two decades. Namely, deregulation,
13
14 technological change and the globalization of goods and financial markets, the financial crisis
15
16 of the 1990s the global economic crisis of 2008-2009 and the European debt crisis of 2011-
Fo
17
18 2012 have all had an impact on banks’ loan quality.
19
rP
20 One of the most common indicators used to identify the banks’ loan quality is the
21
22 ratio of non-performing loans (NPLs). An increase in this ratio may signal a deterioration in
23
ee

24
25
banking sector results (Mörttinen, Poloni, and Vesala 2005). Experience shows that a rapid
26
rR

27 build‐up of NPLs has a crucial role in banking crises (Demirgüç-Kunt and Detragiache
28
29 1998), (González‐Hermosillo 1999).
ev

30
31 This experience was confirmed during the last few years, that is, since the onset of the
32
iew

33 global financial crisis in 2007-2008, when the levels of NPLs have significantly increased
34
35
across countries. In fact, according to analysts, the amount of NPLs is expected to increase
36
37
dramatically in the coming years, affecting the liquidity and profitability of banks and thereby
On

38
39
40 the financial stability of the banking systems (Makri, Tsagkanos, and Bellas 2014). However,
41
ly

42 although after 2007-2008 almost all countries in the world were faced with rapid growth of
43
44 NPLs, the growth varied significantly among the different groups of countries, and among
45
46 countries in the same group. For example, in 2008 the amount of NPLs as a share of total
47
48 loans in high income OECD countries was 3%, increasing to 8% in 2014, whereas in Central
49
50
51
and Southeastern Europe it was 4% in 2002, and has reached almost 15% in 2014. In
52
53 countries such as Macedonia, Poland, and Lithuania the NPLs in 2002 were 6.7%, 2.8% and
54
55 6.1%, respectively, increasing to 11.3%, 5%, and 10%, respectively, in 2014. The large
56
57
58
2
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 5 of 28 Eastern European Economics

1
2
3 disparity across countries raises questions about what causes this variation and thus what are
4
5 the determinants for NPLs. There is a rapidly growing number of empirical studies - which
6
7 analyze factors that influence the NPLs (De Nicoló, et al. 2003; Quagliariello 2003;
8
9 Hoggarth, Logan, and Zicchino 2005; Fofack 2005; Babouček and Jančar 2005; Espinoza,
10
11
12
and Prasad 2010; Vogiazas and Nikolaidu 2011; Klein 2013). All these authors have
13
14 proposed a variety of different macroeconomic and institutional factors as possible
15
16 determinants of NPLs.
Fo
17
18 Considering the ratio of NPLs to total loans as a measure for NPLs, the aim of this
19
rP
20 study is to evaluate the macroeconomic and bank-specific determinants of NPLs in the Baltic
21
22 States. We focused on the Baltic countries (Estonia, Latvia, and Lithuania), for the period
23
ee

24
25
2005- 2016, beecause these countries, more than the other countries in Europe, have been
26
rR

27 affected by the subprime mortgage crisis. For example, in the 2009, in Estonia the NPLs were
28
29 5.9 %, while in Latvia in the same year the percentage of NPLs was 14.3. The greatest rise of
ev

30
31 the NPL ratio was recorded in Lithuania, from 6% in 2008 to almost 24% in 2009. Also,
32
iew

33 according to the (Tang, Zoli, and Klytchnikova 2000; Koivu 2002).these countries are
34
35 relatively homogenous and have adopted a broadly common macroeconomic model; they
36
37
have similar institutions and economies, and their banking sectors share important
On

38
39
40 commonalities. Although, as we have seen, there are many studies that analyze the
41
ly

42 determinants of NPLs in many countries and regions, not many authors include the banking
43
44 sectors of the Baltic States in their research. According to the authors, six studies examine
45
46 the determinants of NPLs in the Baltic countries (Festic and Repina 2009; Kavkler and
47
48 Mejra Festic 2010; Fainstein and Novikov 2011;Klein 2013; Donath, Cerna, and Oprea
49
50
2014; Kjosevski and Petkovski 2017) Almost all of them have certain similar characteristics.
51
52
53 Namely, in these studies some of the authors were using only macroeconomic variables
54
55 (Fainstein and Novikov 2011; Donath, Cerna, and Oprea 2014) or have analyzed each
56
57
58
3
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 6 of 28

1
2
3 country individually, not in a panel model (Festic and Repina 2009; Kavkler and Festic 2010;
4
5 Fainstein and Novikov 2011; Donath, Cerna, and Oprea 2014). The majority of studies,
6
7 except the articles of (Klein,2013 Kjosevski and Petkovski 2017), which have analyzed
8
9 determinants of NPLs in Baltic States, have used aggregate data for the whole banking
10
11
12
system of each country and not disaggregated data (examination of individual data for each
13
14 bank). All of them have a relatively short time series of the data.
15
16 Bearing in mind the aforementioned studies this study offers some novelties. As we
Fo
17
18 previously mentioned,the study of Klein (2013) analyzes the Baltic States in a panel data set,
19
rP
20 but within the countries of the Central, Eastern and Southeastern Europe (CESEE), and not as
21
22 a single region. On the other hand, the article of Kjosevski and Petkovski (2017) use
23
ee

24
25
unbalanced panel for the banks in the Baltic States, but apply only one estimation technique
26
rR

27 (difference Generalized Method of Moments). To avoid the risk of providing unconsistent

28
29 and biased results by using only one estimation technique, in our study we implemented three
ev

30
31 alternative estimation models (fixed- effects model, difference Generalized Method of
32
iew

33 Moments and system Generalized Method of Moments). We also analyzee longer period,
34
35 from 2005 to 2016.
36
37
Among the main advantages of panel data, compared to other types of data, is that the
On

38
39
40 approach allows testing and adjustment of the assumptions that are implicit in cross-sectional
41
ly

42 analysis (Maddala 2001). The short time series, poor availability and quality of the data have
43
44 been the common reasons to refrain from the analysis of Baltic countries. We have addressed
45
46 these concerns by selecting a more recent time period, including the ups and downs of the
47
48 economic and credit cycles, while making use of better data availability. We also employ a
49
50
thorough data preparation process by eliminating inconsistencies, consolidating the existing
51
52
53 information and filling in the data gaps for banks with more significant market share by using
54
55 the banks’ public reports. On the basis of the studies of (De Bock and Demyanets 2012;
56
57
58
4
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 7 of 28 Eastern European Economics

1
2
3 Louzis, Vouldis, and Metaxas 2010), we have applied a dynamic panel data approach to
4
5 explain the determinants of the NPL in Baltic States. In order to provide consistent and
6
7 unbiased results, we implemented the three alternative estimation techniques fixed effects
8
9 model, difference Generalized Method of Moments and system GMM.
10
11
12
The structure of the article is as follows. After the Introduction, Section 1 gives an
13
14 overview of the literature on empirical findings relevant to the determinants for NPLs.
15
16 Sources of the data used , as well as the methodology are presented in Section 2. Section 3
Fo
17
18 shows the empirical results of determinants and Section 4 concludes the article and gives
19
rP
20 policy recommendations.
21
22
23
ee

24
25
1. Literature Review
26
rR

27 In this section we present a brief sublimate of literature, dealing with the empirical
28
29 findings relevant to the determinants for NPLs. Moreover, given the huge number of
ev

30
31 published studies on this topic (Quagliarello 2007; Boudriga, Boulila, and Jellouli 2009; Dash
32
iew

33 and Kabra 2010; Espinoza and Prasad 2010; Louzis, Vouldis, and Metaxas 2010 Nkusu
34
35 (2011) Castro 2012; Klein 2013), are some of the most relevant studies, we focus on
36
37
empirical literature that examines the determinants of NPLs only in the Baltic States.
On

38
39
40 Empirical results of the above-mentioned studies differ, because of the differences in
41
ly

42 databases, time periods, and the different specifics of each of the countries. However, there
43
44 are some common elements that allow categorizing the determinants of banks’ NPLs are
45
46 usually measured by the ratio of NPLs to total loans. The internal determinants usually
47
48 include: bank-specific variables, such as size of the bank, equity to total assets ratio (ETA),
49
50
return on assets (ROA) or return on equity (ROE) and growth of gross loans (GGL). The
51
52
53 macroeconomic determinants include the GDP growth, unemployment, exchange rate
54
55 interest rate and inflation.
56
57
58
5
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 8 of 28

1
2
3 In their study, Festic and Repina (2009) using panel regression for the period from
4
5 1998Q1 to 2008Q3 examine the impact of macroeconomic and bank-specific determinants
6
7 for NPLs in Baltic States . Their results show that a slowdown in economic activity
8
9 accelerates the growth of the NPLs. Also, the results suggest that rapid growth of credit
10
11
12
harms loan performance, most likely due to soft-loan constraints and macroeconomic
13
14 overheating.
15
16 Kavkler and Festic (2010) applied ordinary least squares method for the period 1997–
Fo
17
18 2007. They analysed effects of 12 financial and macroeconomic variables as predictors for
19
rP
20 NPLs in the Baltic States. . Their results indicated that a sharpslowdown in economic activity
21
22 (GDP, net exports, investment and savings growth) would likely deteriorate the loan
23
ee

24
25
portfolio quality in the Baltic States.
26
rR

27 Fainstein and Novikov (2011) applying a separate vector-error-correction model

28
29 (VECM) for each of these three Baltic countries, used quarterly data for the periods from
ev

30
31 1997 Q3/ 2002Q1 /2004Q1 to 2009Q4 (depending on the country). Their results showed that
32
iew

33 real GDP growth was the most significant determinant of NPLs' growth in all three countries
34
35 and that real estate market growth played an important role in two of these countries (Latvia
36
37
and Lithuania).
On

38
39
40 The research of Klein (2013) examines factors that affected NPLs in CESEE (Bosnia
41
ly

42 and Herzegovina, Bulgaria, Hungary, Croatia, Czech Republic, Estonia, Latvia and
43
44 Lithuania) for the period 1998–2011. He used three alternative estimation techniques: fixed
45
46 effect model, difference GMM and system GMM. The results show that NPLs respond to
47
48 macroeconomic conditions, such as GDP growth, unemployment and inflation.
49
50
Using macroeconomic indicators such as GDP, inflation, unemployment and lending
51
52
53 interest rates for the period 2000–2013, Donath, Cerna, and Oprea (2014) estimated the
54
55 evolution of bad loans ratio in the Baltic States and Romania . Their results show that NPLs'
56
57
58
6
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 9 of 28 Eastern European Economics

1
2
3 had a significant negative correlation with GDP growth in all four countries. The inflation
4
5 rate exhibited a negative correlation with the NPLs in all countries except Lithuania. The
6
7 lending interest rate was positively correlated with the NPLs in all countries except Romania,
8
9 whereas the unemployment positively correlated with the NPLs in each of the four countries.
10
11
12
13
14 2. Data and Methodology
15
16
Fo
17
18 This section identifies the sources of our data, presents the data and describes the
19
rP
20 regression model that we used to investigate the effects of macroeconomic and bank-specific
21
22 determinants on NPLs.
23
ee

24
25
26
rR

27 2.1 Data Source and Sample Characteristics

28
29
ev

30
31 For our research we used an unbalanced panel with 21 banks in the Baltic States. Data
32
iew

33 are based on annual frequency for 2005–2016, and the number of banks was chosen in
34
35 accordance with the avialability of data. Unbalanced panel data are able to have more
36
37
observations and their results are less dependent on a particular period (Rinaldi and Sanchis-
On

38
39
40 Arellano 2006).
41
ly

42 The data for the bank-specific determinants (ROA, ROE, and growth of gross loans)
43
44 were collected from balance sheets, income statements and notes from the annual reports
45
46 from the Bankscope database.The data for macroeconomic-determinants (GDP growth
47
48 unemployment, percentage of total labour force, inflation, consumer prices and public debt
49
50
(percentage of GDP) were obtained from the World Development Indicators (WDI) database.
51
52
53 Before attempting to identify potential internal and external determinants of NPLs, it
54
55 is necessary to identify the dependent determinant. In the literature to date there is no
56
57
58
7
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 10 of 28

1
2
3 internationally harmonized definition that has been applied in all or most countries of the
4
5 world for a considerable period of time. In this context it is worth mentioning that Bankscope
6
7 reports the level of “impaired loans”, which may be different than the official classification of
8
9 NPLs . “Impaired loans” is an accounting concept, which reflects cases in which it is
10
11
12
probable that the creditor will not be able to collect the full amount that is specified in the
13
14 loan agreement, whereas “NPL” is a regulatory concept, which primarily reflects loans that
15
16 are more than 90 days past due Report of the Working Group on NPLs in Central, Eastern
Fo
17
18 and Southeastern Europe (CESEE, 2012). Bearing these differences in mind, we will follow
19
rP
20 (Klein 2013) and treat “impaired loans” as NPLs. In this analysis our dependent variable will
21
22 be the ratio of impaired (NPLs) to total (gross) loans.
23
ee

24
25
The determinants that we used as control determinants, which may explain the NPLs
26
rR

27 of banks, are:
28
29 - Macroeconomic determinants: public debt-PD; GDP growth - GDPG; inflation –
ev

30
31 INF; unemployment – UN;;
32
iew

33 - Bank-specific determinants: ratio of equity to total assets – ETA; return on assets-

34
35 ROA; return on equity- ROE; growth of gross loans - GGL;
36
37
On

38
2.1.1 Macroeconomic Determinants
39
40
41
ly

42
43 At the core of all the previously mentioned studies, the variables related to gross
44
45 domestic product (GDP) are the main macroeconomic determinants of NPLs. In this context,
46
47 several variations of this determinant, such as the annual growth rate of real GDP, the
48
49 production gap, the growth of income per capita, and so on., are well known in the literature.
50
51
However, the real GDP growth rate is by far the most common macroeconomic determinant
52
53
54
used as an example in the studies of: (Gasha and Morales 2004; Jimenez and Saurina 2006;
55
56 Quagliarello 2007; Marcucci and Quagliariello 2008, 2009; Castro 2012; Nkusu 2011; Klein
57
58
8
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 11 of 28 Eastern European Economics

1
2
3 2013; Beck, Jakubik and Piloiu 2013). Hence, we also include the annual growth rate of real
4
5 GDP in our analysis. Thus, we want to examine the effect of the cycle in which the economy
6
7 is, on the credit risk. According to (Nkusu 2011), the growing economy associated with the
8
9 growth of the general level of income and reduced financial stress, and hence GDP growth,
10
11
12
should be negatively correlated with NPLs.
13
14 To reflect the price stability in the model, we follow (Kavkler and Festic 2010;
15
16 Donath, Cerna, and Oprea 2014) and we include the inflation as the general consumer prices
Fo
17
18 rate, but its impact on NPLs is not clear. On the one hand, higher inflation can make debt
19
rP
20 servicing easier by reducing the real value of outstanding loans, but on the other hand, it can
21
22 also weaken borrowers’ ability to service debt by reducing their real income. Gunsel, (2008),
23
ee

24
25
(Rinaldi and Sanchis-Arellano 2006) find a positive correlation between the inflation rate and
26
rR

27 NPLs in North Cyprus and Euro Zone countries. Also in the articles of (Kavkler and Festic
28
29 2010) and (Donath et al. 2014) in the case of Baltic States the results indicates that inflation
ev

30
31 was significant and a positive determinant of NPLs in the Baltic States. In the opposite
32
iew

33 direction, (Sofoklis and Nikolaidu 2011) in the case of Tunisian and Romanian banking
34
35 sectors, found a negative correlation between inflation and credit risk. The study of (Boštjan
36
37
and Aver 2008; Bofondi and Ropele 2011) in the case of Slovenian and Italian banking
On

38
39
40 systems, did not find any influence of inflation on credit risk. Therefore, the relationship
41
ly

42 between inflation and NPLs may be ambiguous.

43
44 Regarding unemployment, it is rational to suppose that an increase in the
45
46 unemployment should influence negatively the cash flow streams of households and increase
47
48 the debt burden. With regard to firms, increases in unemployment may signal a decrease in
49
50
production as a consequence of a drop in effective demand. This may lead to a decrease in
51
52
53 revenues and a fragile debt condition. Several empirical studies have investigated the relation
54
55 between unemployment and NPLs and they have found it to be positive (Bofondi and Ropele
56
57
58
9
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 12 of 28

1
2
3 2011; Godlewski 2008; Makri, Tsagkanos, and Bellas 2014). Therefore, we expect that an
4
5 increase in the unemployment will lead to an increase in the NPLs.
6
7 When the global economic crisis started in 2008, it has significantly affected government
8
9 finances and then it has extended its negative impact to the banks. Taking this point into
10
11
12
consideration,we will follow (Makri, Tsagkanos, and Bellas 2014) and we will also include
13
14 public debt as a determinant of NPLs. We anticipate a positive association between NPLs and
15
16 public debt.
Fo
17
18 2.2.2 Bank-Specific Determinants
19
rP
20
21
22 The share of equity in total assets is an important determinant of NPLs . According to the
23
ee

24
25
“moral hazard” hypothesis, discussed by (Keeton and Morris 1987) banks with relatively low
26
rR

27 capital respond to moral hazard incentives by increasing the riskiness of their loan portfolio,
28
29 which in turn results in higher non-performing loans on average in the future. In this case, the
ev

30
31 connection with NPLs is negative (Berger and DeYoung 1997; Salas and Saurina 2002; Klein
32
iew

33 2013). On the other hand, according to (Quagliarello 2007), as the risk appetite of the bank is
34
35 higher, the greater is the share of capital to existing shareholders invested in the bank, in
36
37
order to convince other shareholders to invest and support the bank. And hence the
On

38
39
40 connection can be positive. Positive connection was discovered in the studies of (Rajan and
41
ly

42 Dahl 2003; Boudriga, Boulila, and Jellouli 2009; Espinoza and Prasad 2010). According to
43
44 empirical research and theory, with these determinants we expected an ambiguous correlation
45
46 with NPLs .
47
48 A number of authors consider influence of banks’ past performance measured by
49
50
profitability (ROA –ROE ) on future problem loans ratio. It is expected banks which are more
51
52
53 profitable to have lower level of NPLs (Swamy 2012) and hence the connection is negative.
54
55 According to Boudriga, Boulila and Jellouli 2009), inefficient banks with lower profitability
56
57
58
10
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 13 of 28 Eastern European Economics

1
2
3 are tempted to resort to less reliable and risky placements to increase profitability and/or to
4
5 meet the demands of regulatory authorities. The negative correlation between bank
6
7 performance (profitability) and credit risk is confirmed by (Godlewski 2004). In this area
8
9 again we will return to (Berger and DeYoung 1997), who explain the second hypothesis of
10
11
12
“bad management” by return on assets. Namely, poor performance of the company can be
13
14 linked with characteristics of managers that result in decreased profitability (expressed by the
15
16 low return on assets or equity). This further motivates managers to lend to riskier borrowers,
Fo
17
18 which in the end leads to growth of NPLs . Apart from these factors, we will follow (Makri,
19
rP
20 Tsagkanos, and Bellas 2014) and we will also examine the two profitability ratios (ROA and
21
22 ROE). Banks’ profitability is linked to the risk-taking behaviour of banks. As highly
23
ee

24
25
profitable banks have fewer incentives to engage in high-risk activities, ROA and ROE are
26
rR

27 expected to display a negative sign.

28
29 The credit policy of the bank plays an essential role in determining the subsequent
ev

30
31 levels of NPLs. To maximize the short run benefits, managers seek to rapidly expand credit
32
iew

33 activities and may hence take inadequate credit exposures (Castro 2012: Beck, Jakubik, and
34
35 Piloiu 2013; Klein 2013). Several studies indicate the presence of positive correlation
36
37
between credit growth and NPLs such as the study of (Dash and Kabra 2010). However,
On

38
39
40 there are studies such as (Salas and Saurina 2002; Quagliarello 2007; Boudriga, Boulila, and
41
ly

42 Jellouli 2009; Dash and Kabra 2010; Swamy 2012) which found a negative correlation
43
44 between these two determinants, which may be the result of some specificity, regulation and
45
46 background in different banking systems that make banks more conservative and cautious in
47
48 the spread of credit supply (Quagliarello 2007). Therefore the effect of individual credit
49
50
growth can be in both directions.
51
52
53 Apart from the actual determinants in the empirical model, we included two dummy
54
55 variables. Thereby, DUM2008/2009 marked the global economic crisis that had the value of 1
56
57
58
11
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 14 of 28

1
2
3 for the period from 2008 to 2009 and 0 for all other periods. Because to the consequential
4
5 deterioration of economic activity, borrowers had more difficulties paying off their debts,
6
7 therefore increasing the rate of NPLs; hence, we expected a positive and significant sign for
8
9 the coefficient on these dummies.
10
11
12
One of the assumptions of the linear regression model is that there is no
13
14 multicollinearity among the explanatorydeterminants. multicollinearity is a problem when the
15
16 correlation is above 0.80 (Kennedy 2008). The correlation among the five variables is
Fo
17
18 broadly in line with economic theory: The highest correlation coefficient was between ROA
19
rP
20 and ROE, which is both logical and expected because net income is a component of both the
21
22 ROA and the ROE of the banks. Furthemore, NPLs were negatively correlated with GDP
23
ee

24
25
growth and inflation, and positively correlated with the change of unemployment, and public
26
rR

27 debth
28
29 The correlation matrix shows that, in our sample multicollinearity problems were either
ev

30
31 not severe or non-existent.
32
iew

33 Insert Table 1 here

34
35
36
37
2.2 Methodology
On

38
39
40
41
ly

42 In order to analyse the determinants that affected the NPLs in the Baltic States we
43
44 adapted panel data analysis, using NPL and UN in logarithmic differences. Hsiao (2003) list
45
46 several benefits of using panel data 1. The use of panel data enables us to control for
47
48 individual heterogeneity. 2. Panels provides more informative data, more variability, less
49
50
collinearity among the variables, greater degree of freedom, and more efficiency. 3. With
51
52
53 panel data, one is better able to study the dynamics of adjustment. 4. Panel data are more
54
55 suitable for identifying and measuring effects that are simply not detectable in pure cross-
56
57
58
12
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 15 of 28 Eastern European Economics

1
2
3 sections or pure time-series data. 5. Panel data models allow us to construct and test more
4
5 complicated behavioural models than pure cross-section or time data models.
6
7 According to (Rinaldi and Sanchis-Arellano 2006; Louzis and Metaxas 2012)
8
9 empirical evidence suggests that the NPL ratio may follow a unit root process hinting at a
10
11
12
possible cointegrating relation with macroeconomic variables To avoid the problem which
13
14 may arise because of existence of non stationary variables, one might have to identify the
15
16 order of integration of variables. (Campbell and Perron 1991). suggest that standard unit root
Fo
17
18 tests can have low power against stationary alternatives for important cases, Bearing this in
19
rP
20 mind in this article, we tested for stationarity of the panel, using Maddala and Wu Fisher tests
21
22 for unbalanced panels.
23
ee

24
25
In the literature which investigates the determinants of NPLs in the Baltic States, the
26
rR

27 authors usually applied ordinary least squares methods or a fixed- effects model (Festic and
28
29 Repina 2009; Kavkler and Festic 2010; Tanasković and Jandrić 2015). Also, the difference
ev

30
31 Generalized Method of Moments was applied by Kjosevski and Petkovski (2017). In order to
32
iew

33 provide consistent and unbiased results, we implemented three alternative estimation

34
35 techniques. The first one is a fixed- effects model, which allows controlling for unobserved
36
37
heterogeneity across banks. Although this approach is rather simple and intuitive, it may give
On

38
39
40 rise to “dynamic panel bias”, which results from the possible endogeneity of the lagged
41
ly

42 variable and the fixed effects in the error term. To provide consistent and unbiased results,
43
44 we implemented the difference Generalized Method of Moments (difference GMM)
45
46 estimation, which is based on first differences and was introduced by (Arellano and Bond
47
48 1991). Arellano and Bond proposed one and two-step estimators. In this article, we use the
49
50
one-step GMM estimator since Monte Carlo studies have found that this estimator
51
52
53 outperforms the two-step estimator both in terms of producing a smaller bias and a smaller
54
55 standard deviation (Ruth and Owen 1999). Thus, following (Louzis, and Metaxas 2012) we
56
57
58
13
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 16 of 28

1
2
3 instrumented macroeconomic variables by themselves, whereas the bank-specific
4
5 determinants were instrumented with current and lagged values of the regressors. In order to
6
7 avoid the problem of too many instruments in comparison to the number of groups (Roodman
8
9 2009), the number of instruments are kept lower than the number of banks. In the standard
10
11
12
(un-collapsed) form, each instrumenting variable creates one instrument for each time period
13
14 and the lag available to that period, whereas in the collapsed form not a whole matrix of
15
16 instruments, but a single column vector of instruments is created. Although collapsing can
Fo
17
18 reduce statistical efficiency in large samples, it can be very helpful as a tool to avoid the bias
19
rP
20 in finite samples, which are usually characterized by instrument proliferation. In other words,
21
22 we control the number of instruments by limiting our analysis to 1 lag. This helps to avoid
23
ee

24
25
bias due to too many instruments in a relatively small sample. The validity of chosen
26
rR

27 instruments for parameters estimation can be tested using the Hansen test. Accepting the null
28
29 hypothesis means that the chosen instruments are valid. One drawback of this approach,
ev

30
31 however, is that in samples with a limited time dimension (small T) and high persistence, the
32
iew

33 estimation has low precision (Blundell and Bond 1998). Namely, according to (Blundell and
34
35 Bond 1998) if the lagged dependent variables are persistent during the time or tend to be
36
37
random walk, lagged levels of these variables will be weak instruments in the first difference
On

38
39
40 equation of regression. Therefore, we also estimate a “system GMM” developed by (Arellano
41
ly

42 and Bover 1995) and (Blundell and Bond 1998), which addresses this concern. Under this
43
44 approach, the lagged bank level variables were modeled as pre-determined (thus instrumented
45
46 GMM-style in the same way as the lagged dependent variable) whereas the macro variables
47
48 were treated as strictly exogenous (instrumented by themselves as “IV style” instrument,
49
50
(Roodman 2009).
51
52
53 Following (Salas and Saurina 2002; Merkl and Stolz 2009; Louzis and Metaxas 2012:
54
55 Klein 2013; Abid, Med, and Zouari-Ghorbel 2013), we assume that the share of NPLs in the
56
57
58
14
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 17 of 28 Eastern European Economics

1
2
3 loan portfolio is closely related to its values in the previous periods, because NPLs cannot be
4
5 immediately written off and may remain on banks' balance sheets up to several years. In
6
7 other words, NPL ratio shows a tendency to persist over time. To test the persistence of
8
9 NPLs , we use the previous year’s NPLs rate (NPLt-1) as an independent variable and we
10
11
12
expect a positive correlation. The inclusion of lagged terms of the dependent variable on the
13
14 right hand side of the equation violates the exogeneity assumption for regressors. These
15
16 dynamic relations are given by the following equation:
Fo
17
18
19
rP
20
21 yit = α i yi ,t −1 + α i Bi ,t + α i M i ,t + DUM 2008 / 2009 + ε it (1)
22
23
ee

24
25
26 where yit denotes the aggregate NPLs to total gross loans, B denotes the bank-specific
rR

27
28 variables and M denotes the macroeconomic factors, DUM is a dummy variable. Note that i
29
ev

30 corresponds to the examined bank of the sample and t to the year, whereas ε denotes the
31
32 error term.
iew

33
34
Furthermore, following (Beck, Jakubik, and Piloiu 2013; Makri, Tsagkanos, and
35
36
37 Bellas 2014) we consider the macroeconomic variables as strictly exogenous, we treat the
On

38
39 bank-specific regressors as weakly exogenous and we use one lag for both bank-specific and
40
41 macroeconomic regressors, targeting to capture the dynamics of explanatory variables over
ly

42
43 the previous year. Therefore, our next econometric model is expressed as follows:
44
45
46
47
48 yit = α i yi ,t −1 + α i Bi ,t −1 + α i M i ,t −1 + DUM 2008 + DUM 2009 + ε it −1
49
50
51 (2)
52
53
54
55
56
57
58
15
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 18 of 28

1
2
3 To test the validity of chosen instruments we will use the Hansen test. Accepting the
4
5 null hypothesis means that the chosen instruments are valid. Furthemore, we will test serial
6
7 correlations in the differenced residuals (first-order [AR1]) and second-order [AR2] serial
8
9 correlations). According to (Arellano and Bond 1991) the first-order autocorrelation in the
10
11
12
differenced residuals does not imply that the estimates are inconsistent. However, the
13
14 second-order autocorrelation would imply that the estimates are inconsistent. We also report
15
16 Wald tests of the joint significance of both the coefficients and the dummies, which validates
Fo
17
18 the use of such determinants in our equation.
19
rP
20 To obtain deeper insight into the relevance of explanatory variables, we estimate
21
22 Equation (2) in three different versions; we begin by examining only macro determinants as
23
ee

24
25
regressors (model 1), then only bank-specific determinants (model 2), and finally both bank-
26
rR

27 specific and macro determinants (model 3).

28
29
ev

30
31 3. Empirical Results
32
iew

33 In this section, we begin with analysis of the results of the panel unit root tests. The
34
35 results of this test are presented in Table 3. The unit root analysis, according to ADF and PP
36
37
Fisher-type tests, indicates that null hypothesis of non-stationarity can be rejected for all our
On

38
39
40 determinants. The results of Breitung test indicates that hypothesis of non-stationarity cannot
41
ly

42 be rejected for the three determinants (NPL, DCPS and ETA). However, bearing in mind that
43
44 the other two unit tests (ADF and PP Fisher-type) show that these determinants were
45
46 stationary at their levels, we include NPL, DCPS and ETA in our models, and we treat them
47
48 as non-stationary variables at their levels.
49
50
51
52
53 Insert Table 2 here
54
55
56
57
58
16
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 19 of 28 Eastern European Economics

1
2
3 Next, in Table 3 we report the empirical results of fixed effects model, difference
4
5 GMM and system GMM. Despite their difference, all approaches arrive at essentially similar
6
7 results as to the sign, and the statistical significance of most variables in the regression
8
9 specification. This confirms that our results are robust to different specifications, although the
10
11
12
precision of the estimated coefficients differs across different methods that we have used in
13
14 our study.
15
16
Fo
17
18 Insert Table 3 here
19
rP
20
21
22
23
ee

24
25
The results presented in Table 3 broadly confirm that both bank-level and
26
rR

27 macroeconomic factors play a role in affecting the banks’ asset quality. The models seem to
28
29 fit the panel data reasonably well, having fairly stable coefficients,whereas the Wald test
ev

30
31 indicates a fine goodness of fit. The Hansen test shows that the chosen instruments are valid
32
iew

33 (with a p-value of 0.48 for model 1, 0.52 for model 2, and 0.46 for the third model) in the
34
35 difference GMM and (with p-value of 0.54 for model 1, 0.62 for model 2, and 0.72 for the
36
37
third model) in the system GMM. The estimator ensures efficiency and consistency provided
On

38
39
40 that the residuals do not show serial correlation of order two (even though the equations
41
ly

42 indicate that negative first order autocorrelation is present, this does not imply that the
43
44 estimates are inconsistent). Inconsistency would be implied if second-order autocorrelation
45
46 was present (Arellano and Bond 1991), but this case is rejected by the test for AR(2) errors.
47
48 The lagged dependent variable is statistically significant and have positive value in all
49
50
three models, confirming the dynamic character of the models. The values of this determinant
51
52
53 are between 0.51 and 0.87, suggesting that a shock to NPLs is likely to have a prolonged
54
55 effect on the banking system. These results are similar to those found by previous studies
56
57
58
17
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 20 of 28

1
2
3 (Jimenez and Saurina 2005) for Spain where lagged NPLs value was 0.55 and (Kjosevski and
4
5 Petkovski 2017) for Baltic States where values of lagged NPLs were between 0.33 and 0.49.
6
7 From the macroeconomic determinants we found evidence that public debt was
8
9 significant in both models when we used fixed effect model and difference GMM. The
10
11
12
positive signs are in line with the literature. This relationship highlights that higher public
13
14 debt in Baltic States might lead to an important rise of NPLs. Our results are consistent with
15
16 the findings of (Makri, Tsagkanos, and Bellas 2014) where the results were between 0.11 and
Fo
17
18 0.12 for 14 countries from Eurozone.
19
rP
20 The positive effect of unemployment emphasized in the literature is confirmed by the
21
22 results of this study. Obviously, when a person loses his source of income he cannot return
23
ee

24
25
his loan, or regarding to enterprises, the rise of unemployment could lead to a decline in
26
rR

27 production due to the fall in effective demand. In both cases that will lead to higher NPLs.
28
29 Because we used annual data, the significant impact of unemployment NPLs was in the
ev

30
31 current period. Namely, a rise of unemployment affects households’ ability to service their
32
iew

33 debts, and firms cut their labor costs with a three-month time delay (Louzis, Vouldis and
34
35 Metaxas 2010). Our results are consistent with the findings of (Nkusu 2011) where the
36
37
results were between 0.20 and 0.24 ) for 26 advanced economies and (Makri, Tsagkanos, and
On

38
39
40 Bellas 2014) between 0.23 and 0.09, for 14 countries from Eurozone.
41
ly

42 The results of our article show that growth of gross domestic product has a
43
44 significant and negative impact on NPLs. They confirm that change in economic activity
45
46 affects the NPLs. The values were between -0.16 and – 1.58 and were significant in all three
47
48 estimation models. These results are consistent with the results of (Louzis, Vouldis, and
49
50
Metaxas 2010) where values of GDP growth were between 0.25 and 0.46 for Greece, (Nkusu
51
52
53 2011), (0.18-0.04) for 26 advanced economies. (Klein 2013) (0.05-0.08) for 16 countries
54
55 from Central and Eastern Europe, (Beck, Jakubik, and Piloiu 2013), (1.50-1.52), for 75
56
57
58
18
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 21 of 28 Eastern European Economics

1
2
3 countries of the world, (Makri, Tsagkanos, and Bellas 2014), (0.053-0.071) for 14 countries
4
5 from Eurozone.
6
7 As for inflation, our results show a positive impact of inflation indicates that higher
8
9 inflation probably was anticipated by the banks management, which in turn implies that
10
11
12
interest rates have been appropriately adjusted. Actually, inflation reduces the capacity of
13
14 banks’ borrowers to repay the loans through the income channel or due to the falling value of
15
16 income amid rising inflation. The positive results were also found in the studies of (Rinaldi
Fo
17
18 and Sanchis-Arellano 2006) with the value 0.23 for seven countries of the euro area, (Klein
19
rP
20 2013) with the values for inflation between 0.006 and 0.38, for 16 countries from Central and
21
22 Eastern Europe and (Makri, Tsagkanos, and Bellas 2014) with values of inflation between
23
ee

24
25
0.039 and 0.045 for 14 countries from Eurozone.
26
rR

27 The results from the bank-specific determinants are in line with literature. The
28
29 empirical results of return on assets (ROA) and return on equity (ROE) show that profitability
ev

30
31 has a significant negative impact on NPLs.This relationship, is consistent with the empirical
32
iew

33 results from (Makri, Tsagkanos, and Bellas 2014) with values between -0.052 and -0.038, for
34
35 14 countries from Eurozone. The results for ROA were confirmed in the study of (Erdinc and
36
37
Abazi 2014) with values between -0.34 and -0.55, for 20 emerging European countries and
On

38
39
40 (Makri, Tsagkanos, and Bellas 2014) with values between -0.62 and -0.38, for 14 countries
41
ly

42 from Eurozone. These results demonstrate the validity of the hypothesis of “bad
43
44 management”, reflected in the reduced profitability, which in turn motivates managers to go
45
46 for an increased risk exposure, therefore creating growth of bad loans.
47
48 As we expected, risk behavior of banks- ETA, is statistically significant, but only in
49
50
the two models (fixed and difference GMM) and has a negative sign. This relationship is
51
52
53 also confirmed by (Klein 2013) with results between -0.04 and -0.06 for 16 countries from
54
55 Central and Eastern Europe and (Makri, Tsagkanos, and Bellas 2014) where the results were
56
57
58
19
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 22 of 28

1
2
3 between -0.01 and -0.13, for 14 countries from Eurozone.
4
5 As for the dummy variables, which are introduced to cover the global economic
6
7 crisis,we found that both DUM 2008 and DUM 2009 are statistically significant at the 1%
8
9 level of significance, in all three models. These results confirmed that the rise of the global
10
11
12
economic crisis led to a deterioration in the quality of bank loans to enterprises and
13
14 households. This was not surprising, because the banks in Baltic States were among those
15
16 that were most affected by the crisis and as a result they had the largest increase of NPLs.
Fo
17
18 4. Conclusions
19
rP
20 Using three alternative estimation techniques, fixed effects model, difference
21
22 Generalized Method of Moments and system Generalized Method of Moments, with data
23
ee

24
25
ranging from 2005 to 2016, in this article, we have analyzed the macroeconomic and bank-
26
rR

27 specific determinants of NPLs for a panel of 21 banks from the three Baltic States. Our
28
29 findings are largely consistent with the literature. The results provide evidence that from
ev

30
31 among the macroeconomic determinants in our baseline model, public debt, growth of GDP,
32
iew

33 inflation and unemployment have the strongest effect on NPLs. Furthermore, we have found
34
35 that ETA, ROA, ROE and GGL as bank - specific determinants have an influence on NPLs.
36
37
Also, we found that the dummy variables that we introduced to cover the global economic
On

38
39
40 crisis, DUM 2008 and DUM 2009, have the positive impact on the growth of NPLs of Baltic
41
ly

42 States.
43
44 The paper’s findings offer several policy implications. First, the regulatory authorities
45
46 could use the results of this study to detect banks with potential for sharp build up of NPLs
47
48 in the future. Second, regulators should place greater emphasis on risk management systems
49
50
and procedures followed by banks, to avert future financial instability. Third, they need to
51
52
53 streamline banks to better manage risk, taking into account individual characteristics of
54
55 individual banks. A better understanding of the individual factors that make some banks more
56
57
58
20
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 23 of 28 Eastern European Economics

1
2
3 resilient than the others to adverse economic trends can prevent a rise of credit risk and thus
4
5 reduce negative feedback between the financial sector and the real economy.
6
7 The future research may broaden the scope of examination. First, there is a lack of
8
9 available data on selected determinants for a longer period. The existence of long time series
10
11
12
of data would enable obtaining more accurate and more reliable results. Second, the future
13
14 research could be made by taking into account the situation in some other Central and Eastern
15
16 European countries. Third, .in this paper the distribution of loans between household and
Fo
17
18 enterprise loans is not taken into consideration Finally, the research may be improved by
19
rP
20 including other macroeconomic determinants (monetary aggregates, stock prices and
21
22 exchange rate) or bank-specific factors (size, loans-to-assets ratio, and so on .).
23
ee

24
25
References
26
rR

27
28
29
ev

30 Abid, L.; Ouertani, N., and Zouari-Ghorbel, S. 2013. “Macroeconomic and Bank-Specific
31
32 Determinants of Household’s Non-Performing Loans in Tunisia: a Dynamic Panel Data.”1st
iew

33
34 TSFS Finance Conference, TSFS 2013, 12-14 December 2013, Sousse, Tunisia
35
36 Arellano, M., and Bond, S. 1991. “Some tests of specification for panel data. Monte Carlo
37
On

38
evidence and an application to employment equations.” Review of Economic Studies 58: 277-
39
40
41 297.
ly

42
43 Aver, B. 2008. “An empirical analysis of credit risk factors of the Slovenian banking
44
45 system.” Managing Global Transitions 6 (3): 317-334.
46
47 Babouček, I., and Jančar, M. 2005. “Effects of Macroeconomic Shock to the Quality of the
48
49 Aggregate Loan Portfolio.” Working Paper Series, No. 1: 1 – 62.
50
51
Baltagi, B. 2001. Econometric Analysis of Panel Data (second edition). John Wiley & Sons,
52
53
54
Chichester.
55
56
57
58
21
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 24 of 28

1
2
3 Beck, R.; Jakubik, P., and Piloiu, A. 2013. “Non-performing loans What matters in addition
4
5 to the economic cycle?”Working Paper Series, No 1515.
6
7 Berger, A., and DeYoung, R. 1997. “Problem Loans and Cost Efficiency in Commercial
8
9 Banks.” Journal of Banking and Finance, 21: 849–870.
10
11
12
Blaschke, W., and Jones, M. 2001. “Stress testing of financial systems: An overview of
13
14 issues, methodologies and FSAP experiences. ” IMF Working Paper, 01/88.
15
16 Blundell, R., and Bond, S. 1998. “Initial conditions and moment restrictions in dynamic panel
Fo
17
18 data models.”Journal of Econometrics, 87(1): 115-143.
19
rP
20 Bofondi, M., and Ropele, T. 2011. “Macroeconomic determinants of bad loans: evidence
21
22 from Italian banks.” Occasional Paper (89).
23
ee

24
25
Boudriga, A.; Boulila, N., and Jellouli, S. 2009. “Does bank supervision impact
26
rR

27 nonperforming loans: cross-country determinants using aggregate data? ” MPRA Paper No.
28
29 18068.
ev

30
31 Castro, V. 2012. “Macroeconomic determinants of the credit risk in the banking system: the
32
iew

33 case of the GIPSI.” NIPE WP 11/ 2012

34
35 Campbell, J., and Perron, P. 1991. “Pitfalls and opportunities: what macroeconomists should
36
37
know about unit roots.” Macroeconomics Annual, National Bureau of Economic
On

38
39
40 Research:141-201.
41
ly

42 Dash, M., and Gaurav, K. 2010. “The determinants of non-performing assets in Indian
43
44 commercial bank: An econometric study. ” Middle Eastern Finance and Economics, 7: 94-
45
46 106.
47
48 De Bock, R., and Demyanets, A. 2012. “Bank Asset Quality in Emerging Markets:
49
50
Determinants and Spillovers.” IMF Working Paper 12/71.
51
52
53
54
55
56
57
58
22
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 25 of 28 Eastern European Economics

1
2
3 De Nicoló, G.; Batholomew, P.; Zaman, J., and Mary, Z. 2003. “Bank Consolidation,
4
5 Conglomeration and Internationalization: Trends and Implications for Financial Risk.” IMF
6
7 Working Paper #03/158.
8
9 Demirgüç-Kunt, A., and Enrica, D. 1998. “The determinants of banking crises in developing
10
11
12
and developed countries.” IMF Staff Papers, 45 (1): 81-109.
13
14 Donath, L.; Cerna, V., and Oprea, I. 2014. “Macroeconomic Determinants of Bad Loans in
15
16 Baltic Countries and Romania.”SEA - Practical Application of Science vol. II, Issue 4 (6.)
Fo
17
18 European Banking Coordination Vienna Initiative. (2012). Report of the working group on
19
rP
20 NPLs in Central, Eastern and Southeastern Europe.
21
22 Erdinc, D., and Abazi, E. 2014. “The Determinants of NPLs in Emerging Europe, 2000-
23
ee

24
25
2011.” Journal of Economics and Political Economy Vol 1 Issue 2.
26
rR

27 Espinoza, R., and Prasad, A. 2010. “Nonperforming Loans in the GCC Banking Systems and
28
29 their Macroeconomic Effects. ” IMF Working Paper 10/224.
ev

30
31 Fainstein, G., and Novikov, I. 2011. “The Comparative Analysis of Credit Risk Determinants
32
iew

33 In the Banking Sector of the Baltic States.” Review of Economics & Finance, vol. 1: 20-45.
34
35 Festic, M., and Repina, S. 2009. “Financial Stability in the Baltics.” Czech Journal of
36
37
Economics and Finance, 59, no. 6.
On

38
39
40 Fofack, H. 2005. “Non-Performing Loans in Sub-Saharan Africa: Causal Analysis and
41
ly

42 Macroeconomic Implications.” Working Paper No. WP 3769.

43
44 Hoggarth, G.; Logan, A., and Zicchino, L. 2005. “Macro stress tests of UK banks.” BIS
45
46 papers, 22: 392-408.
47
48 Hsiao, C. 2003. “Analysis of Panel DataCambridge University Press
49
50
Gasha, G., and Morales, A. 2004. “Identifying Threshold Effects in Credit Risk Stress
51
52
53 Testing.” IMF Working Paper, 150.
54
55 Godlewski., C. 2004. “Capital Regulation and Credit Risk Taking : Empirical Evidence from
56
57
58
23
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 26 of 28

1
2
3 Banks in Emerging Market Economies.” Institut d'Etudes Politiques, Université Robert
4
5 Schuman, Strasbourg 3.
6
7 Gonzalez Hermosillo, B. 1999. “Determinants of ex-ante banking system distress: A micro-
8
9 micro empirical exploration of some recent episodes.” IMF working paper, WP/99/33.
10
11
12
Gunsel, N. 2008. “Micro and Macro determinants of bank fragility in North Cyprus
13
14 economy.” International Research Journal of Finance and Economics ISSN 1450- 2887
15
16 Jakubík, P., and Reininger, T. 2013. “Determinants of Nonperforming Loans in Central,
Fo
17
18 Eastern and Southeastern Europe.” Focus on European Economic Integration, issue 3: 48-66.
19
rP
20 Kavkler, A., and Festic, M. 2010. “The banking sector in the Baltics”, Banks and Bank
21
22 Systems, Vol. 5.
23
ee

24
25
Keeton, W, and Morris, C. 1987. “Why Do Banks’ Loan Losses Differ?”, Economic
26
rR

27 Review:3–21.
28
29 Kennedy, P. 2008. “A Guide to Econometric 6th edition”. Malden, Massachusetts: Blackwell
ev

30
31 Publishing.
32
iew

33 Kjosevski, J., and Petkovski, M. 2017. Non-performing loans in Baltic States: determinants
34
35 and macroeconomic effects, Baltic Journal of Economics, 17:1, 25-44,
36
37
Klein, N. 2013. “Non-Performing Loans in CESEE; Determinants and Impact on
On

38
39
40 Macroeconomic Performance”, IMF Working Papers 13/72
41
ly

42 Koivu, T. 2002. “Banking and finance in the Baltic countries”, BOFIT, no. 11.
43
44 Jappelli, T., Pagano, M., and di Maggio, M 2008. “Households’ Indebtedness and Financial
45
46 Fragility.” CSEF Working Paper 208.
47
48 Jimenez, G., and Saurina, J. 2005. “Credit cycles, credit risk, and prudential regulation.”
49
50
Banco de Espana.
51
52
53 Judson, R., and Owen, A. 1999. “Estimating dynamic panel data models: A guide for
54
55 macroeconomics”. Economic Letters 65: 9-15.
56
57
58
24
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
Page 27 of 28 Eastern European Economics

1
2
3 Louzis, D.; Vouldis, A., and Metaxas, V. 2010. “Macroeconomic and bank-specific
4
5 determinants of non performing loans in Greece: a comparative study of mortgage, business
6
7 and consumer loan portfolios,” Working Paper, no. 118.
8
9 Louzis, D., and Metaxas, V 2012. “Macroeconomic and Bank-Specific Determinants of Non-
10
11
12
performing Loans in Greece: A Comparative Study of Mortgage, Business and Consumer
13
14 Loan Portfolios.” Journal of Banking & Finance, 36 (4):1012–1027.
15
16 Maddala, G.S. 2001. Introduction to Econometrics. 3rd edition, Chichester, John Wiley and
Fo
17
18 Sons.
19
rP
20 Makri, V.; Tsagkanos, A., and Bellas, A/ 2014. “Determinants of Non-Performing Loans:
21
22 The Case of Eurozone.” Panoeconomicus: 93-206
23
ee

24
25
Marcucci, J., and Quagliariello, M. 2008. “Credit risk and business cycle over different
26
rR

27 regimes”, Temi di discussione no.670.

28
29 Marcucci, J., and Quagliariello, M. 2009. “Asymmetric effects of the business cycle on bank
ev

30
31 credit risk.” Journal of Banking & Finance, vol. 33, 9:1624-1635.
32
iew

33 Merkl, C., and Stolz, S. 2009. “Banks Regulatory Bbuffers, Liquidity Networks and
34
35 Monetary Policy.” Transmission Applied Economics, 41: 2013–2024
36
37
Mörttinen, L.; Poloni, P., and Vesala, S. 2005. “Analysing banking sector conditions: how to
On

38
39
40 use macro-prudential indicators.” ECB Occasional Paper, no. 26
41
ly

42 Nkusu, M. 2011. “Nonperforming Loans and Macrofinancial Vulnerabilities in Advanced

43
44 Economies.” IMF Working Paper 11/161.
45
46 Roodman, D. 2009. “A Note on the Theme of Too Many Instruments.” Oxford Bulletin of
47
48 Economics and Statistics, 71: 135–158..
49
50
Quagliariello, M. 2003. “Are Macroeconomic Indicators Useful in Predicting Bank Loan
51
52
53 Quality? ” Evidence from Italy. Bank of Italy, Paper Series, (1)
54
55
56
57
58
25
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]
 Eastern European Economics Page 28 of 28

1
2
3 Quagliarello, M. 2007. “Banks’ Riskiness Over the Business Cycle: a Panel Analysis on
4
5 Italian Intermediaries.” Applied Financial Economics, 17(2): 119-138.
6
7 Rajan, R., and Dahl, C. 2003. “Non-performing Loans and Terms of Credit of Public Sector
8
9 Banks in India: An Empirical Assessment.” Occasional Papers, Vol. 24: 81–121.
10
11
12
Rinaldi, L., and Sanchis-Arellano, A. 2006. “Household Debt Sustainability, what Explains
13
14 Household Non-Performing Loans? An Empirical Analysis.” European Central Bank
15
16 Working Paper Series 570.
Fo
17
18 Salas V., and Jesús, S. 2002. “Credit Risk in Two Institutional Regimes: Spanish Commercial
19
rP
20 and Savings Banks.” Journal of Financial Services Research, vol. 22: 203–224.
21
22 Swamy, V. 2012. “Impact of Macroeconomic and Endogenous Factors on Non- Performing
23
ee

24
25
Banks Assets.” International Journal of Banking and Finance, Vol 9, Issue 1, Article 2
26
rR

27 Tang, H.; Zoli, E., and Klytchnikova, I. 2000. “Banking crises in transition economies: fiscal
28
29 costs and related issues.” The World Bank Policy Research Working Paper, no. 2484.
ev

30
31 Tanasković, S., and Jandrić, M. 2015. “Macroeconomic and Institutional Determinants of
32
iew

33 Non-performing Loans.” Journal of Central Banking Theory and Practice, 1: 47-62.

34
35 Vogiazas, S., and Nikolaidu, E. 2011. “Investigating the Determinants of Nonperforming
36
37
Loans in the Romanian Banking System: An Empirical Study with Reference to the Greek
On

38
39
40 Crisis.”Economics Research International. vol. 2011.
41
ly

42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
26
59
60 URL: https://mc.manuscriptcentral.com/meee E-mail: [email protected]