You are on page 1of 16

Journal of International Money and Finance 32 (2013) 1–16

Contents lists available at SciVerse ScienceDirect

Journal of International Money
and Finance
journal homepage: www.elsevier.com/locate/jimf

The impact of banking sector stability on the real economyq
Terhi Jokipii a, Pierre Monnin b, *
a
b

Swiss National Bank, Financial Stability, Bundesplatz 1, 3001 Bern, Switzerland
Swiss National Bank, Financial Market Analysis, Boersentrasse 15, 8022 Zurich, Switzerland

a b s t r a c t
JEL classification:
E20
E44
G21
Keywords:
Banking sector stability
Real output growth
Output growth forecasts

This article studies the relationship between the degree of banking
sector stability and the subsequent evolution of real output growth
and inflation. Adopting a panel VAR methodology for a sample of
18 OECD countries, we find a positive link between banking sector
stability and real output growth. This finding is predominantly
driven by periods of instability rather than by very stable periods.
In addition, we show that an unstable banking sector increases
uncertainty about future output growth. No clear link between
banking sector stability and inflation seems to exist. We then argue
that the link between banking stability and real output growth can
be used to improve output growth forecasts. Using Fed forecast
errors, we show that banking sector stability (instability) results in
a significant underestimation (overestimation) of GDP growth in
the subsequent quarters.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction
The increased incidence of banking and financial crises over the last quarter century has triggered
an active research agenda, not only on the underlying causes of crises, but also on their impact on the
real economy. Literature in this field has mainly followed two directions: the first set of authors have
focussed on understanding the links between banking sector characteristics and long-term growth
while a second set of authors have been concerned with quantifying the costs of banking sector crises
in terms of real output losses. Within the first strand of literature, Levine (1997, 2001) assesses the link
q The authors thank Signe Krogstrup for collecting data on inflation and growth forecasts. We also thank an anonymous
referee and the members of the SNB Financial Stability group for their useful comments. The opinions expressed herein are
those of the authors and do not necessarily reflect the views of the Swiss National Bank.
* Corresponding author. Tel.: þ41 44 631 36 69; fax: þ41 44 631 81 39.
E-mail addresses: terhi.jokipii@snb.ch (T. Jokipii), pierre.monnin@snb.ch (P. Monnin).
0261-5606/$ – see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jimonfin.2012.02.008

the extent to which macroeconomic policies and banking system soundness depend on one another (Benink and Benston. we show that banking sector instability is followed by higher uncertainty about output growth. However.4 The novelty of our set-up is that it allows the variance of output growth to be dependant of the state of the banking sector. and are interested in shedding light on the impact of banking sector instability on real output growth. Periods of bank sector stability are generally followed by an increase in real output growth and vice versa. consider a banking sector which suffers credit losses in a business cycle downturn but which is still able to function without external help. Gupta..e. The focus on crises vs. 1986) provide further evidence that the degree of development in the financial sector acts as an important contributor to economic growth. In addition. 2006. It remains unclear whether ‘normal’ reductions in banking sector stability – i. without examining its causes. 2005. 3 For example.. 1997. The link between banking sector stability and economic activity is of particular interest to policy makers which base their monetary policy decisions on economic forecasts. we contribute to the second strand of literature discussed above by empirically assessing whether the degree of banking sector stability has an impact on the real economy in normal times. These empirical conclusions have prompted an increased interest by both policy makers and academics to assess.. Monnin / Journal of International Money and Finance 32 (2013) 1–16 between the degree of banking sector openness and economic growth. the choice of threshold for defining a crisis is highly discretionary. 2005. however it is a single country study that does not make use of the panel VAR methodology adopted here. we explore the impact of banking sector stability on real output growth and inflation in the subsequent quarters with a panel vector autoregressive model. 1997. 2 . We take banking sector instability as given. 2005). ignoring the banking sector’s impact during more stable times.2 T. 2007. Deutsche Bundesbank. This measure has the advantage of being continuous by nature and thus captures a continuum of states. it is not in a fully fledged crisis either. but that does not translate into a banking crisis – have a significant impact on growth. Goodhart et al. cited above. reaches a clear conclusion: banking crises have usually coincided with. by improving the functioning of domestic financial markets and banks. 4 Allenspach and Perrez (2008) study interactions between the banking sector and the real economy for Switzerland via a VAR approach. Kroszner et al. European Central Bank. rather than classification as either crisis or non-crisis. bank crises are relatively rare events. 1998. Demirgüç-Kunt and Detragiache. Following Aspachs et al. This captures the impact of banking sector instability on output growth as well as on its uncertainty. Levine et al. (2007). He shows that international financial liberalization. that have worked toward quantifying the costs of banking sector crises in terms of real output losses have focussed solely on the loss in output growth during or after banking crises.b. Demirgüç-Kunt and Maksimovic. Several other authors (Levine.. Serwa. (2008) who show a positive link between average growth and banking sector instability. 2005.2 The main reason for this is that many authors have chosen to work with binary dependent variables (crisis vs. Our results can be summarized as follows: banking sector stability is an important driver of future GDP growth. as it is difficult to separate cause and effect in the financial sector real economy nexus (Kaminsky and Reinhart. 2002. One exception to this is Ranicère et al. Research focusing on understanding whether financial stability should be considered in the setting of monetary policy has 1 This literature is reviewed in Levine (2004). King and Levine. several drawbacks are associated with adopting this approach. 2000. focussed on quantifying costs of crises. but since it is still functioning. accelerates economic growth. Hillbers et al. The literature is however. 1999. Dell’ariccia et al. Rajan and Zingales. Boyd et al. 2006). 2006.. or preceded. 2005. This contrasts with the binary variable approach usually adopted in this context. Second. Jokipii. Several studies. a level of instability that can regularly be observed. from a theoretical point of view. P. non-crisis). 1993a. First..1 The second strand of literature.. non-crisis periods has consequently resulted in a lack of research assessing the impact of the banking sector on the real economy in less extreme times. We measure banking sector stability based on the banking sector’s probability of default. The stability of such a banking sector has clearly decreased after credit losses.3 In this paper. Finally. a substantial economic slowdown (see among others Hogart et al. far less clear regarding whether or not the banking sector is the main trigger of the economic slowdown. binary variables impose the unrealistic assumption that a banking sector that is not experiencing a crisis is necessarily healthy. 1998. a finding which appears to be driven predominantly by periods of relative instability rather than by periods of stability. 2007. 2008).

With these two distribution functions. the total assets At þ 1 of the banks are smaller than their total debt Dt þ 1). then the firm cannot repay its debt and it becomes insolvent. If the assets are greater than the debt.T. Peek et al. between t and t þ 1. We contribute to this literature by extending our analysis further and investigating the results uncovered thus far. and in particular whether bank related information can be used to improve macro forecasts. computed as the aggregate balance sheet of banking institutions within country i. Stability index for the banking sector 2. 2. the firm’s equity is equal to: maxðAtþ1  Dtþ1 . Section 5 explores the relationship between banking sector stability and central bank forecast errors. we are able to compute the probability that. We also assume that. the value of the firm’s assets is smaller than its debt. In this case. In particular. if at time t þ 1. t þ 1 in our case).e. at debt maturity. We assume that the asset value of the banking sector follows a geometric Brownian motion characterized by: dAt ¼ mAt dt þ st At dw (1) where m is the instantaneous growth rate of assets. The rest of this paper is organized as follows. of the aggregate banking sector in our case). Default probability and distance-to-default for the banking sector To estimate the default probability of a firm (i. Banking sector stability: definition and assumptions Banking sector instability is defined as the probability of the banking sector becoming insolvent within the next quarter. Using Fed’s forecasts data. 2003). which implies the following one-period transition function: Dtþ1 ¼ Dt er (2) 2. We consider a banking sector to be insolvent if. Section 2 defines our measure of banking sector stability. st is their instantaneous volatility rate and dw is a Wiener process. Jokipii. Section 6 concludes.2. at the end of the quarter. and that in fact. conditional on their initial values at time t.1. the asset value falls below the debt value. We continue by assessing whether additional information embedded in our stability index might help to improve output growth forecasts. To compute the default probability of the banking sector of country i at time t. the firm’s equity is worth zero. Merton’s method is based on the idea that if. at time tþ1. P. (1999.e. two elements are required: the distribution of both the asset value and the debt value at the end of the period. the market value of the assets owned by all the banks of a country is not sufficient to repay its total debt (i. Merton (1974) suggests considering the firm’s equity as a call option on its assets. 0Þ (3) . then the firm repays its debt and the market value of the firm is equal to the difference between the asset value and the debt. at debt maturity (i. Section 4 estimates the impact of banking sector stability on output and inflation. our results indicated an apparent importance of banking sector stability on output growth. This approach makes the implicit assumption that the banking sector of country i is equal to a single bank.e. Thus Merton’s model states that. debt grows at the (continuously compounded) growth rate r. we show that banking sector stability (instability) results in a significant underestimation (overestimation) of GDP growth in the subsequent quarters. the Federal Open Market Committee (FOMC) consider this information when setting monetary policy. This finding indicates that additional information embedded in our stability measure has the potential to further improve economic forecasts. Our measure therefore abstracts from any competitiveness considerations. Monnin / Journal of International Money and Finance 32 (2013) 1–16 3 investigated the informational advantages of the central bank. Section 3 describes the data used. Romer and Romer (2000) show that incorporating confidential supervisory information about bank health improves central bank forecasts of both unemployment and inflation.

qÞ ¼  5 T1 X This assumption implies that bank’s bonds will yield the risk free rate even in case of default for the bank. only debt is directly observable. (1) to get the one-period transition density of At. the probability of default (i. k. Therefore. For the rest of our analysis. Duan (1994. The distance-to-default measure has the advantage of being unbounded. 2000) shows how to estimate the unobserved evolution of the asset value and its variance making use of the observed evolution of market capitalization. 2. Duan (1994. we assume that the variance follows a GARCH(1. that the initial price Et of the banking sector’s equity is: Et ¼ At Fðdt þ st Þ  Dt Fðdt Þ (4) lnðAt =Dt Þ  s2t =2 (5) with dt ¼ st where F(∙) is the cumulative normal distribution. A lower distance-to-default always implies a higher probability of default. Furthermore. 2000) shows that the log likelihood function takes the form: T1 T1 X X ^t þ st Þ  T  1lnð2pÞ  1 ^ lnFðd lns2t  lnA tþ1 2 2 t¼1 t¼1 t¼1  2  2 ^ X DlnA 1 T1 tþ1  m  st =2  2 t¼1 s2t LðEt . which is equal to: DlnAtþ1 ¼ m  s2t =2 þ ut (6) where DlnAt þ 1 ¼ lnAt þ 1lnAt and ut is white noise with variance s2t . This is why dt is often called the distance-to-default. b. Jokipii. the probability of not exercising the option) is equal to 1  F (dt). To compute the distance-to-default using Eq. Note that these two measures are equivalent since they are linked by a strictly continuously decreasing function. Eq. The default probability on the other hand. Monnin / Journal of International Money and Finance 32 (2013) 1–16 By assuming.5 Merton (1974) shows. In his original paper. its volatility rate st and the debt Dt.e. a. 2000) works with a constant variance. using the Black-Scholes option pricing formula.1) process characterized by: s2t ¼ k þ aε2t1 þ bs2t1 (7) The set of unknown parameters q include m.e. and its variance rate we make use of Eq.e. To obtain the asset market value.e. (4) which links the unobserved assets with the observed market capitalization of the banking sector (i. P. We start by rewriting Eq. as stated in Section 2. One can show (see Bichsel and Blum. its equity value Et). we will not work with the default probability directly but we will rather use the distance-to-default dt. 2004) that dt measures how far. we need to know the asset market value At. the probability that the assets At þ 1 are greater than the debt Dt þ 1). Estimation of the distance-to-defaults with equity market value Duan (1994. the point where assets are equal to debt). in terms of standard deviations. Note that F(dt) in Eq. is bounded between zero and one. 2000) makes use of the fact that for any given set of parameters q. (4) is a one-to-one mapping between Et and At. (4) corresponds to the probability of exercising the option (i. Duan (1994.1 that At is log normally distributed and that debt is insured.3. which is convenient for empirical estimations.4 T. In practice however. . the banking sector is from its default point (i. we extend his method to estimate a time-varying variance. The parameter set q can therefore be estimated by maximizing a log likelihood function defined on Et instead of on At. (5).

1. We make use of a pVAR approach predominantly for the reason that the interactions between financial distress and the real economy have not been rigorously identified theoretically. The debt data are annual and have been transformed into quarterly data by linear interpolation. Monnin / Journal of International Money and Finance 32 (2013) 1–16 5 ^t correspond to d with ^ t is the unique solution to Eq. Similarly. Belgium (12 banks). Ireland (5 banks). United Kingdom (14 banks) and United States (37 banks). 7 We additionally experiment by using the average market value observed during the quarter. We consider a banking sector to be unstable if its distance-to-default is in the first 20% of all distance-to-defaults observed internationally over the period 1980–2007. These countries are Australia (17 banks). we obtain debt and market value data from Datastream. given that the parameters q and d where A t ^ At instead of At . Norway (32 banks). We assume that in each country. Iceland.7 Fig. Japan (110 banks). we make use of real GDP growth. Austria (10 banks). (4) for Et. Fig. 3. Finally. Luxembourg.2 to define the thresholds of stability on which we base our assessment. Other data In addition to the data used to calculate the distance-to-default index for each country. we also controlled for other exogenous variables such as the US real interest rate. Canada (10 banks). Netherlands and New Zealand have been excluded because we could not get data for at least 5 banks. The first term on the right-hand side of the equation corrects the traditional log likelihood function to account for the fact that we obtain the asset value At from the observable equity price Et via Eq. 1 presents the estimated distance-to-defaults for each country. the global price of stocks and the global growth rate. None of these variables were significant and changed the conclusion significantly.2. Greece (21 banks). France (47 banks). 1990Q1 – 1993Q2 (housing crises in several countries). However the results are not significantly affected and are therefore not presented here for brevity. We can observe four periods in which at more than third of the countries simultaneously experience instability in their banking sector: 1987Q4 (Black Monday and the following stock market crisis).8 The seasonality component of price movements is removed using the X12 method in Eviews.T. Spain (23 banks). The maximization of the log likelihood function can be performed with ordinary maximization techniques. Switzerland (33 banks).6 To calculate the distance-to-default index outlined in Section 2. Portugal (9 banks). Germany (32 banks). Jokipii. Distance to default Our sample consists of 521 banks. banking sector stability is measured by the relative degree of distance-to-default deviations from its own historical mean. Data 3. 8 In some preliminary estimation. oil price levels and inflation between 1980Q1 and 2008Q4. Finland. 4. covering 18 OECD countries. We decided not to include these variables in our final estimation in order to achieve higher degree of freedom and thus narrower confidence intervals for our variables of interest. stable and very stable. In other words. Using a VAR therefore allows us to 6 We selected countries that joined the OECD before 1975 and that are classified as high-income countries by the World Bank. . Note that the three last terms on the right-hand side of the equation would correspond to the log likelihood functions of a GARCH process if the asset value were directly observable. we consider a banking sector to be very stable if its distance-todefault is in the last quartile of the distribution. 3. (4). The horizontal lines represent the first and third quartiles of the distance-to-defaults of all countries. an unusually unstable banking sector corresponds to a distanceto-default that is substantially below the country historical mean. Denmark (52 banks). Italy (50 banks). Sweden (7 banks). a banking sector is stable in all others cases. 2 documents the number of countries with an unstable banking sector in each period. We define three possible states for the banking sector: unstable. These quartiles are used in Section 4. P. 1998Q3 (Russian and LTCM crises) and in 2008 (subprime crisis). The quarterly market value is the minimum of daily market values observed during the quarter. Impact of banking sector stability real output growth and inflation In order to assess whether banking sector stability is linked to real output growth and inflation we adopt a panel vector autoregression (pVAR) approach with state-dependent variance.

allows us. when compared to a traditional VAR approach. The choice of a state-dependent variance derives from 14 12 10 8 6 4 2 0 1980 1985 1990 1995 2000 Fig.6 T. 1. This is particularly useful in our context since episodes of banking sector instability are relatively rare. by using cross sectional data. Estimated distance-to-defaults. impose as little a priori theorizing as possible. a panel VAR. Jokipii. to increase the number of observations and thus the precision of our estimations. Monnin / Journal of International Money and Finance 32 (2013) 1–16 Australi a Austria 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Belgiu m 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 France 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Germany 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Greece 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Japan 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Norway 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Switzerla nd Ireland 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Portugal 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 United Kingdom 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Canada 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Spai n 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Denmark 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Italy 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Sweden 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 United States 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 32 28 24 20 16 12 8 4 0 1980 1985 1990 1995 2000 2005 Fig. 2. 2005 . Moreover. Number of countries with weak banking sector. P.

The log likelihood function is then: l ¼ T X I X N 1 1  lnð2pÞ  lnjHit j  ε0it Hit1 εit 2 2 2 t¼1 i¼1 (14) The log likelihood can be estimated with traditional maximization techniques. in addition to studying the impact of banking sector stability on the levels of GDP growth and inflation. The covariance matrix can be decomposed as: Hit ¼ Dit Ri Dit (9) with 0 1 B fi21 Ri ¼ B @ « 1 fi12 fi13 / fi1N fi23 / fi2N C C 1 « « fiN1 fiN2 fiN3 1 / « A 1 (10) where fijk is a constant country-specific correlation coefficient between variables j and k in yit and 0 s2it1 B 0 D2it ¼ B @ « 0 0 s2it2 « 0 0 0 « 0 1 / 0 / 0 C C 1 « A / s2itN (11) where s2itn is the variance of residuals for country i at time t for the nth variables of yit. xit is a vector of exogenous variables (i. In addition. The model is estimated using real GDP growth. Jokipii. We use a multivariate version of the estimation methodology proposed by Harvey (1976). which is function of a set of dependent variables zt  1 (see below). Monnin / Journal of International Money and Finance 32 (2013) 1–16 7 a preliminary analysis showing that the residuals of a simple pVAR were heteroscedastic. inflation and banking sector stability). price levels and oil price growth data between 1980Q1 and 2008Q4. P. Thus. mi is a fixed effect for each country. I is the number of countries and N is the number of variables in each vector yit.e. output growth.e. To model Hit. . oil prices). we standardize9 the distance-to-default for each country to eliminate the differences in distance-to-default levels and volatility between countries. The pVAR model has the following specification: 1=2 yit ¼ Ayit1 þ Bxit þ mi þ Hit εit (8) where yit is a vector of endogenous variables for country i at time t (i. our framework allows us to explore its impact on both GDP growth and inflation uncertainty. we adopt a strategy similar to the conditional constant correlation (CCC) of Bollerslev (1990). Note that zit can contain some or all variables in yit and xit.T. The likelihood function of the pVAR given q is: L ¼   1 0 1 exp  H ε ε N=2 2 it it it jHit j1=2 i ¼ 1 ð2pÞ T Y I Y t¼1 1 (13) where T is the number of periods. Each variance depends of a set of past variables zit through the following function: s2itn ¼ expðain þ an zit1 Þ (12) where ain is a fixed effect specific to each country and an is a common slope. This is equivalent to 9 A standardized variable is equal to its value subtracted from its mean and divided by its standard deviation. εit is a vector of independent error terms normally distributed and Hit is a country-specific time-varying covariance matrix.

banking sector stability is measured by the relative size of distance-todefault deviations from each country’s historical mean.3483** *(**) indicates that the coefficient at significant at the 5% (1%) level. In what follows.6499** 0.0020** 0. we focus solely on the one-way relationship between banking sector stability and output growth since we wish to shed light on the impact that the banking sector can have on the real economy. This means that a stable banking sector is followed by 10 The number of lags has been chosen with the Akaike Information Criterion. only coefficients related to the endogenous and exogenous variables are presented. an unusually unstable banking sector corresponds to a distance-to-default that is significantly below the individual country historical mean. (8) with GDP growth and banking sector’s distance-to-default as endogenous variables and oil price yield as exogenous variable. The results indicate that both the level of output growth and banking sector stability have a significant impact on their own variance.0011 2. The impulse-response functions also show that the impact of a positive GDP shock on the improvement of banking sector stability is long lasting. DD denotes distance-to-default. 3).1754 0. This is in line with the negative link between variance and growth identified by Ramey and Ramey (1995) and Fatás and Mihov (2003). Level GDPt1 GDPt2 DDt1 DDt2 Oilt Oilt1 Oilt2 Variance GDPt DDt logðs2GDPt Þ logðs2DDt Þ 0. This signifies that uncertainty about future banking sector stability is higher during periods of stability.8 T. Jokipii. banking sector stability is important for subsequent output growth.2861** 6.10 For clarity. Table 1 additionally shows that banking sector stability has a negative and significant impact on output growth variance (column 3. The positive impact of banking sector stability on output growth is observable from the impulseresponse functions of the estimated pVAR (lower left-hand side panel in Fig. in terms of historical standard deviations. Combined with the positive autocorrelation for the banking sector described previously. in each country.0010** 0.1.1982** 0. rows 3 and 4).0060 3. Impact on real output growth Table 1 displays the estimated coefficients of the pVAR model with a lag of two for the endogenous and the exogenous variables.0006 0. uncertainty about future output growth is smaller in booms than in recessions.0397 0.9671** 0.0495 0. Our results suggest that greater stability within the banking sector induces growth in the subsequent periods (column 1.2035** 0. fixed effects are omitted. 4. Variance: estimated coefficients for Eq.1. Linear impact of banking sector stability 4.0961** 0. row 3). 3). 11 .7156 11. left-hand side panels of Fig. P. rows 1).3239** 0.0526* 16. than a stable banking sector is to be followed by another period of stability. Level: estimated coefficients of Eq. Higher output growth is followed by lower output variance in the following quarters (column 3. in different states of the world. assuming that.0093 0.1. In other words. In other words. Real output growth is important for subsequent stability in the banking sector and similarly. The estimated coefficients show that the relationship between banking sector stability and real output growth is a two-way relationship.11 In contrast. Monnin / Journal of International Money and Finance 32 (2013) 1–16 Table 1 Linear impact of banking sector stability on real output growth. Moreover a one standard deviation shock (either to banking sector stability and to output growth) is more persistent in its effect on banking sector stability than on output growth (right-hand side vs.8144 0. rows 3 and 4). (12) with GDP growth and banking sector’s distance-to-default as independent variables. Table 1 additionally presents the estimated coefficients for the output growth and banking sector stability variances. the variance of banking sector stability increases when the banking sector’s distance-to-default is high (column 4. our results imply that an unstable banking sector is more likely to be followed by a further period of instability.2104** 0.0011** 0.

A positive shock to banking sector stability increases its own variance (right-hand side lower panel). Finally. we find that higher uncertainty about the stability of the banking sector generally follows higher output growth (column 4. Combined with the positive link between output growth and banking sector stability. than an unstable banking sector is to be followed by low output growth.13 It shows that neither inflation nor banking sector stability has an impact on the other. Interestingly.2. . Fig. our findings indicate that a stable banking sector is more likely to be followed by high output growth. Jokipii. row 2). Impulse-response function for real output growth and banking sector stability. 5) show that higher inflation increases the uncertainty about next period inflation. a shock to output growth increases banking sector stability variance (right-hand side upper panel). Monnin / Journal of International Money and Finance 32 (2013) 1–16 9 Fig.1. this means that. P. The simulated shocks are equal to one standard deviation either to banking sector stability or to output growth. even-though banking sectors are on average more stable after a growth in output. Combined with the positive link between banking sector stability and output growth. but decreases the variance of distance-to- 12 13 14 The simulated shocks are equal to one standard deviation either to banking sector stability or to output growth. As explained previously. The number of lag is determined by the Akaike Information Criterion. a positive shock (to either output growth or to banking sector stability) reduces the variance of output growth (left-hand side panels). Impact on inflation Table 2 presents the estimated pVAR with inflation and distance-to-default as endogenous variables with a lag of 2. 4. 4 documents the behavior of output growth and banking sector stability variances after shocks. the probability of observing a banking crisis is higher after a period of growth.12 The graph shows the difference between the average variance with and without shocks. The results for the variance equations and for their impulse-response functions14 (Fig. 3. reduced uncertainty about future output growth.T.

10 T. Level Inflationt1 Inflationt2 DDt1 DDt2 Oilt Oilt1 Oilt2 Variance Inflationt DDt logðs2Inflt Þ logðs2DDt Þ 1. Non linear impact of the banking sector stability In the previous section we showed that some empirical evidence of a linear link between banking sector stability and the distribution of output growth exists.2286** 0. As per Section 3.0097 0.5015** 0. Impulse-response functions additionally show that the variance of inflation tends to decrease for a very short period after a positive shock to banking sector’s degree of stability. we briefly investigate this possibility with a simple model. some authors have suggested that the main impact of banking sector stability on the real economy is non-linear. we assess whether real output and inflation distributions are different in periods following different state of banking sector stability.0254 0. Essentially. 4.3480** *(**) indicates that the coefficient at significant at the 5% (1%) level.5770** 0. (8) with inflation and banking sector’s distance-to-default as endogenous variables and oil price yield as exogenous variable. default after one period. we check whether there is an asymmetric reaction of real output growth and inflation to banking sector stability. Variance: estimated coefficients for Eq. Impulse-response function for real output growth and banking sector stability variance.0280 0.0007 0.1427** 0. Jokipii. P. we Table 2 Linear impact of banking sector stability on inflation.0487 0. .0080 0.0330 0.2.0623 0. Monnin / Journal of International Money and Finance 32 (2013) 1–16 Fig.1910** 0.1143* 0.9765** 0.1237** 0.2172** 0.3227** 0. (12) with inflation and banking sector’s distance-to-default as independent variables.0428 0. Level: estimated coefficients of Eq. More specifically. In this section.0458** 0. 4.1309* 0. However.

define three states for the banking sector: unstable. Table 3 presents the estimated coefficients of the new pVAR. We say that the banking sector is unstable if its (standardized) distance-to-default is in the first 20% of all distance-to-default observed internationally over the period 1980–2008. t1 D1. t. (12) with inflation and banking sector’s distance-to-default as independent variables.0012 4.0007 0. Variance: estimated coefficients for Eq.3709** 1. They indicate that only unstable banking sectors have a significant (and Table 3 Non linear impact of banking sector stability on real output growth.3758** 1. Similarly.0004 0.0008 0. Impulse-response function for inflation and banking sector stability variance. 5.2483 7.0877 0. The results for the dummy variables are particularly interesting. t2 D2.2010** 0.0214 17.3040* 0. t2 Oilt Oilt1 Oilt2 Variance GDPt DDt logðs2GDPt Þ logðs2DDt Þ 0.2115** 0.0019** 0.t1 D2.T. and D2.2440** 0.6422** 0. Level: estimated coefficients of Eq.0209 0.0539** 3. is equal to one when the banking sector is very stable and zero otherwise. we say that the banking sector is very stable if its distance-to-default is in the last 20% of the distribution. (8) with inflation and banking sector’s distance-to-default as endogenous variables and oil price yield as exogenous variable. .1747** *(**) indicates that the coefficient at significant at the 5% (1%) level.3551** 0. P. t. (8) by two dummy variables: D1.0067 0. stable and very stable. Level GDPt1 GDPt2 D1.1445 0.4656** 0.1228** 0.3846** 0. Jokipii.0826** 8. Monnin / Journal of International Money and Finance 32 (2013) 1–16 11 Fig.3045* 2. The banking sector is stable in all other cases.1969** 0.0016** 0. we make use of a panel VAR but we replace the lagged distanceto-default of the right-hand side of Eq.0019** 0. is equal to one when the banking sector is unstable and zero otherwise. Similarly to the linear estimations.2373 0.

0098** 0.0654 0.0043 0.2017 0. rows 9 and 11).0847 0.0427 0. In this section.0386 0. It is still a function of the macroeconomic environment as banking sector stability decrease for higher short term interest rates and higher investments (column 2.0485 0. Particularly stable banking sectors have no impact on real output growth.0173 0. Monnin / Journal of International Money and Finance 32 (2013) 1–16 Table 4 Non linear impact of banking sector stability on inflation. t1 D1. Table 4 presents the estimated coefficients of the non-linear pVAR with inflation.1034 0.3. t2 D2.0757 0. both unstable and very stable banking sectors have a significant impact on real output growth variance. As previously. money growth (DM2t) growth and investments (DKt) in the sample. real output growth is also dependent on banking sector stability. The reasons for that are mainly that we are not interested in the long term structural influence of the banking sector on growth but more in the transitory impact of the fluctuation of its stability.6516** 0.1 is very simplistic since it assumes that real output is function of its own lags.g.0815 0.0985* 0. Expanding the set of macroeconomic variables The model estimated in Section 4.1259** 0. t2 Oilt Oilt1 Oilt2 Variance Inflationt DDt logðs2Inflt Þ logðs2DDt Þ 1. The introduction of the new variables changes the results for distance-to-default variance: the uncertainty about banking sector stability decreases with higher real output growth.3533** 1. Indeed. They show that the linear links uncovered are predominantly driven by periods of banking sector instability rather than by a smooth and continuous link between banking sector stability and the real economy.12 T.2236 *(**) indicates that the coefficient at significant at the 5% (1%) level.0161 0. of banking sector stability and of oil prices only. The non-linear model provides some interesting insights into the results from the linear model for output growth. Variance: estimated coefficients for Eq. several other economic variables can impact output growth. Greece and Sweden are missing from the initial sample. In reality. short term interest rates (STIRt).3387** 1. (8) with inflation and banking sector’s distance-to-default as endogenous variables and oil price yield as exogenous variable. In contrast. 15 Australia.1. However.0497** 0. after adding these variables. Table 5 also shows that real output growth variance is still a negative function of banking sector stability – i.3342** 0. we add consumption growth (DCt). political environment. uncertainty about real output growth increases with banking sector instability. and contrary to the results in Section 4.4536** 0.e.0224 0. The results are less clear than for the output growth.0189 0.2686* 0. Jokipii. Table 5 shows that real output growth and distance-to-default are still functions of their own lag.1281** 0. Level: estimated coefficients of Eq. t1 D2. P. we extend the set of exogenous macroeconomic variables included in the pVAR to capture their influence on real output growth. (12) with inflation and banking sector’s distance-to-default as independent variables. We chose to present these results in a separate section and not directly in the main section because the variables that we include are not available for all the countries studied or for the same period. etc.1336* 0. . the sample is reduced to 15 countries instead of 1815 and the number of observation shrinks from 1757 data points to 1052. but it seems that unstable banking sectors slightly increase inflation (column 1. 1999) and not long term growth rates (like human capital. row 4). Level Inflationt1 Inflationt2 D1.0421 0. 4. We chose variables that are likely to influence transitory business cycles of GDP (see e. More precisely.).5162** 0. negative) impact on real output growth. banking sector stability does not depend on real output growth anymore.0456 0. Stock and Watson.

8445** 0.0000** 2. and in turn. Level: estimated coefficients of Eq.0074 0.0011 0.6176** 0.9067 0. it is possible that considering the state of the banking sector in a systematic way would improve forecasts. we extend our analysis by assessing whether additional information embedded in our stability index might help to improve output growth forecasts. banking sector stability brings 16 http://www. our dataset for the remainder of the paper consists of quarterly data between 1980 and 2001.T.0136* 0.1029* 0.3182 0. The correlations are presented in Table 6. (8) with GDP growth and banking sector’s distance-to-default as endogenous variables and oil price yield as exogenous variable. implicitly assuming that there is a one-way causal effect from the degree of stability to the forecast errors. Essentially. Banking sector stability and central bank forecast errors The recent global financial crisis highlighted the tight interdependence between monetary and financial stability. We base our analysis on the United States since forecast data is publicly available on the Fed website back to1965.philadelphiafed. Since the forecast error is calculated as the difference between the actual value and the forecasted rate. regardless of whether the path of effect is direct or indirect. Since economic activity and interest rates affect financial sector risks. three-period and fourperiod ahead forecasts on distance to default. make no a priori assumptions as to whether one variable is dependent on the other. We therefore investigate whether Fed growth forecasts are indeed making use of information contained in our measure of banking sector stability. additionally plays a key role (see Reifschneider et al. the result indicates that banking sector stability today results in a significant underestimation of GDP growth in the subsequent quarters. Despite playing an important role in Fed projections for macroeconomic growth. For comparability with our previous estimations. based on additional information not systematically included within the models. We therefore proceed to describe the dependence of forecast errors on the banking sector stability. Variance: estimated coefficients for Eq. . however. To do this. however. The results are presented in Panel a of Table 7.0081 0. A substantial degree of discretion. adopting an ordinary least squares (OLS) framework. The link between banking stability and economic activity is of particular interest to policy makers which base their monetary policy decisions on economic forecasts.0000** 12. We find that banking sector stability has a positive and significant impact on the two-and three-period ahead Fed forecasts. Jokipii. two-period. Correlations between the stability indicator and the forecast errors are significant for current forecasts as well as for two-period and three-period ahead forecasts indicating that some degree of association between these two measures exists. (12) with GDP growth and banking sector’s distance-to-default as independent variables. econometric models are not the sole input in the regular forecast process. we regress one-period. 1997). P.org/econ/forecast/greenbook-data/index.5655 0.0008* 0.0011 0.9168 0.2489 0.3829 0. 5.0007 0.0848* 0. In this section.2344** 0. the financial sector affects the real economy.0003** 0.16 with a five year delay.2032** 0.0200** 0. unclear what information is considered and whether judgment based adjustments to forecasts are efficient.2313** 0.5079* 9.8984** 0. It is. emphasizing the need for policy frameworks to be adjusted accordingly. Monnin / Journal of International Money and Finance 32 (2013) 1–16 13 Table 5 Linear impact of banking sector stability on real output growth: extended sample.0008** 0..cfm.0304 14. Simple correlations. Level GDPt1 GDPt2 DDt1 DDt2 Oilt Oilt1 Oilt2 DM2t1 STIRt1 DCt1 DKt1 Variance GDPt DDt logðs2GDPt Þ logðs2DDt Þ 0.0155 0.3826** *(**) indicates that the coefficient at significant at the 5% (1%) level. We start by computing simple correlations between the current distance to default and the period ahead forecasts.

12* 0.10 0. P. Monnin / Journal of International Money and Finance 32 (2013) 1–16 Table 6 Correlations: distance to default with forecast errors. coefficients on the two-period (and three-period) ahead forecasts are negative and significant.32* 0. Specifically.34* 0.29** 0. DD(C).43 0. We capture the degree of stability by estimating the distance to default for each country’s banking sector at each point in time.26 0. Similarly. Focusing on data from the Fed.12 0. 6. Periods of stability are generally followed by an increase in real output growth. Conclusion In this paper.21 D2 Two-period ahead forecast Three-period ahead forecast 0.29** 0.30** 0.24** 0. we explore the relationship between banking sector stability and the subsequent evolution of real output growth and inflation. adopting OLS we regress the one-period. We find that the relationship is asymmetric by nature.18 Four-period ahead forecast 0. confirming. Two dummy variables D1 and D2 are created as per Section 3. we show .41 0.41** 0. D1 is equal to one when the banking sector is unstable and zero otherwise. show that the relationship between period ahead forecasts and distance to default is driven by periods of instability.31 0.39 0.34** 0. the one-period ahead forecast etc. in that it is driven predominantly by periods of instability.25 0. three-period and four-period ahead forecasts on D1 and D2.31 0. Dependent variable One-period ahead forecast Panel a: linear estimations Distance to default 0. refers to the current forecast. more growth than expected. presented in Panel b of Table 7.37* 0.34* 0.12** 0. Table 7 Estimation: distance to default and Fed forecast errors.39 0. DD(L1).28** 0. stability with a two-period lag. DD(L2) denote the current stability.26** 0. Again. D1 and D2 are dummies capturing periods of instability and stability respectively. Our analysis is further extended to assess whether additional information embedded in our stability index might help to improve central bank growth forecasts.29 0.37 0. defining periods of instability. D2 is equal to one when the banking sector is very stable and zero otherwise.47** 0.19 Panel b: non-linear estimations 0.31 0.18** 0. Jokipii. The results.39 0.43** 0.10 D1 0. while instability corresponds with subsequent periods of reduced growth.50*** 0.27 0.38** 0.51*** 0. one.44 0.23** 0. stability with a one-period lag.29 0.41 0.28 0. two-period. current.29 0.28 Note: *(**) denotes significance at the 5% (1%) level. We additionally experiment with asymmetric effects in this regard. We also show that a stable banking sector reduces real output growth uncertainty. stable and very stable periods relative to both historical and international average. Forecast errors are calculated as the difference between actual and predicted GDP growth. Adopting a panel VAR methodology to assess both linear and non-linear relationships between our variables of interest we show that banking sector stability is an important driver of GDP growth. Current One Two Three Five Six Seven Eight DD(C) DD(L1) DD(L2) DD(L3) DD(L4) 0. To do this we distinguish between unstable.04*** 0.14 T.44 0.23* 0.41* Note: *(**) denotes significance at the 5% (1%) level.34 0. We further explore the possibility of an asymmetric impact of banking sector stability on the real economy. stability and high stability as per Section 3. On the vertical axis. capturing periods of instability and periods of stability respectively. This finding indicates that additional information embedded in our stability index measure has the potential to improve Fed growth forecasts.29 0. In particular. Forecast errors are calculated as the difference between actual and predicted GDP growth.

Does the federal reserve possess an exploitable informational advantage? Journal of Monetary Economics 50. Applied Financial Economics 14. P. 498–505. American Economic Review 89. 629–655. 1138–1151. Johnsen.. R. In: Aghion. 825–855. The determinants of banking crises in developing and developed countries. The case for restricting fiscal policy discretion... Law. Credit. Zicchino. Review of Economics and Statistics 72. A. 1999.. 89–112. 2000.. T. H.. Our findings have several important implications for policy makers. Finance and growth: Schumpeter might be right. R. Aftermath of banking crises: effects on real and monetary variables. Peek. Bichsel. has the potential to further improve economic forecasts. Pazarbasioglu.. 1976. 596–620.. Merton. 05/96. Is bank supervision central to central banking? Quarterly Journal of Economics 114. Perrez. 1–7. R... A. Institutions and Instruments 14... A. Journal of Economic Literature 35. 449–470. 688–702. Gupta. King. Reis. L. R. Journal of International Money and Finance 24. Saporta. The relationship between risk and capital in Swiss commercial banks: a panel study.T. Tootell. Financial intermediation and growth: causality and causes.. 461–465. In addition... Kaminsky. Journal of Financial Intermediation 17.. R.. B. 513–542. Journal of Banking and Finance 26. D. J. American Economic Review 88. 2003. E.A. Mathematical Finance 10. C. Zingales. Duan. Levine. Levine.. On the pricing of corporate debt: the risk structure of interest rates. Towards a measure of financial fragility. 2107–2137. A. 559–586. Country-Level Macro Stress-Testing Practices. Cross-country evidence on the link between volatility and growth. Mimeo..-C. I. Levine. Journal of Financial Economics 108. Kwak. J. Goodhart. Levine. 1974. P.. J. Reinhart. Otker-robe. 2005. Rosengren... D. 2007.. J. Journal of Money. Journal of Monetary Economics 32.A.. Levine. 155–167. Quarterly Journal of Economics 118. financial dependence. E. Jokipii. G. Cross-Country Empirical Studies of Systemic Bank Distress: A Survey. and firm growth.. 2000. (Eds. 2007. Hillbers.. Duan. R. banking sector stability appears to be an important driver of GDP growth in subsequent periods. Credit. R. Levine. 289–328. R. 2008. This result is in line with the notion that additional information. The legal environment. The twin crises: the causes of banking and balance of payments problems. P. The real effect of banking crises. G.. G. 2001. Our findings indicate that banking sector stability (instability) results in a significant underestimation (overestimation) of GDP growth in the subsequent quarters. D. Modeling the coherence in short-run nominal exchange rates: a multivariate generalized ARCH model.. 31–77. G. The Interaction Between the Swiss Banking Sector and the Macroeconomy: Results from a VARAnalysis.. Demirgüç-Kunt. G. C. Klingebiel. Rosengren. 2006. and growth. 1999. 1998. 2008. Boyd. Beck. 2005. Kroszner. highlighting the need for policy makers to consider banking sector measures in their forecast models. Harvey. Correction: maximum likelihood estimation using price data of the derivative contract. 473–500. I. R.. R. Maximum likelihood estimation using price data of the derivative contract. Mihov. R.H.. Sunirand. Tsomocos. rajan.. and Banking 37 (4). 187–225. 2004. Financial Stability Review.. G. Laeven. 461–462. European Central Bank. Deutsche Bundesbank. References Allenspach.P.. L. entrepreneurship. G.B. Elsevier. Bollerslev. Smith. 1995. . P.P.. N. 1997.. 688–726. R.. C. 2003. banks. 817–839. IMF Working Paper. Estimating regression models with multiplicative heteroscedasticity. Maksimovic. 717–738. L. Dell’ariccia. 1993a.-C. Costs of banking system instability: some empirical evidence. Peek.S. 2005. 1990. 147–154. 1993b. 675–691. P. Economic Theory 27. G.. Demirgüç-Kunt. International financial liberalization and economic growth. and growth: theory and evidence. Fatás. Detragiache.G. Goodhart. S. pp. 05/151. 37–74....S. E. 2005. Annals of Finance 3. J. 2006. 1997. 2006. J. A model to analyse financial fragility. T. Finance and growth: theory. Finance. Blum. O. Benston.G. R. Tootell. 2005....M. Detragiche. 1419–1447. Financial development and economic growth: views and agenda. Journal of Monetary Economics 35. Detragiache. American Economic Review 85. Financial dependence and growth. embedded in our stability index measure.. highlighting the need for greater attention to be paid to banking sector soundness in the implementation of economic policy. Review of International Economics 9. Durlauf. First. Ramey. Handbook of Economic Growth. Financial Markets.G. 1998. Mathematical Finance 4.. evidence...... IMF Working Paper...). A. Journal of Money. we find that additional information contained in our banking sector measure could help forecasters to reduce forecast errors. 1994.A. E. we show that for a sample of 18 OECD countries. Mathematical Finance 53. Levine. V. E. Aspachs. King. Ramey.. and mechanisms.M. Benink... Econometrica 44 (3). 977–999. The real output losses associated with modern banking crises. S... Swiss National Bank. Demirgüç-Kunt. 2002. The future of banking regulation in developed countries: lessons from and for Europe.B. IMF Staff Papers 45. and long -run economic growth. Monnin / Journal of International Money and Finance 32 (2013) 1–16 15 that output forecast errors are correlated with our stability measure.. 1986. Banking crises. 2004.. finance. R. Loayza. V. Assessing and Managing Rapid Credit Growth and the Role of Supervisory and Prudential Policies. 81–109.. Journal of Financial Economics 84.. N. Tsomocos. Hogart. J.. Rajan.. V..C.. C. R.S. Journal of Finance 29.. Financial Stability Report. and Banking 30.

D.. Westermann. R.H. Jokipii. D. 2007. Larger Crises Cost More: Impact of Banking Sector Instability on Output Growth. MPRA Paper.J. Watson..D. F. (Eds. Quarterly Journal of Economics 123.. Woodford. Econometric models and the monetary policy process. 1999.W. D. J. 2008.S. Stock. 1–37. A. 1997... 359–406.. In: Taylor. 5101. Wilcox. Serwa. Elsevier.L.16 T. . D. M.. Systemic crises and growth. Business cycle fluctuations in U. Monnin / Journal of International Money and Finance 32 (2013) 1–16 Rancière. Romer. 429–457. Romer. Tornell. Reifschneider. C. American Economic Review 90.. P. D..)..H.. Handbook of Macroeconomics. Pricing risk-adjusted deposit insurance: an option based model. macroeconomic time series. M. Stockton.W. J. Carnegie-Rochester Conference Series on Public Policy 47. 2000..