Economic Integration and Financial Stability: A European Perspective

Gianni De Nicolo; Alexander F. Tieman

doi:10.5089/9781451865561.001.A001

Jump to Content Jump to Main Navigation

Economic Integration and Financial Stability

A European Perspective

Author: Mr. Gianni De Nicolo and Alexander F. Tieman¹

¹ 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Contributor Notes

Author’s E-Mail Addresses: gdenicolo@imf.org; atieman@imf.org.

Publication Date:: 01 Dec 2006

eISBN:: 9781451865561

ISBN:: 9781451865561

Language:: English

Keywords:: WP; IPG rate; bank risk profile; insurance risk profile; profiles of bank; risks increase

Download PDF (339.6 KB)

This paper assesses changes in synchronization of real activity and financial market integration in Western Europe and evaluates their implications for financial stability. We find increased synchronization of real activity since the early 1980s and increased equity markets integration since the early 1990s. We also find that measures of systemic risk at large European financial institutions have not declined during the period 1990-2004 and that bank systemic risk profiles have converged. At the same time, the sensitivity of bank and insurance systemic risk measures to common real and financial shocks has increased in most countries. Overall, these results suggest that the integration process does not necessarily entail an unambiguously positive effect on financial stability.

Abstract

I. Introduction

Research on synchronization of real activity and financial integration in Europe has intensified in the past few years. Nevertheless, the implications of integration for financial stability remain largely unexplored. This paper aims at contributing to fill in this gap.

Structural changes in the environment in which financial firms operate, such as increased real synchronization and advances in financial integration, may affect individual and system-wide risk profiles of financial intermediaries differentially (see De Nicolò and Kwast, 2002, and De Nicolò and others, 2004e). On the one hand, enhanced synchronization in real activity may reduce the benefits of cross-country diversification. If either the shocks hitting a set of economies (and the relevant borrowers) become more similar, or the transmission mechanism of country-specific shocks becomes stronger, or both, then the pool of diversifiable (credit and market) risks available to intermediaries may shrink. On the other hand, financial integration may enhance diversification opportunities for individual intermediaries, which can rely on enlarged investment opportunities across activities and borders to enhance expected returns for the same amount of risk. Yet, a set of intermediaries may become less diversified as a whole if their exposures to the same risks increase, either by choice or because the sources of “aggregate” risk have become more similar. Moreover, increased linkages among intermediaries through enhanced common exposures to financial markets may make their exposure to contagion more likely.

Disentangling these possibly countervailing effects on financial stability is the main task of this paper. Accomplishing this task requires first assessing whether increases in real synchronization and advances in financial integration have indeed occurred, since the existing literature does not offer unequivocal answers. Second, it requires constructing measures of systemic risk, and relating them to outcomes of changes in real synchronization and financial integration.

With regard to real synchronization, several studies have attempted to identify a “European business cycle,” but research on the existence of such an object is still ongoing; as a result, few studies have focused on changes in real synchronization.² Moreover, this literature has dealt almost exclusively with fluctuations of GDP and/or industrial production growth rates. As our focus is on the impact of changes in real synchronization on the risk profiles of financial institutions through their portfolio choices, synchronization of volatility of growth rates of real activity may be as important, if not more important, than synchronization in levels.

With regard to financial integration, recent studies have documented increased convergence in prices of money and bond markets, while noting the slower pace of price convergence in retail bank credit markets (Barros and others, 2005; Baele and others, 2004; and Adam and others, 2002). However, the literature exhibits mixed results concerning the integration of equity markets. As stressed by Adjaouté and Danthine (2004), a difficulty in assessing integration lies in disentangling pricing effects from changes in fundamentals. Yet, we view an assessment of advances in equity market integration as a robust gauge of advances in financial integration more generally, since equity markets are ones in which claims on a large variety of countries’ investment opportunities are traded, and integration in such markets does not necessarily follow mechanically from cross-country convergence of interest rates.

We proceed in three steps. First, we assess cross-country convergence of the first and second moments of output growth. The consideration of second moments is novel, and turns out to be informative on the changing nature of common versus country-specific driving forces of the dynamics of real activity. Second, we test whether cross-country convergence of estimates of a discount factor used to price “idiosyncratic” risks in equity markets has occurred, employing a version of the methodology introduced by Flood and Rose (2005).

Finally, we document the dynamics of proxy measures of systemic risk based on data for a set of large European banks and insurance companies in the past 15 years. We test convergence in both levels and volatility of these dynamics, and assess whether the risk profiles of these financial institutions have become more sensitive to common real and financial shocks. In doing so, we view the sensitivity of financial institutions’ risk profiles to common real and financial shocks as a useful metric to gauge the implications of increased synchronization in real activity and advances of financial integration through the overall exposure of intermediaries to “common” market and credit risks.

Our investigation yields three main sets of results. First, we find evidence of increased synchronization in the dynamics of real activity since the early 1980s, in the form of declining trends in the cross-country dispersion in the mean and volatility of industrial production monthly growth rates. These declining trends are found after controlling for common shocks, whose magnitude has become smaller, and are mainly driven by business cycle synchronization. Second, we find evidence of increased equity markets integration since the early 1990s, in the form of a declining trend in the cross-country dispersion of expected discount factors estimated in each of the European equity markets considered.

Third, we find lack of evidence of a decline in risk profiles for European banks and insurance companies during the period 1990-2004. Importantly, we find that these risk profiles have converged, and that the sensitivity of bank risk profiles to both common real and financial shocks has significantly increased. An interpretation of these findings is that increased synchronization in real activity and advances in financial integration may have reduced the benefits of cross-country diversification.

The remainder of the paper consists of three sections. Section II assesses synchronization in real activity, while Section III considers integration of equity markets. Section IV constructs indicators of system-wide financial risk for a set of systemically important banks and insurance companies in a large set of European countries, and relates the dynamics of these measures to the outcomes of increased real synchronization and advanced financial integration. Section V concludes.

II. Synchronization of Real Activity

We gauge changes in synchronization of real activity by the evolution of the cross-country dispersion in the first and second conditional moments of seasonally adjusted monthly industrial production growth (IPG) for the 10 European countries for which we have complete data for the period 1961:01-2004:12. ³ Subject to an important qualification detailed below, a downward trend in the dynamics of the cross-country variance of IPG and its country-specific volatility may capture increased synchronization in real activity, as this indicates increased correlation of the IPG series in the sample.⁴ A statistical model for such dynamics is obtained as follows.

Let X_it denote IPG in country i at date t. X_it turns out well described by the following E(xponential)GARCH(1,1) model:

\begin{matrix} X_{it} = α_{it} + β_{i} F_{t} + γ X_{it - 1} + h_{it} ε_{it} & (1) \end{matrix}

\begin{matrix} Ln (h_{it}^{2}) = a_{it} + b ε_{it - 1}^{2} + cLn (h_{it - 1}^{2}) & (2) \end{matrix}

The term F_t in the mean Equation (1) is a risk factor common to all countries. The variance Equation (2) describes the evolution of country-specific volatility. As customary, the innovations ε_it are assumed to be i.i.d. and normally distributed with zero mean and unit variance.

We take a weighted average of IPG rates as a proxy measure of the common component of IPG rates. Thus, we set F_t = Σ_i w_itX_it /N. This common component is measured using the time-varying weights proposed by Lumsdaine and Prasad (2003), where $w_{it} \equiv {\tilde{h}}_{it}^{- 1} / \sum_{i} {\tilde{h}}_{it}^{- 1}$ , and variances ${\tilde{h}}_{it}$ are obtained by estimates of model (1)-(3) without the common factor. This specification embeds the assumption that the relative conditional standard deviation is a measure of the degree of commonality of countries’ real fluctuations.

Stock and Watson (2005) have documented a reduction in the volatility of G-7 business cycles, and argued that this has been associated with a reduction in the magnitude of common shocks. Have reductions of volatility and the magnitude of common shocks occurred in Europe? The answer is affirmative. Note that by construction, F_t is described by an EGARCH(1,1) model. As shown in Figure 1, the estimated conditional variance and the residuals of both series appear to exhibit a downward trend. Furthermore, we tested for such a downward trend by estimating the EGARCH model for F_t with a trend both in the mean and variance equations. As shown in Table 1, the coefficient of the time trend in the conditional mean and variance is negative and highly significant.

Table 1.

EGARCH Estimates for the Common Components of IPG

\begin{array}{l} Mean Equation : F_{t} = A_{0} + A_{1} t + A_{2} F_{t - 1} + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

Table 1.

EGARCH Estimates for the Common Components of IPG

\begin{array}{l} Mean Equation : F_{t} = A_{0} + A_{1} t + A_{2} F_{t - 1} + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

	Coefficients	Standard Error	T-Statistic	p-value
Mean Equation
A0	0.33118	0.07129	4.64574	0.00000
A1	-0.00057	0.00018	-3.22896	0.00124
A2	0.03188	0.03612	0.88267	0.37741
Variance Equation
B0	-0.15833	0.19829	-0.79851	0.42457
B1	-0.00267	0.00076	-3.51361	0.00044
B2	0.38005	0.08970	4.23701	0.00002
B3	-0.12188	0.24345	-0.50065	0.61662

Table 1.

EGARCH Estimates for the Common Components of IPG

\begin{array}{l} Mean Equation : F_{t} = A_{0} + A_{1} t + A_{2} F_{t - 1} + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

	Coefficients	Standard Error	T-Statistic	p-value
Mean Equation
A0	0.33118	0.07129	4.64574	0.00000
A1	-0.00057	0.00018	-3.22896	0.00124
A2	0.03188	0.03612	0.88267	0.37741
Variance Equation
B0	-0.15833	0.19829	-0.79851	0.42457
B1	-0.00267	0.00076	-3.51361	0.00044
B2	0.38005	0.08970	4.23701	0.00002
B3	-0.12188	0.24345	-0.50065	0.61662

To obtain a model for the cross-country variance of IPG and its volatility, note that the conditional mean and variance of X_it are given by m_t−1 (X_it) ≡ α_it + β_iE_{t -1}F_t + γX_it−1 and by ${var}_{t - 1} (X_{it}) \equiv β_{i}^{2} σ_{F}^{2} (t) + h_{it}^{2}$ respectively. Assume that the coefficients {α_it,β_i,α_i} are distributed cross-sectionally with means {α_t,β,α} and variances ${σ_{α t}^{2}, σ_{β}^{2}, σ_{a}^{2}}$ , and that covariances among all these random variables, as well as that of X_it-1 and F_t and each of these are approximately nil. Under these assumptions, the cross-sectional variances of m_t-1(X_it) and $h_{it}^{2}$ are given by

\begin{matrix} σ_{X}^{2} (t) \equiv E {(m_{t - 1} (X_{it}) - {Em}_{t - 1} (X_{it}))}^{2} = σ_{α t}^{2} + σ_{β}^{2} {(E_{t - 1} F_{t})}^{2} + γ^{2} σ_{X}^{2} (t - 1) & (3) \end{matrix}

\begin{matrix} σ_{h^{2}}^{2} (t) \equiv E {(h_{it}^{2} - E h_{it}^{2})}^{2} = σ_{a t}^{2} + b^{2} σ_{ε^{2}}^{2} (t - 1) + c^{2} σ_{h^{2}}^{2} (t - 1) & (4) \end{matrix}

Increased synchronization in real activity occurs if $σ_{α t}^{2}$ and/or $σ_{a t}^{2}$ exhibit a declining path. Note that a decline in $σ_{X}^{2} (t)$ exclusively driven by a decline in the magnitude of common shocks ((E_t-1F_t)²) (which we have shown above has occurred) would not necessarily indicate increased integration, since disconnected economies hit by the same shock would exhibit the same decline.

We estimate the following counterpart of model (3)-(4):

\begin{matrix} {\bar{σ}}_{X}^{2} (t) = A_{0} + A_{1} t + A_{2} {(E_{t - 1} F_{t})}^{2} + A_{3} {\bar{σ}}_{X}^{2} (t - 1) + H_{t} η_{t} & (5) \end{matrix}

\begin{matrix} \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} & (6) \end{matrix}

Two measures for ${\bar{σ}}_{X}^{2} (t)$ are used. The first one is the sample cross-sectional variance of IPG rates. The second one is the variance of deviations of IPG of each country from its own trend, obtained by applying the relevant HP filter.⁵ This second measure is used to gauge the extent to which detrended and non-detrended dynamics differ. For both measures, we test whether A₁ and/or B₁ are negative.

Table 2 reports estimates of model (5)-(6). Both the trend coefficients in the mean equation (6) and the variance equation (7) are negative and significant. Moreover, the estimates of these coefficients remain virtually unchanged when the cross-sectional variance of deviations of IPG from trend is used. Most important, the relevant trend coefficients remain negative but become highly significant. This indicates that business cycle convergence is the primary driver of the decreasing dispersion of IPG rates. Notably, increased synchronization has occurred not only in the form of increased correlation of IPG fluctuations across countries, but also in terms of increased correlation of their (country-specific) volatilities.

Table 2.

EGARCH Estimates for Cross-Country Variances of IPG and De-Trended IPG

\begin{array}{l} Mean Equation : {\bar{σ}}_{X}^{2} (t) = A_{0} + A_{1} t + A_{2} {(E_{t - 1} F_{t})}^{2} + A_{3} {\bar{σ}}_{X}^{2} (t - 1) + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

Table 2.

EGARCH Estimates for Cross-Country Variances of IPG and De-Trended IPG

\begin{array}{l} Mean Equation : {\bar{σ}}_{X}^{2} (t) = A_{0} + A_{1} t + A_{2} {(E_{t - 1} F_{t})}^{2} + A_{3} {\bar{σ}}_{X}^{2} (t - 1) + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

	Coeff	Std Error	T-Stat	Signif
*******************************************************************************
*Mean Equation*
A0	5.832508368	2.292758087	2.54388	0.01096278
A1	-0.007229695	0.004513876	-1.60166	0.10923082
A2	2.918317416	0.859319970	3.39608	0.00068359
A3	0.085131670	0.084094149	1.01234	0.31137665
*Variance Equation*
B0	4.814365906	1.578460677	3.05004	0.00228812
B1	-0.006901057	0.002589526	-2.66499	0.00769909
B2	0.588642141	0.152287022	3.86535	0.00011093
B3	-0.026846689	0.203262590	-0.13208	0.89492192
	De-trended
*Mean Equation*
A0	5.303880223	0.950224807	5.58171	0.00000002
A1	-0.006814521	0.001873768	-3.63680	0.00027605
A2	7.069362688	3.455381054	2.04590	0.04076625
A3	0.233748490	0.069387844	3.36872	0.00075517
*Variance equation*
B0	4.617319622	0.890776496	5.18348	0.00000022
B1	-0.008062957	0.001681891	-4.79398	0.00000164
B2	0.454190765	0.139146278	3.26412	0.00109803
B3	0.226598798	0.114381371	1.98108	0.04758215

Table 2.

EGARCH Estimates for Cross-Country Variances of IPG and De-Trended IPG

\begin{array}{l} Mean Equation : {\bar{σ}}_{X}^{2} (t) = A_{0} + A_{1} t + A_{2} {(E_{t - 1} F_{t})}^{2} + A_{3} {\bar{σ}}_{X}^{2} (t - 1) + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

	Coeff	Std Error	T-Stat	Signif
*******************************************************************************
*Mean Equation*
A0	5.832508368	2.292758087	2.54388	0.01096278
A1	-0.007229695	0.004513876	-1.60166	0.10923082
A2	2.918317416	0.859319970	3.39608	0.00068359
A3	0.085131670	0.084094149	1.01234	0.31137665
*Variance Equation*
B0	4.814365906	1.578460677	3.05004	0.00228812
B1	-0.006901057	0.002589526	-2.66499	0.00769909
B2	0.588642141	0.152287022	3.86535	0.00011093
B3	-0.026846689	0.203262590	-0.13208	0.89492192
	De-trended
*Mean Equation*
A0	5.303880223	0.950224807	5.58171	0.00000002
A1	-0.006814521	0.001873768	-3.63680	0.00027605
A2	7.069362688	3.455381054	2.04590	0.04076625
A3	0.233748490	0.069387844	3.36872	0.00075517
*Variance equation*
B0	4.617319622	0.890776496	5.18348	0.00000022
B1	-0.008062957	0.001681891	-4.79398	0.00000164
B2	0.454190765	0.139146278	3.26412	0.00109803
B3	0.226598798	0.114381371	1.98108	0.04758215

To gauge the approximate timing of increased synchronization, we estimated model (1)-(2) for each country with dummies for different decades in both the mean and variance equations. As shown in Table 3, increased synchronization in the form of a decline of the cross-country variance of IPG rates started to occur in the early 1980s By contrast, the crosscountry dispersion in the variance of IPG volatility does not exhibit a decline. This result, coupled with the significant decline found earlier, suggests that convergence in volatility is either driven by a subset of countries in the sample or has occurred at different points in time in some countries, or both.

Table 3.

Country-by-Country EGARCH Estimates for IPG

\begin{array}{l} Mean Equation : X_{it} = D 60 + D 70 + D 80 + D 9004 + β_{i} {\bar{X}}_{t} + γ X_{it - 1} + h_{it} ε_{it} \\ Variance Equation : Ln (h_{it}^{2}) = D 60 + D 70 + D 80 + D 9004 + b ε_{it - 1}^{2} + cLn (h_{it - 1}^{2}) \end{array}

Table 3.

Country-by-Country EGARCH Estimates for IPG

\begin{array}{l} Mean Equation : X_{it} = D 60 + D 70 + D 80 + D 9004 + β_{i} {\bar{X}}_{t} + γ X_{it - 1} + h_{it} ε_{it} \\ Variance Equation : Ln (h_{it}^{2}) = D 60 + D 70 + D 80 + D 9004 + b ε_{it - 1}^{2} + cLn (h_{it - 1}^{2}) \end{array}

	Mean Equation				Variance Equation
	D60	D70	D80	D9004	D60	D70	D80	D9004
Country
Austria	-0.02	0.36	0.13	0.36	0.76	-1.09	-1.15	-0.82
Belgium	0.09	0.02	0.06	0.13	-0.14	0.27	0.23	0.21
France	-0.01	0.24	0.07	0.07	1.27	-2.35	-2.72	-3.01
Germany	0.21	-0.07	0.03	0.04	0.44	-0.83	-0.58	-1.19
Greece	0.70	0.63	0.22	-0.04	0.51	-0.53	-0.30	-0.29
Italy	0.16	0.29	0.08	0.00	0.13	0.15	-0.21	-0.90
Netherlands	0.50	0.14	0.12	0.10	-0.17	0.44	0.91	0.62
Portugal	0.35	0.67	0.41	0.10	0.61	-0.56	-0.84	-0.89
Sweden	0.17	0.04	0.08	0.20	1.10	-1.68	-0.67	-2.30
U.K.	-0.01	0.04	0.07	-0.03	-0.83	1.59	1.00	0.40
Mean	0.21	0.24	0.13	0.09	0.37	-0.46	-0.44	-0.82
Variance	0.06	0.07	0.01	0.01	0.40	1.28	1.15	1.34

Table 3.

Country-by-Country EGARCH Estimates for IPG

\begin{array}{l} Mean Equation : X_{it} = D 60 + D 70 + D 80 + D 9004 + β_{i} {\bar{X}}_{t} + γ X_{it - 1} + h_{it} ε_{it} \\ Variance Equation : Ln (h_{it}^{2}) = D 60 + D 70 + D 80 + D 9004 + b ε_{it - 1}^{2} + cLn (h_{it - 1}^{2}) \end{array}

	Mean Equation				Variance Equation
	D60	D70	D80	D9004	D60	D70	D80	D9004
Country
Austria	-0.02	0.36	0.13	0.36	0.76	-1.09	-1.15	-0.82
Belgium	0.09	0.02	0.06	0.13	-0.14	0.27	0.23	0.21
France	-0.01	0.24	0.07	0.07	1.27	-2.35	-2.72	-3.01
Germany	0.21	-0.07	0.03	0.04	0.44	-0.83	-0.58	-1.19
Greece	0.70	0.63	0.22	-0.04	0.51	-0.53	-0.30	-0.29
Italy	0.16	0.29	0.08	0.00	0.13	0.15	-0.21	-0.90
Netherlands	0.50	0.14	0.12	0.10	-0.17	0.44	0.91	0.62
Portugal	0.35	0.67	0.41	0.10	0.61	-0.56	-0.84	-0.89
Sweden	0.17	0.04	0.08	0.20	1.10	-1.68	-0.67	-2.30
U.K.	-0.01	0.04	0.07	-0.03	-0.83	1.59	1.00	0.40
Mean	0.21	0.24	0.13	0.09	0.37	-0.46	-0.44	-0.82
Variance	0.06	0.07	0.01	0.01	0.40	1.28	1.15	1.34

Summing up, synchronization in real activity appears to have increased since the early 1980s. It has been primarily driven by increases in business cycle synchronization, and is not the mechanical outcome of a decline in the magnitude of common shocks.

III. Equity Markets Integration

One difficulty in testing advances in equity market integration rests on disentangling pricing effects from the effects of fundamental shocks. Here we tackle this problem by using a version of the methodology proposed by Flood and Rose (2005) (FR henceforth), which exploits the pricing of “idiosyncratic risk,” as opposed to systematic risk. As detailed below, an advantage of this methodology is that it does not require taking a stand on a particular asset pricing model.

Consider the standard intertemporal asset pricing equation:

\begin{matrix} p_{t}^{j} = E_{t} (m_{t + 1} x_{t + 1}^{j}) = {COV}_{t} (m_{t + 1}, x_{t + 1}^{j}) + E_{t} m_{t + 1} E_{t} x_{t + 1}^{j} & (7) \end{matrix}

where $p_{t}^{j}$ is the price of asset j at date t, E_t is the expectations operator, COV_t is the covariance operator conditional on information available at date t, and m_t+1 is the rate used to discount the income $x_{t + 1}^{j}$ accruing to the holder of the asset at date t +1. Equation (8) can be re-written as:

\begin{matrix} x_{t + 1}^{j} = δ_{t} (p_{t}^{j} - {COV}_{t} (m_{t + 1}, x_{t + 1}^{j})) + ε_{t + 1}^{j}, & (8) \end{matrix}

where δ_t ≡ 1/E_tm_t+1 is the inverse of the expected discount factor (IEDF), and $ε_{t + 1}^{j} \equiv x_{t + 1}^{j} - E_{t} x_{t + 1}^{j}$ is a prediction error orthogonal to information available at date t. If all assets traded in a given market are discounted at the same rate, then such asset market is said to be integrated.

As our focus is not on assessing integration per se, but changes in integration, we take the dynamics of the cross-country variance of estimated IEDFs across different equity markets as our measure of changes in integration. If markets become increasingly integrated, this variance should decline. If they moved towards perfect integration, this variance should converge to zero.

Estimates of IEDFs are obtained as follows. Let the time series vector Z_t denote the set of factors that capture all systematic components of (log) returns. Then, $p_{t}^{j} = p_{t - 1}^{j} \exp (β^{'} Z_{t} + v_{t}^{j})$ , where $v_{t}^{j}$ is the idiosyncratic part of asset j return. Following FR, define the “systematic price” ${\tilde{p}}_{t}^{j}$ as the value of $p_{t}^{j}$ conditional on idiosyncratic information available at date t set to zero. Thus, ${\tilde{p}}_{t}^{j} \equiv p_{t - 1}^{j} \exp (β^{'} Z_{t})$ . Normalizing equation (8) by such price yields:

\begin{matrix} x_{t + 1}^{j} / {\tilde{p}}_{t}^{j} = δ_{t} (p_{t}^{j} / {\tilde{p}}_{t}^{j}) - δ_{t} {COV}_{t} (m_{t + 1}, x_{t + 1}^{j} / {\tilde{p}}_{t}^{j})) + ε_{t + 1}^{j} = δ_{t} \exp (v_{t}^{i}) + u_{t + 1}^{j} & (9), \end{matrix}

where $u_{t + 1}^{j} \equiv ε_{t + 1}^{j} - δ_{t} {COV}_{t} (m_{t + 1}, x_{t + 1}^{j} / {\tilde{p}}_{t}^{j})$ .

Estimates of the IEDF δ_t can thus be obtained through regressions (9) by means of the following two-step procedure. In the first step, the “systematic price” ${\tilde{p}}_{t}^{j}$ of value weighted industry portfolios for each stock market is estimated by OLS, using the set of principal components of returns to factor out systematic risks. Their number is identified by standard statistical procedures. In the second step, returns are normalized with the estimated systematic price ${\hat{p}}_{t}^{j}$ , and the IEDF estimates for each equity market and date are obtained by means of OLS cross-sectional regressions of the type:

\begin{matrix} x_{t + 1}^{j} / {\hat{p}}_{t}^{j} = c_{t} + δ_{t} \exp (v_{t}^{i}) + u_{t + 1}^{j} & (10) \end{matrix}

As stressed by Marshall (2005), the constant is needed to control for date-specific (aggregate) shocks that can still be embedded into the IEDF estimates. This estimation procedure was applied to the 13 sectoral portfolios and 9 countries for which monthly data from Datastream were available for the 1975:01-2004:12 period. ⁶

All IEDF estimates exhibit correlation with future returns and first-order autocorrelation not significantly different from zero. They also exhibit high volatility, but standard tests reject the hypothesis of ARCH effects. It turns out that the dynamics of the IEDFs for country i at date t, denoted by M_it, is well represented by a simple random walk with drift. Thus, we gauge whether equity markets integration has advanced by testing the significance of a time trend in the following equation for the cross-country variance of IEDFs, denoted by ${\bar{σ}}_{M}^{2} (t)$ :

\begin{matrix} {\bar{σ}}_{M}^{2} (t) = A_{0} + A_{1} t + η_{t} & (11) \end{matrix}

Table 4 reports estimates of (11), specified with a time trend as well as with a set of decade-long dummies. As shown in Panel A, the trend coefficient is negative and significant. As shown in Panel B, increased integration appears to have occurred approximately since the early 1990s. This is also confirmed by simple tests of differences in slopes associated with decade-long dummies. As shown in Panel C, these results are further validated by estimated time trends for the IEDF for each country, since the cross-country variance of the decade-long dummies exhibits a decline from the early 1990s.

Table 4.

Dependent Variables: Cross-Country Variance of IEDFs and Country IEDFs

\begin{array}{l} Panel A : {\bar{σ}}_{M}^{2} (t) = A_{0} + A_{1} t + η_{t} \\ Panels B : {\bar{σ}}_{M}^{2} (t) = D 702 + D 801 + D 802 + D 901 + D 902 + D 00 + η_{t} \\ Panels C : {EDF}_{i} (t) = D 702 + D 801 + D 802 + D 901 + D 902 + D 00 + η_{t} \end{array}

Table 4.

Dependent Variables: Cross-Country Variance of IEDFs and Country IEDFs

\begin{array}{l} Panel A : {\bar{σ}}_{M}^{2} (t) = A_{0} + A_{1} t + η_{t} \\ Panels B : {\bar{σ}}_{M}^{2} (t) = D 702 + D 801 + D 802 + D 901 + D 902 + D 00 + η_{t} \\ Panels C : {EDF}_{i} (t) = D 702 + D 801 + D 802 + D 901 + D 902 + D 00 + η_{t} \end{array}

		Coeff	Std Error	T-Stat	Signif
**************************************** ***********************************
Panel A
	A0		0.238763917	0.032929671	7.25072	0.00000000
	A1		-0.000240336	0.000106836	-2.24957	0.02447599
Panel B
	D702		0.1727069830	0.0193907237	8.90668	0.00000000
	D801		0.2497064481	0.0464092453	5.38053	0.00000007
	D802		0.1860275694	0.0201381090	9.23759	0.00000000
	D901		0.1888773756	0.0223010268	8.46945	0.00000000
	D902		0.1544279148	0.0150023740	10.29357	0.00000000
	D00		0.1431732182	0.0150554668	9.50972	0.00000000
Panel C
		d702	d801	d802	d901	d902	d00
AUSTRIA		1.20645	1.10617	1.29210	1.16723	1.19304	1.20955
BELGIUM		1.13181	1.17612	1.22143	1.22645	1.17590	1.26687
DENMARK		1.11991	1.27833	1.22076	1.23424	1.26985	1.15297
FRANCE		1.17015	1.19635	1.24437	1.28162	1.22491	1.20120
GERMANY		1.30347	1.25389	1.14590	1.15146	1.12503	1.13513
IRELAND		1.05380	1.03319	1.27659	1.26040	1.28902	1.19700
ITALY		1.30551	1.26973	1.27198	1.28658	1.34620	1.24998
NETHERLANDS		1.20677	1.20968	1.23937	1.24176	1.25930	1.15309
U.K.		1.23885	1.25970	1.13021	1.20395	1.27444	1.26789
	Mean	1.19297	1.19813	1.22697	1.22819	1.23974	1.20374
	Variance	0.00704	0.00684	0.00315	0.00221	0.00449	0.00252

Table 4.

Dependent Variables: Cross-Country Variance of IEDFs and Country IEDFs

\begin{array}{l} Panel A : {\bar{σ}}_{M}^{2} (t) = A_{0} + A_{1} t + η_{t} \\ Panels B : {\bar{σ}}_{M}^{2} (t) = D 702 + D 801 + D 802 + D 901 + D 902 + D 00 + η_{t} \\ Panels C : {EDF}_{i} (t) = D 702 + D 801 + D 802 + D 901 + D 902 + D 00 + η_{t} \end{array}

		Coeff	Std Error	T-Stat	Signif
**************************************** ***********************************
Panel A
	A0		0.238763917	0.032929671	7.25072	0.00000000
	A1		-0.000240336	0.000106836	-2.24957	0.02447599
Panel B
	D702		0.1727069830	0.0193907237	8.90668	0.00000000
	D801		0.2497064481	0.0464092453	5.38053	0.00000007
	D802		0.1860275694	0.0201381090	9.23759	0.00000000
	D901		0.1888773756	0.0223010268	8.46945	0.00000000
	D902		0.1544279148	0.0150023740	10.29357	0.00000000
	D00		0.1431732182	0.0150554668	9.50972	0.00000000
Panel C
		d702	d801	d802	d901	d902	d00
AUSTRIA		1.20645	1.10617	1.29210	1.16723	1.19304	1.20955
BELGIUM		1.13181	1.17612	1.22143	1.22645	1.17590	1.26687
DENMARK		1.11991	1.27833	1.22076	1.23424	1.26985	1.15297
FRANCE		1.17015	1.19635	1.24437	1.28162	1.22491	1.20120
GERMANY		1.30347	1.25389	1.14590	1.15146	1.12503	1.13513
IRELAND		1.05380	1.03319	1.27659	1.26040	1.28902	1.19700
ITALY		1.30551	1.26973	1.27198	1.28658	1.34620	1.24998
NETHERLANDS		1.20677	1.20968	1.23937	1.24176	1.25930	1.15309
U.K.		1.23885	1.25970	1.13021	1.20395	1.27444	1.26789
	Mean	1.19297	1.19813	1.22697	1.22819	1.23974	1.20374
	Variance	0.00704	0.00684	0.00315	0.00221	0.00449	0.00252

In sum, European equity market integration appears to have advanced since the early 1990s. This result is consistent with the conjectures advanced by Adjaouté and Danthine (2004), and with the finding of Baele and others (2004), who document European stock returns and volatility as increasingly affected by ”common” European shocks.

IV. Systemic Risk and Integration

A. Measures of Systemic Risk

Our indicators of systemic risk are distance-to-default (DD) measures of “portfolios” of sets of publicly traded, systemically important European banks and insurance companies. These measures are based on the structural valuation model of Black and Scholes (1973) and Merton (1974) (BSM hereafter).⁷ In the BSM model, the portfolio’s equity is viewed as a call option on the portfolio’s assets, with strike price equal to the current book value of total liabilities. When the value of the portfolio’s assets is less than the strike price, its equity value is zero. The market value of assets is not observable, but can be estimated using equity values and accounting measures of liabilities.

Under the assumption that asset values follow a lognormal process, the DD of portfolio of N firms (indexed by i) is given by:

{DD}_{t} = \frac{Ln (V_{t}^{P} / L_{t}^{P}) + (μ_{P} - 0.5 σ_{P}^{2})}{σ_{P}}

where $V^{P} = \sum_{i} V_{t}^{i}$ and $L_{t}^{P} = \sum_{i = 1}^{N} L_{t}^{i}$ are the total value of assets and liabilities, respectively. The mean and variance of the portfolio are respectively given by $μ_{P} = \sum_{i} w_{t}^{i} μ^{i}$ and $σ_{P} = \sum_{i} \sum_{i} w_{t}^{i} w_{t}^{J} σ_{ij}$ , where $w_{t}^{i} = V_{t}^{i} / \sum_{i} V_{t}^{i}$ and σ _ij is the asset return covariance of firm i and j. Thus, the “portfolio” DD embeds the structure of risk interdependencies among firms. “Default” at date t +1 occurs when $V_{t}^{P} < L_{t}^{P}$ . Thus, the DD indicates how many standard deviations $Ln (V_{t}^{P} / L_{t}^{P})$ has to deviate from its mean in order for default to occur. Since $V_{t}^{P} = L_{t}^{P} + E_{t}^{P}$ , where $E_{t}^{P}$ is the value of equity, declines in $V_{t}^{P} / L_{t}^{P}$ imply declines in capitalization $(E_{t}^{P} / L_{t}^{P})$ .

Lower (higher) levels of the “portfolio” DD imply a higher (lower) probability of firms’ joint failure. Since positive and negative variations in the DD of individual firms are allowed to offset each other, the DD of a portfolio is always higher than the (weighted) sum of the DDs of the individual firms.⁸ As a result, the probability of “failure” associated with the “portfolio” DD is always lower than that associated with the actual probability of joint failures of sets of firms in the portfolio. Thus, the “portfolio” DD can be viewed as tracking the evolution of a lower bound to the joint probabilities of failure of the firms composing a portfolio. For this reason, and because of the distributional assumptions underlying its derivation,⁹ the portfolio DD measure is a conservative measure of systemic risk, that is, it is likely to underestimate systemic risk.

We constructed portfolios for all banks and insurance companies included in the relevant Datastream indexes during the period 1991.01-2004.12.¹⁰ The monthly DD measures were estimated using the methodology described in Vassalou and Xing (2004). At each date, the value of asset, the return on assets and its volatility were derived using the option valuation formula of the BSM model, using one year of daily equity return data preceding the estimation date, and the accounting value of liabilities for the relevant year. This procedure was repeated for each month, and the relevant estimates were used to construct the “portfolio” DD measures at a monthly frequency. As shown by Duan, Gauthier and Simonato (2004), this procedure yields maximum likelihood estimates of implied asset returns and volatility.

B. Trends in Systemic Risk

Figure 2 depicts the evolution of bank systemic risk measures and their trend computed applying an HP filter. Over the period 1990-2004, bank systemic risk does not appear to have declined.

One explanation of these dynamics rests on the changes in the composition of sources of income that have occurred during the period examined. European banks exhibited a substantial increase in noninterest income in the past decade (ECB, 2004). As documented in De Nicolò and others (2005), the volatility of noninterest income growth was significantly larger than that of interest income growth at large banks since 1997. Moreover, the correlation between interest and noninterest income growth has been high (0.79 for the EU-15), indicating decreasing diversification benefits across traditional and nontraditional business lines. As banks’ earnings have increasingly relied on income generated through financial market activity, most large banks have experienced significant increases in asset return volatility since the early 1990s. Substantial increases in capitalization and improvements in returns have occurred as well, but they have not been sufficient to offset increases in asset return volatility. Thus, large European banks may have supported higher-risk/higher-return investments with larger capital buffers. Yet, risk-adjusted asset returns have not increased, and overall risk profiles have not declined. Using a different systemic risk metrics but a smaller sample of European financial institutions, Hartmann, Straetmans and de Vries (2005) provide further evidence consistent with our results.

These findings, and the attendant interpretation, do not characterize only European banks. A similar pattern appears to have characterized the evolution of risk profiles of U.S. banks as well, as documented by De Nicolò, Hayward, and Vir Bhatia (2004) for U.S. large complex banking groups, and by Stiroh (2004), Hartmann, Straetmans, and de Vries (2005), Stiroh and Rumble (2006), and Houston and Stiroh (2006) for a large set of U.S. bank holding companies, using different systemic risk metrics. Using a term aptly coined by Stiroh and Rumble, the “dark side of diversification” has materialized for U.S. banks as well during the same period.

The dynamics of systemic risk profiles for insurance companies is depicted in Figure 3. While in most European countries these dynamics are similar to those of banks, cross-country heterogeneity is more marked. In some instances, the systemic risk measures indicate that a decline has occurred recently.

C. Convergence of Systemic Risk Measures

Hartmann, Straetmans and de Vries (2005) provide evidence of increased systemic risk in the 1990s for a set of large European (and U.S.) banks in the form of increased correlation of measures of extreme realizations of bank stock excess returns. Unlike these measures, our measures of systemic risk are directly related to firms’ joint probability of failures, since they take into account not only the evolution of bank returns, but also their volatility and firms’ capitalization. Do our measures of systemic risk point at increased correlation in the form of convergence? The answer is affirmative.

We test convergence in both levels and volatility of our systemic risk measures. Denote the sample cross-sectional variance of “portfolio” DDs by ${\bar{σ}}_{DD}^{2} (t)$ . We posit and estimate the following EGARCH-type model for ${\bar{σ}}_{DD}^{2} (t)$ :

\begin{matrix} σ_{DD}^{2} (t) = A_{0} + A_{1} t + A_{2} Y_{t}^{2} + A_{3} {\bar{σ}}_{DD}^{2} (t - 1) + H_{t} η_{t} & (12) \end{matrix}

\begin{matrix} \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} & (13) \end{matrix}

As remarked previously, assessing convergence requires controlling for the magnitude of common shocks, since financial firms operating in totally disconnected economies may exhibit increased comovements in their risk profiles just because the economies in which they operate are hit by the same shock. This motivates the introduction of a proxy measure of “common” risks in the mean equation (12), which is simply measured by the average of each country financial sector DD, Y_t = ∑_i DD_it/N. Convergence in systemic risk profiles of the banks and insurance sectors is assessed by testing whether the coefficients A₁ and/or B₁ are negative.

Table 5 reports estimates of model (12)-(13). Bank systemic risk profiles exhibit convergence in both mean and the variance, since the trend coefficients are negative and highly significant in both the mean and variance equations. These results provide strong support to the conjecture that risk interdependencies have increased remarkably. By contrast, insurance risk profiles do not exhibit convergence either in the mean or in the variance. In sum, convergence in systemic risk profiles has occurred for banks, but there is no evidence that it has occurred for insurance companies.

Table 5.

Dependent Variables: Cross-Country Variance of Portfolios’ DDs

\begin{array}{l} Mean Equation : {\bar{σ}}_{DD}^{2} (t) = A_{0} + A_{1} t + A_{2} Y_{t}^{2} + A_{3} {\bar{σ}}_{DD}^{2} (t - 1) + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

Table 5.

Dependent Variables: Cross-Country Variance of Portfolios’ DDs

\begin{array}{l} Mean Equation : {\bar{σ}}_{DD}^{2} (t) = A_{0} + A_{1} t + A_{2} Y_{t}^{2} + A_{3} {\bar{σ}}_{DD}^{2} (t - 1) + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

	Coeff	Std Error	T-Stat	Signif
******************************************************************************
	BANKS
*Mean Equation*
A0	2.130758233	0.796056529	2.67664	0.00743641
A1	-0.004987358	0.001985143	-2.51234	0.01199330
A2	0.005940051	0.003149941	1.88577	0.05932652
A3	0.858994960	0.039575910	21.70500	0.00000000
*Variance Equation*
B0	6.070978192	2.226300055	2.72694	0.00639254
B1	-0.017596387	0.005921549	-2.97158	0.00296267
B2	0.220411658	0.116434721	1.89301	0.05835703
B3	-0.224892329	0.226904246	-0.99113	0.32162055
*******************************************************************************
	INSURANCE
*Mean Equation*
A0	1.103339887	2.667662223	0.41360	0.67916850
A1	0.003887521	0.004944093	0.78630	0.43169399
A2	0.057381532	0.021335058	2.68954	0.00715502
A3	0.719393328	0.103899125	6.92396	0.00000000
*Variance Equation*
B0	9.234640107	9.860517900	0.93653	0.34900198
B1	-0.021483407	0.023667905	-0.90770	0.36403563
B2	0.010679293	0.393523834	0.02714	0.97834998
B3	-0.470348664	0.936137459	-0.50244	0.61536123

Table 5.

Dependent Variables: Cross-Country Variance of Portfolios’ DDs

\begin{array}{l} Mean Equation : {\bar{σ}}_{DD}^{2} (t) = A_{0} + A_{1} t + A_{2} Y_{t}^{2} + A_{3} {\bar{σ}}_{DD}^{2} (t - 1) + H_{t} η_{t} \\ Variance Equation : \log H_{t}^{2} = B_{0} + B_{1} t + B_{2} η_{t - 1}^{2} + B_{3} \log H_{t - 1}^{2} \end{array}

	Coeff	Std Error	T-Stat	Signif
******************************************************************************
	BANKS
*Mean Equation*
A0	2.130758233	0.796056529	2.67664	0.00743641
A1	-0.004987358	0.001985143	-2.51234	0.01199330
A2	0.005940051	0.003149941	1.88577	0.05932652
A3	0.858994960	0.039575910	21.70500	0.00000000
*Variance Equation*
B0	6.070978192	2.226300055	2.72694	0.00639254
B1	-0.017596387	0.005921549	-2.97158	0.00296267
B2	0.220411658	0.116434721	1.89301	0.05835703
B3	-0.224892329	0.226904246	-0.99113	0.32162055
*******************************************************************************
	INSURANCE
*Mean Equation*
A0	1.103339887	2.667662223	0.41360	0.67916850
A1	0.003887521	0.004944093	0.78630	0.43169399
A2	0.057381532	0.021335058	2.68954	0.00715502
A3	0.719393328	0.103899125	6.92396	0.00000000
*Variance Equation*
B0	9.234640107	9.860517900	0.93653	0.34900198
B1	-0.021483407	0.023667905	-0.90770	0.36403563
B2	0.010679293	0.393523834	0.02714	0.97834998
B3	-0.470348664	0.936137459	-0.50244	0.61536123

D. The Sensitivity of Systemic Risk Measures to Real and Financial Shocks

Have increased real synchronization and advances in financial markets integration affected the dynamics of systemic risk profiles? We address this question by assessing whether the sensitivity of risk profiles of banks and insurance companies to both real and financial shocks common to all countries considered has significantly increased. The impact of such shocks is simply proxied by the common component of real activity estimated previously, and by a value-weighted index of European stock market returns.

An increased sensitivity of the dynamics of systemic risk profiles to the common component of real activity would indicate a greater impact of common real shocks to institutions’ exposures, in part due to increased synchronization of real activity. Similarly, an increased sensitivity of these profiles to common financial shocks may be a result of increased financial integration. This is supported by our results on convergence in European IEDFs, as well as by the increased correlation of European stock market returns documented by Bekaert, Hodrick, and Zhang (2005),

Changes in these sensitivities are gauged as follows. Let $ρ_{t}^{i} \equiv {DD}_{it} - {DD}_{it - 1}$ denote monthly changes in the DD for banks and insurance sectors in country i. Standard tests for $ρ_{t}^{i}$ reject ARCH effects for both banks and insurance companies. We estimate a version of the following simple model for $ρ_{t}^{i}$ in each country:

\begin{matrix} ρ_{t}^{i} = α + β_{1} (t) F_{t} + β_{2} (t) R_{t} + γ ρ_{t - 1}^{i} + ε_{t} & (14), \end{matrix}

where F_t denotes the common component of real activity estimated previously, and R_t denotes the Datastream European stock market index. Variables F_t and R_t are likely to capture distinct sources of real and financial shocks, since their contemporaneous correlation was only 0.06 during the entire 1991.1-2004.12 period.

We wish to test whether the coefficients β₁(t) and β₂ (t) have increased. Given the short length of the sample, regressions (14) are simply estimated allowing these coefficients to differ for three periods approximately four-years long: 1991.1-1994.12, 1995.1-1998.12, and 1999.1-2004.12.

Table 6 reports regressions (14) for banks, where the coefficients for the three periods considered are denoted by β_1j and β_2j, j=1,2,3, and the relevant indicator functions I_j, j=1,2,3. The sensitivity of bank systemic risk measures to common shocks has increased in all countries but Austria, Greece, Italy, and Norway, where it has remained approximately constant. In most cases, the relevant coefficients have turned from negative to positive, and significantly so. By contrast, the sensitivity of bank risk profiles to financial shocks has changed differently across countries, increasing in Austria, France, Germany, Greece, Italy, the Netherlands, and Portugal, but decreasing in Belgium, Ireland, Norway, Spain, Sweden, and the United Kingdom.

Table 6.

Dependent Variable: Banks DDs Changes

Models : (1) Δ ρ_{t}^{i} = α + β_{1} F_{t} + β_{2} R_{t} + γ Δ ρ_{t - 1}^{i} + ε_{t}^{i}; (2) Δ ρ_{t}^{i} = α + \sum_{j = 1}^{3} β_{1 j} I_{j} F_{t} + \sum_{j = 1}^{3} β_{2 j} I_{j} R_{t} + γ Δ ρ_{t - 1}^{i} + ε_{t}^{i}

The results for insurance companies are similar to those for banks with respect to common real risks, but exhibit heterogeneity with respect to financial risks. As shown in Table 7, the sensitivity of insurance systemic risk measures to common real shocks has increased in all countries but Austria, Denmark, Italy, and the United Kingdom, where it has remained approximately constant. On the other hand, the sensitivity to financial shocks has increased in France Germany, Italy, and the Netherlands, but has decreased in Denmark, Ireland, and Norway.

Table 7.

Dependent Variable: Insurance DDs Changes

Models : (1) Δ ρ_{t}^{i} = α + β_{1} F_{t} + β_{2} R_{t} + γ Δ ρ_{t - 1}^{i} + ε_{t}^{i}; (2) Δ ρ_{t}^{i} = α + \sum_{j = 1}^{3} β_{1 j} I_{j} F_{t} + \sum_{j = 1}^{3} β_{2 j} I_{j} R_{t} + γ Δ ρ_{t - 1}^{i} + ε_{t}^{i}

Summing up, we find a significant increase in the sensitivity of bank systemic risk profiles to real shocks. In no country have such sensitivities have declined. We also find a significant increase in the sensitivities to financial shocks for institutions located in all large countries but Spain and the United Kingdom. The results for insurance risk profile mirror those for banks with regard to increased exposures to common real shocks, but exhibit heterogeneity with respect to common financial shocks. However, the sensitivity of insurance risk profiles to both common real and financial shocks has increased for most institutions located in the largest countries.

V. Conclusion

In this paper we have assessed whether synchronization in real activity has increased and equity market integration has advanced, and have examined one dimension through which systemic risk profiles at large, systemically important European banks and insurance companies may have been affected by changes in real synchronization and financial markets integration.

We found increased synchronization of real activity starting in the early 1980s, and increased integration in the equities markets starting in the early 1990s. We also found that our measures of bank systemic risk profiles have not declined over the last 15 years, and have converged both in levels and volatility. This evidence suggests that banks may have opted for investment strategies targeting higher-risk/higher-expected returns.

Furthermore, we assessed whether the sensitivity of bank and insurance systemic risk profiles to common real and financial shocks have changed during this period. We found that for banks, the sensitivity of risk profiles to common real and financial shocks has significantly increased in most countries. The sensitivity of risk profiles of insurance companies to common real shocks has significantly increased as well in most countries, while their sensitivity to common financial shocks appears to have increased only for institutions located in the large continental European countries.

Overall, these findings suggest that increased real synchronization and advances in financial integration may not have necessarily resulted in heightened financial stability. Thus, enhanced monitoring of increased interdependencies in risk profiles among institutions and through markets appears an important task European supervisors may face as integration progresses.

Appendix

Table 8.

Banks and Insurance Companies

Table 8.

Banks and Insurance Companies

		Total Assets, end 2004 (In millions of euro)
Austria		Germany		Italy
*Banks*		*Banks*		*Banks*
BANK AU.CREDITANSTALT	145,680	BANKGESELLSCHAFT BERLIN	152,041	UNICREDITO ITALIANO	264,791
ERSTE BANK	139,390	BAYER.HYPO-UND-VBK.	463,496	SAN PAOLO IMI	209,627
*Insurance*		COMMERZBANK	419,674	CAPITALIA	131,163
GENERALI HOLDING	10,430	DEUTSCHE BANK	836,368	BANCA INTESA	273,181
UNIQA	15,125	*Insurance*		BANCA MONTE DEI PASCHI	128,662
WIENER STAEDT VZ	12,525	ALLIANZ	958,579	BCA.NAZ.LAVORO	78,111
		AMB GENERALI HDG.	85,632	*Insurance*
Belgium		AXA KONZERN	38,438	CATTOLICA ASSICURAZIONI	14,773
*Banks*		AXA VERSICHERUNG	7,219	FONDIARIA-SAI	31,381
ALMANIJ	259,629 *	DBV-WINTERTHUR HOLDING	22,398	GENERALI	252,542
DEXIA	349,463	ERGO VERSICHERUNG	108,988	MILANO ASSIC.	9,804
FORTIS (BRU)	521,524	GERLING	4,327	PREMAFIN-HLDG.DI PART.	31,493
KBC BKVS.HDG.	225,587	HANNOVER RUCK.	30,615	RAS	66,229
		KOELN.RUCK.	9,390	UNIPOL	32,201
Denmark		KOELN.VERWALT.GES ELL.	69
*Banks*		MLP	2,378	Netherlands
DANSKE BANK	248,949	MUNCH.RUCK.REGD.	203,501	*Banks*
JYSKE BANK	16,687	WUESTENROT & WUERTT.	54,380	ABN AMRO HOLDING	607,263
SYDBANK	10,453			FORTIS (AMS)	521,524
*Insurance*		Greece		*Insurance*
ALM BRAND	2,897	*Banks*		ING GROEP CERTS.	866,201
CODAN	5,358	ALPHA BANK	32,917
TOPDANMARK	4,516	BANK OF GREECE	32,810	Norway
		BANK OF PIRAEUS	16,591	*Banks*
France		EFG EUROBANK ERGASIAS	31,760	DNB NOR	86,097
*Banks*		EMPORIKI BK.OF GREECE	18,236	SPAREBANKEN ROGALAND	7,152
BNP PARIBAS	905,001	NAT.BK.OF GREECE	52,877	*Insurance*
CREDIT AGRICOLE	814,654	*Insurance*		STOREBRAND	20,871
SOCIETE GENERALE	600,897	ETHNIKI GREEK GEN IN CO	1,621
*Insurance*				Portugal
AGF-ASR.GL.DE FRN.	103,549	Ireland		*Banks*
APRIL GROUP	419	*Banks*		BANCO BPI	26,166
AXA	440,744	ALLIED IRISH BANKS	102,042	BNC.ESPR.SANTO (BESCL)R	45,894
EULER HERMES	5,716	ANG.IR.BK.	34,305	BNC.TOTTA &ACORES	28,824 *
FINAXA	4,973	BANK OF IRELAND	106,431	BCP R	71,678
SCOR	12,542	*Insurance*
		FBD HOLDINGS	1,394

Total Assets, end 2004 (In millions of euro)
Spain
*Banks*
BANCO ESPANOL DE CREDITO	66,257
BANCO POPULAR ESPANOL	62,522
BANCO SANTANDER CENTRAL HISPANO	569,795
BBV ARGENTARIA	306,218
*Insurance*
CORP.MAPFRE ‘R’	18,033
GRUPO CATALANA OCCIDENTE	3,536
Sweden
*Banks*
FNSPK. A	112,758
NORDEA BANK	262,523
SEB A	175,698
SVENSKA HANDBKN	149,071
UK
*Banks*
BARCLAYS	735,337
HBOS	623,776
HSBC HDG.	935,023
LLOYDS TSB GP.	394,145
RYL.BK.OF SCTL.	821,784
STD.CHARTERED	103,769
*Insurance*
ADMIRAL GROUP	708
ALEA GP.HDG.(BERMUDA)	4,646
AMLIN	2,858
BEAZLEY	1,136
BENFIELD GROUP	6,052
BRIT INSURANCE HOLDINGS	3,563
CATLIN GROUP	2,437
CHAUCER HOLDINGS	1,074
COX IN.HOLDINGS	1,104
DOMESTIC & GENERAL GP.	474
HISCOX	2,269
JARDINE LLOYD THOMPSON	3,171
KILN	707
ROYAL & SUN ALL.IN.	28,279
WELLINGTON UNDERWRITING	1,841

Table 8.

Banks and Insurance Companies

		Total Assets, end 2004 (In millions of euro)
Austria		Germany		Italy
*Banks*		*Banks*		*Banks*
BANK AU.CREDITANSTALT	145,680	BANKGESELLSCHAFT BERLIN	152,041	UNICREDITO ITALIANO	264,791
ERSTE BANK	139,390	BAYER.HYPO-UND-VBK.	463,496	SAN PAOLO IMI	209,627
*Insurance*		COMMERZBANK	419,674	CAPITALIA	131,163
GENERALI HOLDING	10,430	DEUTSCHE BANK	836,368	BANCA INTESA	273,181
UNIQA	15,125	*Insurance*		BANCA MONTE DEI PASCHI	128,662
WIENER STAEDT VZ	12,525	ALLIANZ	958,579	BCA.NAZ.LAVORO	78,111
		AMB GENERALI HDG.	85,632	*Insurance*
Belgium		AXA KONZERN	38,438	CATTOLICA ASSICURAZIONI	14,773
*Banks*		AXA VERSICHERUNG	7,219	FONDIARIA-SAI	31,381
ALMANIJ	259,629 *	DBV-WINTERTHUR HOLDING	22,398	GENERALI	252,542
DEXIA	349,463	ERGO VERSICHERUNG	108,988	MILANO ASSIC.	9,804
FORTIS (BRU)	521,524	GERLING	4,327	PREMAFIN-HLDG.DI PART.	31,493
KBC BKVS.HDG.	225,587	HANNOVER RUCK.	30,615	RAS	66,229
		KOELN.RUCK.	9,390	UNIPOL	32,201
Denmark		KOELN.VERWALT.GES ELL.	69
*Banks*		MLP	2,378	Netherlands
DANSKE BANK	248,949	MUNCH.RUCK.REGD.	203,501	*Banks*
JYSKE BANK	16,687	WUESTENROT & WUERTT.	54,380	ABN AMRO HOLDING	607,263
SYDBANK	10,453			FORTIS (AMS)	521,524
*Insurance*		Greece		*Insurance*
ALM BRAND	2,897	*Banks*		ING GROEP CERTS.	866,201
CODAN	5,358	ALPHA BANK	32,917
TOPDANMARK	4,516	BANK OF GREECE	32,810	Norway
		BANK OF PIRAEUS	16,591	*Banks*
France		EFG EUROBANK ERGASIAS	31,760	DNB NOR	86,097
*Banks*		EMPORIKI BK.OF GREECE	18,236	SPAREBANKEN ROGALAND	7,152
BNP PARIBAS	905,001	NAT.BK.OF GREECE	52,877	*Insurance*
CREDIT AGRICOLE	814,654	*Insurance*		STOREBRAND	20,871
SOCIETE GENERALE	600,897	ETHNIKI GREEK GEN IN CO	1,621
*Insurance*				Portugal
AGF-ASR.GL.DE FRN.	103,549	Ireland		*Banks*
APRIL GROUP	419	*Banks*		BANCO BPI	26,166
AXA	440,744	ALLIED IRISH BANKS	102,042	BNC.ESPR.SANTO (BESCL)R	45,894
EULER HERMES	5,716	ANG.IR.BK.	34,305	BNC.TOTTA &ACORES	28,824 *
FINAXA	4,973	BANK OF IRELAND	106,431	BCP R	71,678
SCOR	12,542	*Insurance*
		FBD HOLDINGS	1,394

Total Assets, end 2004 (In millions of euro)
Spain
*Banks*
BANCO ESPANOL DE CREDITO	66,257
BANCO POPULAR ESPANOL	62,522
BANCO SANTANDER CENTRAL HISPANO	569,795
BBV ARGENTARIA	306,218
*Insurance*
CORP.MAPFRE ‘R’	18,033
GRUPO CATALANA OCCIDENTE	3,536
Sweden
*Banks*
FNSPK. A	112,758
NORDEA BANK	262,523
SEB A	175,698
SVENSKA HANDBKN	149,071
UK
*Banks*
BARCLAYS	735,337
HBOS	623,776
HSBC HDG.	935,023
LLOYDS TSB GP.	394,145
RYL.BK.OF SCTL.	821,784
STD.CHARTERED	103,769
*Insurance*
ADMIRAL GROUP	708
ALEA GP.HDG.(BERMUDA)	4,646
AMLIN	2,858
BEAZLEY	1,136
BENFIELD GROUP	6,052
BRIT INSURANCE HOLDINGS	3,563
CATLIN GROUP	2,437
CHAUCER HOLDINGS	1,074
COX IN.HOLDINGS	1,104
DOMESTIC & GENERAL GP.	474
HISCOX	2,269
JARDINE LLOYD THOMPSON	3,171
KILN	707
ROYAL & SUN ALL.IN.	28,279
WELLINGTON UNDERWRITING	1,841

References

Adam, Klause, Tullio Jappelli, Annamaria Menichini, Mario Padula, and Marco Pagano, 2002, “Analyze, Compare and Apply Alternative Indicators and Monitoring Methodologies to Measure the Evolution of Capital Market Integration in the European Union”, Report to the European Commission (Brussels: European Commission).
- Search Google Scholar
- Export Citation
Adjaouté, Kpate, and Jean-Pierre Danthine, 2004, “Equity Returns and Integration: Is Europe Changing?” Oxford Review of Economic Policy, Vol. 20, No. 4, pp. 555-70.
- Search Google Scholar
- Export Citation
Arora, Navneet, Jeffrey R. Bohn and Fanlin Zhu, 2005, “Reduced Form versus Structural Models of Credit Risk: A Case Study of Three Models,” Moody’s KMV White Paper, February,(New York: Moody’s).
- Search Google Scholar
- Export Citation
Artis, Michael J., 2003, “Is There A European Business Cycle?” CESifo Working Paper Series No. 1053, (Munich: CESifo)
- Search Google Scholar
- Export Citation
Barros, Pedro Pita, Erik Berglöf, Paolo Fulghieri, Jordi Gual, Colin Mayer, and Xavier Vives, 2005, “Integration of European Banking: The Way Forward,” CEPR, MED 2.
- Search Google Scholar
- Export Citation
Baele, Lieven, Annalisa Ferrando, Peter Hördahl, Elizaveta Krylova, and Cyril Monnet, 2004, “Measuring Financial Integration in the Euro Area,” Oxford Review of Economic Policy, Vol. 20, No. 4, pp. 509-530.
- Search Google Scholar
- Export Citation
Bekaert, Geert, Robert J. Hodrick and Xiaoyan Zhang, 2005, “International Stock Return Comovements,” NBER Working Paper No. 11906, (Cambridge: NBER).
- Search Google Scholar
- Export Citation
Black, F., and M. Scholes, 1973, “The Pricing of Options and Corporate Liabilities,” Journal of Political Economy, Vol. 81, No. 3, pp. 637–54.
- Search Google Scholar
- Export Citation
De Nicolò, Gianni, Robert Corker, Alexander Tieman and Jan Willem Van-derVossen, 2005a, “European Financial Integration, Stability and Supervision, IMF Country Report 05/266, August, pp. 113-146 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
De Nicolò, Gianni, Alexander Tieman, and Robert Corker, 2005, “Large Banking Groups and Financial System Soundness”, IMF Country Report 05/116, March, pp. 102-111 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
De Nicolò, Gianni, Phillip Bartholomew, Jahanara Zaman and Mary Zephirin, 2004, “Bank Consolidation, Internationalization and Conglomeration: Trends and Implications for Financial Risk,” Financial Markets, Institutions & Instruments, Vol. 13, No. 4, pp. 173-217.
- Search Google Scholar
- Export Citation
De Nicolò, Gianni, Peter Hayward, and Ashok Vir Bhatia, 2004, “U.S. Large Complex Banking Groups: Business Strategies, Risks and Surveillance Issues,” IMF Country Report 04/228, July, pp. 72-86 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
De Nicolò, Gianni and Myron Kwast, 2002, “Financial Risk and Consolidation: Are They Related?” Journal of Banking and Finance, Vol. 26, No. 5, pp. 861–880.
- Search Google Scholar
- Export Citation
Duan, Jin-Chuan, Genevieve Gauthier and Jean-Guy Simonato, 2004, “On the Equivalence of the KMV and Maximum Likelihood Methods for Structural Credit Risk Models” (unpublished; Toronto: University of Toronto).
- Search Google Scholar
- Export Citation
European Central Bank, 2004, “Accounting for the Resilience of the EU Banking Sector,” Monthly Bulletin, April (Frankfurt European Central Bank).
- Search Google Scholar
- Export Citation
Flood, Robert, and Andrew Rose, 2005, “Estimating the Expected Marginal Rate of Substitution: A Systematic Exploitation of Idiosyncratic Risk,” Journal of Monetary Economics, Vol. 52, No.5, pp. 951-969.
- Search Google Scholar
- Export Citation
Gropp, R.E., J. Vesala, and G. Vulpes, 2006, “Equity and Bond Market Signals as Leading Indicators of Bank Fragility,” Journal of Money, Credit and Banking, Vol.32, No.2, pp. 399-428.
- Search Google Scholar
- Export Citation
Hartmann, Philipp, Stefan Straetsman, and Casper de Vries, 2005, “Banking System Stability: A Cross-Atlantic Perspective,” Working Paper No. 527, September (Frankfurt: European Central Bank).
- Search Google Scholar
- Export Citation
Houston, Joel F., and Kevin J Stiroh, 2006, “Three Decades of Financial Sector Risk,” U.S. Financial Holding Companies,” July (unpublished; New York: Federal Reserve Bank of New York).
- Search Google Scholar
- Export Citation
Kose, Ayhan, Christopher Otrok, and Charles Whiteman, “Understanding the Evolution of World Business Cycles”, IMF Working Paper 05/211, November (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
Krainer, J. and J.A. Lopez, 2001, “Incorporating Equity Market Information into Supervisory Monitoring Models,” FRBSF Working Paper 2001-14 (San Francisco: Federal Reserve Bank).
- Search Google Scholar
- Export Citation
Lumsdaine, Robert L. and Eswar S. Prasad, 2003, “Identifying the Common Component of International Economic Fluctuations: A New Approach,” Economic Journal, Vol. 113 (January), pp.101-127.
- Search Google Scholar
- Export Citation
Marshall, David, 2005, “Comment on Estimating the Expected Marginal Rate of Substitution”, Journal of Monetary Economics, Vol. 52, No. 5, pp. 971-979.
- Search Google Scholar
- Export Citation
Merton, R.C., 1974, “On the Pricing of Corporate Debt: The Risk Structure of Interest Rates,” Journal of Finance, Vol. 29, No. 2, pp. 449–70.
- Search Google Scholar
- Export Citation
Ravn, Morten O. and Harald Uhlig, 2002, “On Adjusting the Hodrick-Prescott Filter for the Frequency of Observations,” Review of Economics and Statistics, Vol. 84, No. 2, pp. 371-76.
- Search Google Scholar
- Export Citation
Solnik, Bruno, and Jacques Roulet, 2000, “Dispersion as Cross-Sectional Correlation,” Financial Analyst Journal, Vol. 56, No. 1, pp. 54-61.
- Search Google Scholar
- Export Citation
Stiroh, Kevin J, 2004, “Diversification in Banking: Is Noninterest Income the Answer?” Journal of Money, Credit and Banking, Vol. 36, No. 5, pp. 853-882.
- Search Google Scholar
- Export Citation
Stiroh, Kevin J, and Adrienne Rumble, 2006, “The Dark Side of Diversification: The Case of U.S. Financial Holding Companies,” Journal of Banking and Finance, Vol. 30, No. 8, pp 2131-2161.
- Search Google Scholar
- Export Citation
Stock, James H., and Mark W. Watson, 2005, “Understanding Changes in International Business Cycle Dynamics,” Journal of the European Economic Association, Vol. 3, No. 5, September, pp. 968-1006.
- Search Google Scholar
- Export Citation
Tarashev, Nikola A., 2005, “An Empirical Evaluation of Structural Credit Risk Models”, BIS Working Paper, No. 179, July, (Basel: Bank for International Settlements).
- Search Google Scholar
- Export Citation
Vassalou, M., and Y. Xing, 2004, “Default Risk in Equity Returns,” Journal of Finance, Vol. 59, No. 2, pp. 831-68.
- Search Google Scholar
- Export Citation

Research Department and Monetary and Capital Markets Department, respectively. We thank Robert Corker, Bob Flood, Ayan Kose, Til Schuermann, Boris Vujcic, and participants to IMF seminars, the DG-ECFIN 2005 Research Conference, and the XIV Tor Vergata Conference, for comments and suggestions. Excellent research assistance by Marianne El-Khoury is gratefully acknowledged.

For recent reviews of the literature, see Stock and Watson (2005) and Kose, Otrok, and Whiteman (2005). Persistence in business cycles heterogeneity within Europe is stressed by Artis (2003).

These countries are: Austria, Belgium, France, Germany, Greece, Italy, Netherlands, Portugal, Sweden, and the United Kingdom. All results we obtain for first moments hold when we use quarterly GDP growth data instead of IPG data.

As shown in Solnik and Roulet (2000), under the assumption that the data generating process for a sufficiently large set of variables is described by a factor model, the evolution of their cross-sectional dispersion is inversely related to their pairwise correlation.

For data at monthly frequency, here and in the sequel we use the value of the smoothing parameter for the HP filter derived by Ravn and Uhlig (2002), equal to 129,600.

The sectors covered are: Banks, Insurance, Financials, Non-Financials, Basic Industries, Cyclical Consumer Goods, Cyclical Services, General Industrials, Information Technology, Non-Cyclical Services, Non-Cyclical Consumer Goods, Resources, and Utilities. The countries are: Austria, Belgium, Denmark, France, Germany, Ireland, Italy, Netherlands, and the United Kingdom. In all first step regressions, the first two principal components were sufficient to capture common risks according to standard statistical criteria.

Risk measures obtained from structural models, such as the BSM model, have been shown to have predictive power for supervisory ratings, bond spreads, and rating agencies’ downgrades as well as actual defaults (see Krainer and Lopez, 2001; Gropp, Vesala, and Vulpes, 2006; Arora, Bohn and Zhu, 2005; and Tarashev, 2005).

This fact can be also seen as an implication of Jensen’s inequality.

Recall that under the assumption that asset values follow a lognormal process, DD measures do not necessarily capture extreme events adequately. In addition, the simplifying assumption that the liability structure is composed of only equity and debt with fixed maturity for all firms, and no rollover of debt, may lead to underestimates of interest rate risks and other risks associated with contingent (derivative) liabilities.

As of end-2004, the Datastream indices included 63 banks and 53 insurance companies, whose identities and asset sizes are reported in the Appendix.

Other Publishers

Save
Cite
Email this content

Share Link

Copy this link, or click below to email it to a friend
Email this content
or copy the link directly:

https://www.elibrary.imf.org/view/journals/001/2006/296/article-A001-en.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully

Economic Integration and Financial Stability: A European Perspective

Author: Mr. Gianni De Nicolo and Alexander F. Tieman

Volume/Issue:: Volume 2006: Issue 296

Publisher:: International Monetary Fund

ISBN:: 9781451865561

ISSN:: 1018-5941

Pages:: 28

DOI:: https://doi.org/10.5089/9781451865561.001

Keywords
I. Introduction
II. Synchronization of Real Activity
III. Equity Markets Integration
IV. Systemic Risk and Integration
V. Conclusion
Appendix
References

View raw image

EGARCH estimates of IPG Common Component

View raw image

Bank DDs

View raw image

Insurance DDs

	β₁	β₁₁	β₁₂	β₁₃	β₂	β₂₁	β₂₂	β₂₃
AUSTRIA	0.22914	0.19882	0.24446	0.19005	3.06991	-2.58253	5.08501	4.14558
t-stat	1.12680	0.48277	0.74375	0.86648	0.90713	-0.31624	0.88052	2.02906
BELGIUM	-0.05524	-0.14530	-0.18260	0.12222	6.79376	7.18137	6.00944	7.22582
t-stat	-0.44064	-0.83982	-0.48629	0.89236	3.71733	2.21134	1.41646	4.12920
DENMARK	0.18350	-0.02602	0.08777	0.50247	5.39206	11.29102	5.53111	1.83205
t-stat	1.72894	-0.21526	0.35315	2.98560	4.54716	5.23163	2.81885	1.05308
FRANCE	-0.06485	-0.21530	-0.07057	0.06281	8.51009	6.23099	10.61516	7.25105
t-stat	-0.86576	-2.08099	-0.37863	0.58821	6.68370	2.43581	5.36341	4.64868
GERMANY	0.00863	-0.08972	-0.12743	0.14603	6.01768	1.50707	5.94823	8.47721
t-stat	0.10507	-0.52650	-0.64834	1.93410	4.15413	0.24985	3.36277	5.78293
GREECE	0.00882	-0.10041	0.14396	0.02024	6.84393	4.94287	9.23776	5.07508
t-stat	0.10323	-0.90929	0.64083	0.15302	5.54959	2.61422	5.00959	2.54594
IRELAND	-0.08240	-0.26604	-0.08318	0.10216	5.67918	6.03562	5.79232	4.84149
t-stat	-1.01429	-1.88952	-0.49246	0.83259	4.19292	3.09929	2.26681	2.32658
ITALY	0.21024	0.25366	0.11337	0.17804	7.30423	2.14074	10.67059	7.15589
t-stat	2.50871	1.80220	0.77431	1.41244	6.47755	0.86133	7.00629	4.21428
NETHERLANDS	-0.02262	-0.23540	-0.03518	0.14693	5.91218	1.99414	6.59478	6.83980
t-stat	-0.26792	-1.62760	-0.16432	1.81176	4.79348	0.59233	3.15099	4.37547
NORWAY	0.17415	0.20640	0.10871	0.18771	5.42550	5.60672	5.62960	5.31900
t-stat	1.95444	1.09798	0.72036	1.61691	4.36156	1.73398	2.78827	3.66942
PORTUGAL	0.14310	-0.14880	0.35741	0.23421	7.23641	2.33354	8.10958	7.98958
t-stat	1.55419	-1.09203	2.06801	1.41386	5.71466	0.82094	3.66198	4.03926
SPAIN	-0.00595	-0.06114	-0.14175	0.15393	8.99014	10.51475	9.79983	7.39503
t-stat	-0.07638	-0.45176	-0.78550	1.52710	6.44645	2.56563	4.43565	4.20194
SWEDEN	0.11139	-0.05850	0.01928	0.35596	5.62820	7.54971	6.35453	3.66768
t-stat	1.44782	-0.47774	0.11013	3.77615	4.85987	3.18056	2.92426	3.14802
U.K.	-0.11513	-0.40599	-0.09382	0.15302	4.95129	5.10395	5.11531	3.94157
t-stat	-1.39310	-2.93993	-0.62166	1.18845	4.43553	2.00226	2.62350	2.81685

	β₁	β₁₁	β₁₂	β₁₃	β₂	β₂₁	β₂₂	β₂₃
AUSTRIA	0.22914	0.19882	0.24446	0.19005	3.06991	-2.58253	5.08501	4.14558
t-stat	1.12680	0.48277	0.74375	0.86648	0.90713	-0.31624	0.88052	2.02906
BELGIUM	-0.05524	-0.14530	-0.18260	0.12222	6.79376	7.18137	6.00944	7.22582
t-stat	-0.44064	-0.83982	-0.48629	0.89236	3.71733	2.21134	1.41646	4.12920
DENMARK	0.18350	-0.02602	0.08777	0.50247	5.39206	11.29102	5.53111	1.83205
t-stat	1.72894	-0.21526	0.35315	2.98560	4.54716	5.23163	2.81885	1.05308
FRANCE	-0.06485	-0.21530	-0.07057	0.06281	8.51009	6.23099	10.61516	7.25105
t-stat	-0.86576	-2.08099	-0.37863	0.58821	6.68370	2.43581	5.36341	4.64868
GERMANY	0.00863	-0.08972	-0.12743	0.14603	6.01768	1.50707	5.94823	8.47721
t-stat	0.10507	-0.52650	-0.64834	1.93410	4.15413	0.24985	3.36277	5.78293
GREECE	0.00882	-0.10041	0.14396	0.02024	6.84393	4.94287	9.23776	5.07508
t-stat	0.10323	-0.90929	0.64083	0.15302	5.54959	2.61422	5.00959	2.54594
IRELAND	-0.08240	-0.26604	-0.08318	0.10216	5.67918	6.03562	5.79232	4.84149
t-stat	-1.01429	-1.88952	-0.49246	0.83259	4.19292	3.09929	2.26681	2.32658
ITALY	0.21024	0.25366	0.11337	0.17804	7.30423	2.14074	10.67059	7.15589
t-stat	2.50871	1.80220	0.77431	1.41244	6.47755	0.86133	7.00629	4.21428
NETHERLANDS	-0.02262	-0.23540	-0.03518	0.14693	5.91218	1.99414	6.59478	6.83980
t-stat	-0.26792	-1.62760	-0.16432	1.81176	4.79348	0.59233	3.15099	4.37547
NORWAY	0.17415	0.20640	0.10871	0.18771	5.42550	5.60672	5.62960	5.31900
t-stat	1.95444	1.09798	0.72036	1.61691	4.36156	1.73398	2.78827	3.66942
PORTUGAL	0.14310	-0.14880	0.35741	0.23421	7.23641	2.33354	8.10958	7.98958
t-stat	1.55419	-1.09203	2.06801	1.41386	5.71466	0.82094	3.66198	4.03926
SPAIN	-0.00595	-0.06114	-0.14175	0.15393	8.99014	10.51475	9.79983	7.39503
t-stat	-0.07638	-0.45176	-0.78550	1.52710	6.44645	2.56563	4.43565	4.20194
SWEDEN	0.11139	-0.05850	0.01928	0.35596	5.62820	7.54971	6.35453	3.66768
t-stat	1.44782	-0.47774	0.11013	3.77615	4.85987	3.18056	2.92426	3.14802
U.K.	-0.11513	-0.40599	-0.09382	0.15302	4.95129	5.10395	5.11531	3.94157
t-stat	-1.39310	-2.93993	-0.62166	1.18845	4.43553	2.00226	2.62350	2.81685

	β₁	β₁₁	β₁₂	β₁₃	β₂	β₂₁	β₂₂	β₂₃
AUSTRIA	-0.07401	-0.13646	-0.02889	-0.05242	5.56877	4.79820	6.68891	4.74129
t-stat	-0.87164	-0.79823	-0.18099	-0.41560	3.93998	1.12355	3.95795	2.21433
DENMARK	0.35582	1.05606	0.12418	-0.00905	11.31496	28.15192	7.08159	8.84322
t-stat	1.32135	1.36990	0.52799	-0.04400	3.67787	1.89994	3.42040	2.21464
FRANCE	-0.19178	-0.48789	-0.20667	0.08307	10.62458	8.60484	12.58048	9.19347
t-stat	-2.11906	-2.58970	-1.10573	0.89256	7.15602	2.15961	4.77712	5.59050
GERMANY	0.02595	-0.03629	0.01163	0.08253	6.17503	4.73298	3.90193	9.01931
t-stat	0.29904	-0.18959	0.06824	0.84674	3.85809	0.77777	1.72722	6.90630
GREECE	-0.00128	-0.16637	0.31959	-0.05217	5.78269	4.96145	5.78071	5.30061
t-stat	-0.01313	-1.81309	1.48146	-0.33157	4.71844	2.88784	2.51895	2.65559
IRELAND	-0.00002	-0.12271	0.04893	0.09791	4.27298	5.76075	6.61989	0.92890
t-stat	-0.00143	-0.92617	0.13156	0.39044	2.58317	1.78139	1.97196	0.44966
ITALY	0.13099	0.12708	0.05859	0.12922	5.67917	0.60276	7.62463	6.70161
t-stat	1.99638	1.09772	0.50392	1.24325	4.58143	0.19621	3.87340	4.68354
NETHERLANDS	0.02564	-0.14301	0.01596	0.12346	5.43702	-1.57370	8.48470	5.82751
t-stat	0.30738	-1.06652	0.07301	1.42235	3.59210	-0.41880	3.98888	3.69937
NORWAY	0.04063	0.00638	0.08005	0.08555	5.45010	8.69135	4.54006	4.48679
t-stat	0.46890	0.04253	0.41735	0.73726	4.95299	2.80595	2.57227	2.96369
SPAIN	0.02762	-0.10887	0.12016	0.09305	5.97838	5.90317	7..39693	4.33736
t-stat	0.39628	-1.16996	0.75689	0.84544	5.60565	4.09959	3.59187	3.03786
U.K.	-0.14484	-0.28911	0.17216	-0.19506	5.39866	7.48984	2.26799	6.49323
t-stat	-1.57067	-2.35482	0.66822	-1.30092	3.21487	3.25397	0.89055	2.04962

	β₁	β₁₁	β₁₂	β₁₃	β₂	β₂₁	β₂₂	β₂₃
AUSTRIA	-0.07401	-0.13646	-0.02889	-0.05242	5.56877	4.79820	6.68891	4.74129
t-stat	-0.87164	-0.79823	-0.18099	-0.41560	3.93998	1.12355	3.95795	2.21433
DENMARK	0.35582	1.05606	0.12418	-0.00905	11.31496	28.15192	7.08159	8.84322
t-stat	1.32135	1.36990	0.52799	-0.04400	3.67787	1.89994	3.42040	2.21464
FRANCE	-0.19178	-0.48789	-0.20667	0.08307	10.62458	8.60484	12.58048	9.19347
t-stat	-2.11906	-2.58970	-1.10573	0.89256	7.15602	2.15961	4.77712	5.59050
GERMANY	0.02595	-0.03629	0.01163	0.08253	6.17503	4.73298	3.90193	9.01931
t-stat	0.29904	-0.18959	0.06824	0.84674	3.85809	0.77777	1.72722	6.90630
GREECE	-0.00128	-0.16637	0.31959	-0.05217	5.78269	4.96145	5.78071	5.30061
t-stat	-0.01313	-1.81309	1.48146	-0.33157	4.71844	2.88784	2.51895	2.65559
IRELAND	-0.00002	-0.12271	0.04893	0.09791	4.27298	5.76075	6.61989	0.92890
t-stat	-0.00143	-0.92617	0.13156	0.39044	2.58317	1.78139	1.97196	0.44966
ITALY	0.13099	0.12708	0.05859	0.12922	5.67917	0.60276	7.62463	6.70161
t-stat	1.99638	1.09772	0.50392	1.24325	4.58143	0.19621	3.87340	4.68354
NETHERLANDS	0.02564	-0.14301	0.01596	0.12346	5.43702	-1.57370	8.48470	5.82751
t-stat	0.30738	-1.06652	0.07301	1.42235	3.59210	-0.41880	3.98888	3.69937
NORWAY	0.04063	0.00638	0.08005	0.08555	5.45010	8.69135	4.54006	4.48679
t-stat	0.46890	0.04253	0.41735	0.73726	4.95299	2.80595	2.57227	2.96369
SPAIN	0.02762	-0.10887	0.12016	0.09305	5.97838	5.90317	7..39693	4.33736
t-stat	0.39628	-1.16996	0.75689	0.84544	5.60565	4.09959	3.59187	3.03786
U.K.	-0.14484	-0.28911	0.17216	-0.19506	5.39866	7.48984	2.26799	6.49323
t-stat	-1.57067	-2.35482	0.66822	-1.30092	3.21487	3.25397	0.89055	2.04962

Topics

Countries

Series

About

Resources

Contributor Notes

Abstract

I. Introduction

II. Synchronization of Real Activity

III. Equity Markets Integration

IV. Systemic Risk and Integration

A. Measures of Systemic Risk

B. Trends in Systemic Risk

C. Convergence of Systemic Risk Measures

D. The Sensitivity of Systemic Risk Measures to Real and Financial Shocks

V. Conclusion

Appendix

References

Same Series

Other IMF Content

Other Publishers

Asian Development Bank

Inter-American Development Bank

Nordic Council of Ministers

The World Bank

Share Link

Table of Contents

Headings

Figures

Metrics