bond market and economic growth

Published on February 2017 | Categories: Documents | Downloads: 50 | Comments: 0 | Views: 647
of 35
Download PDF   Embed   Report

Comments

Content

FORSCHUNGSINSTITUT FÜR EUROPAFRAGEN RESEARCH INSTITUTE FOR EUROPEAN AFFAIRS WIRTSCHAFTSUNIVERSITÄT WIEN UNIVERSITY OF ECONOMICS AND BUSINESS ADMINISTRATION VIENNA

Working Papers
IEF Working Paper Nr. 49

GERHARD FINK / PETER HAISS / SIRMA HRISTOFOROVA

Bond Markets and Economic Growth April 2003

Althanstraße 39 - 45, A - 1090 Wien / Vienna Österreich / Austria Tel.: ++43 / 1 / 31336 / 4135, 4134, 4133 Fax.: ++43 / 1 / 31336 / 758, 756 e-mail: [email protected]

Electronic copy available at: http://ssrn.com/abstract=1003763

Impressum: Die IEF Working Papers sind Diskussionspapiere von MitarbeiterInnen und Gästen des Forschungsinstituts für Europafragen an der Wirtschaftsuniversität Wien, die dazu dienen sollen, neue Forschungsergebnisse im Fachkreis zur Diskussion zu stellen. Die Working Papers geben nicht notwendigerweise die offizielle Meinung des Instituts wieder. Sie sind gegen einen Unkostenbeitrag von € 7,20 am Institut erhältlich. Kommentare sind an die jeweiligen AutorInnen zu richten.
Medieninhaber, Eigentümer Herausgeber und Verleger: Forschungsinstitut für Europafragen der Wirtschaftsuniversität Wien, Althanstraße 3945, A1090 Wien; Für den Inhalt verantwortlich: Univ.-Prof. Dr. Stefan Griller, Althanstraße 3945, A1090 Wien.

Nachdruck nur auszugsweise und mit genauer Quellenangabe gestattet.

Electronic copy available at: http://ssrn.com/abstract=1003763

Bond Markets and Economic Growth

Bond Markets and Economic Growth
Gerhard FINK, Peter HAISS, Sirma HRISTOFOROVA

April 2003
Abstract This paper examines the relationship between the development of the aggregate bond markets and real GDP in 13 highly developed economies. The recent interest in the ties between the real and the financial sector has usually been on the banking sector and the stock markets, rather ignoring the bond markets as a third essential source of external finance. We fill this gap by providing empirical evidence for causality patterns supporting the supply-leading approach in the USA, UK, Switzerland, Germany, Austria, the Netherlands and Spain over the 1950 to 2000 period. In the cases of Japan, Finland and Italy we find evidence of interdependence between bond market capitalization growth and real output growth. Granger causality test and co-integration approach are employed to support this conjecture.

JEL Classification: E44, O16, O40. Keywords: bond markets, financial markets, economic development, Granger causality.

3

IEF Working Paper Nr.49

1. Introduction The continuous attention, which attracts the nexus between the financial and the real sector, reflects the importance of this topic for economic development.1 However, the growing literature on the subject has largely ignored the bond markets despite their role as an essential source of external finance (Wachtel, 2001; DeBondt, 2002). The few previous research efforts on bond markets concentrated only on certain regions and/or involved rather short time horizons (DeBondt, 2002), dealt with financial crisis situations rather than the whole business cycle (Herring and Chatusripitak, 2000; Batten and Kim, 2000), or linked GDP growth to the term structure of interest rates in order to forecast recessions (Harvey, 1989, 1991; Gamber, 1996; Gerlach and Smets, 1997; Ahrens, 2002). We close this gap and provide first empirical, multi-country time series evidence on the link between bond markets and real GDP growth. Specifically, we model the causal relationship from the bond market capitalization to GDP growth for the USA, Japan, Switzerland and most European Economic Area (EEA) countries over the last fifty years (data-permitting). The paper is organized as follows: section two presents descriptive statistics of the bond markets and the rationale for this paper, section three reviews the most recent studies which partially fit the line of our research and provides a tabular overview of the empirical literature on the finance-growth nexus, section four outlines the methodology used, section five presents the empirical findings and section six discusses the results and sketches some ideas for further research. Section seven concludes. A description of the data set is provided in the Data Appendix. 2. Rationale In comparison with bank loans and trade credits, which are the most important sources of external finance in the European Union (ECB, 2001:42 ff.), stocks and bonds play a less important role as financing instruments. Even in the USA, whose system is commonly classified as securities-oriented, external financial transactions of non-financial corporations tend to be dominated by credits and loans (ECB, 2001:44). Notwithstanding this assertion, a closer analysis of the financial markets shows that the aggregate bond market capitalization is as important as stock market capitalization (both measured by the amounts outstanding) and amounts in total to about one-third of the entire financial assets in the USA (32.8 %), and in the EU (EU15: 33.1%, Euro-12: 35.1%).

1

To name but two large research networks: Research teams of the World Bank (http://www.worldbank.org/research/projects/finstructure) specialise on financial structures issues and CEPR (http://www.cepr.org/research/networks/fertn) on the legal and political framework of finance and their implications for real growth.

4

Bond Markets and Economic Growth

In ten out of the nineteen countries in our sample, the volume of debt securities (government and corporate bonds) is higher than the stock amounts outstanding measured as a percentage of total financial assets (TFA): Japan, Norway, Germany, Italy, Belgium, Denmark, Austria, Luxembourg, Ireland and Portugal. Luxembourg, which serves as a hub for international bond issues, has the highest bonds to TFA ratio (68.5%), whereas in Switzerland this ratio amounts to only 14%. In six bondoriented EEA members the debt securities sector makes up the largest share of TFA and exceeds both the credit and the stock markets: Norway, Italy, Belgium, Denmark, Luxembourg and Ireland. Table 1: Bond amounts outstanding and relative size of major financial sectors in 1999
Bond amounts outstanding in mn $ EU15+4 31,599,500 USA 15,442,700 JPN 6,570,300 SUI 185,700 NOR 94,500 EU15 9,306,300 Euro-12 7,179,000 GER 2,160,500 UK 1,504,900 ITA 1,451,300 FRA 1,381,300 NL 676,500 E 393,600 BEL 368,400 SWE 333,200 DNK 289,200 AUT 228,000 GRE 162,426 FIN 118,100 IRL 94,900 LUX 79,366 POR 77,800 Country Bonds 149 166 136 76 64 116 114 108 104 130 102 180 70 157 144 176 115 99 97 108 627 72 Financial assets in % of GDP Credit Stocks 194 145 162 179 331 92 185 285 58 43 121 113 120 91 147 72 127 198 96 65 101 111 141 185 116 77 140 78 114 161 57 64 125 17 94 170 58 286 107 78 92 197 122 63 TFA 488 507 558 546 165 350 326 328 430 291 314 506 263 375 420 297 257 363 441 293 916 258 Financial sectors in % of TFA Bonds Credit Stocks 30.4 39.8 29.8 32.8 31.9 35.3 24.3 59.2 16.5 14.0 33.9 52.2 38.7 35.2 26.1 33.1 34.6 32.2 35.1 36.9 28.0 33.1 44.9 22.0 24.3 29.5 46.1 44.7 32.8 22.5 32.5 32.2 35.3 35.6 27.8 36.6 26.7 44.1 29.2 41.8 37.3 20.9 34.3 27.3 38.4 59.2 19.3 21.5 44.8 48.7 6.5 27.2 26.0 46.8 21.9 13.1 64.9 36.8 36.5 26.7 68.5 10.0 21.5 28.0 47.4 24.5

Notes: Calculations are based on real 1999 data. Countries are ranked in the descending order of the size of the bond market in absolute terms. TFA: Total financial assets = bond amounts outstanding + stock amounts outstanding + domestic credit provided by banks and MFIs. Sources: IFS, BIS, FIBV, OECD, national sources.

Despite the growing corporate bond issuance in Europe, especially after the introduction of the euro as a tangible currency in the European Union (Fink and Fenz, 2002; Neaime and Paschakis, 2002), bond financing has not attracted the deserved attention by the empirical literature. Denomination in euro facilitates companies to issue corporate bonds on other than their domestic markets. The outstanding amount of bonds has been steadily increasing over time and since 1998 at a relatively faster pace than before. Net issuance rose sharply in the first two quarters of 1998. After some temporarily decline it reached its peak in 1999 (more than 360 billions of US dollars) and remained at relatively high levels. Considering the floating regime since its inception in 1973 until 2000, the average growth rates of the bond markets in the countries covered in our study are very impressive. Unfortunately, the time series for Norway, Denmark, Greece, Ireland and Luxembourg only start in 1990, what disqualifies

5

IEF Working Paper Nr.49

them for further econometric analysis. However, we can provide average annual growth rates for all the EU15 and the USA, Japan, Switzerland and Norway from 1981-1990 and 1991-2000 (Table 2).
k

The growth rates are calculated as (

xt + k − 1 )* 100 , where k is the number of years, and x t denotes xt

the bonds amount outstanding in the starting year. Netherlands’ bond markets display the same growth rate in both periods, whereas the bond markets in the other countries grew at a different pace during the sub-periods.2 Table 2: Bond Market Development: Average annual bond market growth in percent between 1981-1990 and 1991-2000
Average growth POR FIN Nl FRA rate p.a. 1981-90 50.72 13.00 11.75 9.52 1991-00 5.21 4.86 11.52 5.97 USA 9.07 4.97 SUI 9.03 1.00 AUT 7.27 7.60 E 7.04 6.80 SWE JPN 5.71 5.06 6.37 5.02 ITA 5.83 3.83 BEL GER 4.78 2.97 UK LUX NOR GRE DNK IRL Time series start in 1990.

3.85 -1.36

9.58 11.33 8.82 7.70 6.75 2.89 10.43

Note: Values are given in percent. Countries are ranked in the descending order of the first period’s growth rates. Source: own calculations.

3. Literature Review 3.1. Possible causal relationships between the financial and real sector Before turning to the overview of the vast literature on the finance-growth nexus, it is useful to discuss briefly the possible hypotheses describing the causality patterns between economic activity and finance. Five possible hypotheses with respect to the causal ties between the bond and the real sectors can be derived: (1) supply-leading; (2) demand-leading; (3) interdependence; (4) no causal relation; and (5) negative causality from finance to growth. These five hypotheses can be combined in a number of syntheses, basically with respect to evolutionary sequencing or different needs of industries and/or enterprises. The supply-leading hypothesis maintains that accumulation of financial assets triggers economic growth. The modern branch of the supply-leading finance literature unarguably begins with the works of Mc Kinnon (1973) and Shaw (1973). Since then, the major consensus is that finance positively influences real economic performance. The demand-leading hypothesis assumes that real growth drives the emergence and establishment of financial centers. This hypothesis regards financial development as endogenously determined by the real economy or its needs - a view that is consistent with the Coase (1956) theorem and the new institutional economics, which argue that institutions adjust to market imperfections in a way that maximizes individual utilities.

2

The reason for the astonishing growth rate of average 51% p.a. during the first sub-period in Portugal is probably attributable to regulation changes in course of the EU accession as bond markets start to develop rapidly after the EU accession. From 1981 until 1987 the real bond market capitalization measured in 1995 USD grew from 894.41 to 2,348.05 millions, i.e. at an annual growth rate of 15% on average, and within only a year, from 1987 to 1988, the value reached 45,538.17 millions.

6

Bond Markets and Economic Growth

The view that there is no causal link between the financial sector and real economic development is expressed most prominently by Lucas (1988). This assertion holds only in a neo-classical world of zero transaction costs (Graff, 2000) and perfect information. In such a world, the Modigliani-Miller theorem (1958) holds and institutions, particularly financial institutions, do not matter. The irrelevance hypothesis has come under attack from various sides. Economists increasingly deny the existence of frictionless markets, primarily due to informational and agency problems, transaction costs, etc. Generally, the importance of institutions is now more acknowledged than in the past and a large number of empirical studies present strong evidence for the relevance of finance for real growth. Most of them use as impulse variables one or more financial variables, such as various money aggregates, usually relative to (real) GDP, credits to non-financial institutions, and bank assets as a measure of the banking sector. Market capitalization and sometimes stock market turnover are drawn to describe the capital markets. Researchers have only recently attempted to include internal financing (notably Rivaud, Dunset, Dubocage, Salais, 2001, and to a lesser extent Claessens and Laeven, 2002). Control variables and conditioning sets of empirical studies cover a wide range of macroeconomic, institutional and educational indicators. In the studies reviewed here, the dependent variable is mainly real per capita GDP growth. In some cases investment-related variables are used also. Studies operating on the industry-level additionally use value added as a dependent variable. 3.2. Empirical studies on bank intermediation, stock markets and growth While cross-country studies normally test the a priori assumption that finance influences the real economy (Levine, Renelt, 1992; King, Levine, 1992; Atje, Jovanovic, 1993; Sala-i-Martin, 1997; Berthélemy, Varoudakis, 1997; Graff, 2000; Rivaud-Danset, Dubocage, Salais, 2001; Leahy, Schich, Wehinger, Pelgrin, Thorgeirsson, 2001; Claessens, Laeven, 2002, Hahn, 2002), time-series analyses address the question of causality mostly via Granger causality tests (Arestis and Demetriades, 1997; Neusser and Kugler, 1998; Darrat, 1999; Shan, Morris and Sun, 2001). There is weak empirical evidence in favor of the supply-leading pattern in the early stages of financial and general economic development, as well as of demand-leading causality in the very longrun. The supply-leading pattern can be traced back to Joseph Schumpeter (1911), who asserted that a well-functioning banking system promotes technological innovation by selectively funding the most innovative and creative enterprises having the best prospects of success. Most studies using time-series analysis focus on the banking sector and the stock markets in order to evaluate the extent to which bank intermediation and efficient stock markets positively influence real growth. The evidence provided is heterogeneous. The most prominent scholar in cross country analysis is Ross Levine. Levine (1997) summarizes the existing theories into the “finance-growth nexus” framework and assesses the quantitative importance of the financial system to economic growth, though he also restricts himself to banking

7

IEF Working Paper Nr.49

intermediation and stock markets only. He advocates the functional approach of the role of financial systems, which focuses upon the relationships between growth and the quality of the functions provided by the financial system (these include monitoring managers, facilitating trading risk, allocating capital, mobilizing savings, etc.). Third factors, such as a country’s legal system and political institutions, are further believed to play a role for the financial and economic development.3 The statistical evidence he presents is based on cross-country comparisons and individual country studies for 35 countries from 1960 to 1989. Four measures of the degree of smooth functioning of the financial system relative to real per capita GDP in 1985 are used: the ratio of liquid liabilities to GDP; the ratio of bank credit to bank credit plus central bank domestic assets; the ratio of credit allocated to private enterprises to total domestic credit (excluding credit to banks); and credit to private enterprises relative to GDP. A positive, statistically significant correlation is found between real per capita GDP and the extent to which loans are directed to the private sector. In a more recent empirical paper Levine and Zervos (1998) examine the role of stock markets and banks in spurring economic growth. Using data for 47 countries over the 1976 to 1993 period they find that both stock market liquidity and banking development together positively predict real growth, capital accumulation and productivity improvements. A selection of cross-section studies with contrasting findings is provided in table 3. Broader overviews can be found in Graff (2000) and Blum et al. (2002). King and Levine (1993), Leahy et al. (2001) and Claessens and Laeven (2002) find a positive impact on growth, to which a finance-friendly legal framework can also contribute. Harris (1997) and Hahn (2002) find minor causal links. RivaudDanset et al. (2001) find no correlation between financial variables and industry performance, and Berthelemy and Varoudakis (1997) find a negative relationship between financial depth and real growth. The studies that employ stock market capitalization as explanatory variable indicate a positive influence on economic growth. Using data for up to 47 countries during the 1976-1993 period, Levine and Zervos (1998) find that when entered together in regressions, both stock market liquidity and banking development positively predict real growth, capital accumulation and productivity improvements.

3

For extensive elaboration on this notion see Levine (1999), Law, Finance, and Economic Growth, Journal of Financial Intermediation, also Levine (1998), The Legal Environment, Bank, and Long-Run Economic Growth, Journal of Money, Credit, and Banking, Vol. 30, No. 3, pp.597-620, Levine, Loayza, Beck (2000), Financial intermediation and growth: Causality and causes, Journal of Monetary Economics 46, pp.31-77, and La Porta, Lopez-de-Silanes, Shleifer and Vishny (1998), Law and Finance, Journal of Political Economy, Vol. 106, No. 6, pp.1113-1115.

8

Bond Markets and Economic Growth

Table 3: Cross-country studies on the links between financial intermediation, stocks and real growth ordered chronologically by year of publication
Authors/Year of publication Country sample Time span EFV I: EFV II: Stocks Banking sector Liquid liabilities/GDP; Assets of commercial & Output growth; central banks/GDP; capital stock Private credit/GDP; growth; Credits to private productivity enterprises/GDP; growth Credits to private, public enterprises & local governments Real income per capita growth ln (M2/GDP) Dependent variable Other explanatory variables Control/ other variables Method Investigated links between Major findings Additional results Case studies and firm-level studies: liberalisation of the financial market leads to higher growth rates, seems to be related to the financial crisis that occurred in many countries 3-5 years after financial liberalisation

Up to 80 King, Levine countries (up 1960-1989 to 19 EU+; up 1993 to 3 ACC)

Initial income; School Banking enrolment rate; Panel Trade exposure; analysis: 5- intermediation & Government year periods economic growth spending/GDP Inflation

Positive relationships between financial intermediation and economic growth

Berthélemy, Varoudakis 1997

85 countries (details not available)

1960-1990

Convergence; investment rate; inflation; education; trade exposure; etc. Investment ratio*stock market turnover/ GDP

Panel Financial depth regression and economic analysis: 5growth year periods

Negative relationship between financial depth and real growth Very limited positive relationship between capital markets & growth only for highincome countries, contrasting to Atje, Jovanovic (1993) Strong positive relationship between both financial segments and economic growth; no negative effects caused by share volatility or international capital market integration Financial development and finance-friendly legal framework enhance growth via innovation All financial sectors have a positive impact on growth, but shares have the greatest

Harris 1997

49 countries: 19 EU+; 2 ACC.

1980-1991

Real GDP per capita growth

Investment rate; 2SLS population regression growth analysis

Financial sectors and economic growth

Levine, Zervos 1998

Up to 47 countries

1976-1993

Real GDP growth; Capital stock & productivity Savings

Bank credit/GDP (bank credit = loans by commercial banks and other deposit-taking banks)

Capitalisation of domestic listed companies/ GDP; Value traded/ GDP; Share returns volatility

Initial output; Enrolment rate; Number of revolutions & other social & political variables

Crosscountry regression analysis

Links between financial sectors and economic growth

Leahy, Schich, Wehinger, 16 EU+ Pelgrin, Thorgeirsson 2001

1970-1997

Gross investment

Private credit from deposit banks/GDP; Liquid liabilities/GDP

Real long-term interest rate; output growth; inflation & Stock market its variation; public capitalisation/GDP revenue & spending; trade exposure; legal indicators

Panel error correction approach building on an ARDL model

Financial development, financial system, innovation & economic growth

9

IEF Working Paper Nr.49

Authors/Year of publication

Country sample

Time span

Dependent variable Mark-up Value added Return on investment

EFV I: Banking sector

EFV II: Stocks

Rivaud-Danset, Dubocage, 9 EU+ Salais 2001

1990-1996

Claessens, Laeven 2002

Up to 51 countries

1980-1989

Real value Private credit/GDP added growth

Other explanatory Control/ other variables variables Own funds Variables Leverage controlling for Financial debt structure company size Liquid capital Legal indicators, esp. dealing with property rights; CountryCash flow specific & industryspecific characteristics; Stock market capitalisation Value traded & turnover (to control for forwardorientation of financial markets)

Method

Investigated links between Companies' financing structures & performance

Major findings No correlation between financial variables & industry performance

Additional results Companies' financing structures depend on country characteristics

Cluster analysis

Panel regression (industry level)

Financial development & property rights, and growth of companies depending on external finance

Companies depending on external finance grow faster in economies with developed financial systems & high property protection

Asset allocation effect as important as finance effect. Poor (intellectual) property protection leads to less investment in intangible assets

Financial Minor causal links development, between financial financial system & 23 countries: development and Gross economic growth Hahn 19 EU+; 1 1970-2000 economic growth, investment controlling for 2002 ACC. strong relationship due forward-looking to forward-looking price effects of price effects financial markets Notes: Papers are ordered chronologically by year of publication. Slashes ("/") in table texts denote division, asterisk denotes multiplication; ”EU+” refers to EU plus 5 OECD countries. EFV denotes explanatory financial variables. ARDL: an autoregressive distributed lag model method, which builds on time series methods, but a panel is analysed. Bank credit to the private Stock market sector/GDP; capitalisation/GDP Liquid liabilities/GDP Panel error correction approach building on an ARDL model

10

Bond Markets and Economic Growth

After having offered a concise selection of the existing cross country studies, we address the time-series analyses on the finance-growth nexus (see table 4). As mentioned before, previous timeseries studies concentrate only on the banking sector and stock markets and neglect the debt markets, but they partly employ a similar testing methodology as utilized here. The focus on the banking and the stock market segments is attributable to the traditionally dominant market position of these financial sectors and the availability of data. The most commonly found pattern supports the supplyleading hypothesis, and in some cases bi-directional causality. The study of Hannson and Jonung (1997) is unique as it examines a single country (Sweden) over a very long time period, which makes it possible to test long-term causality patterns in the sense of Patrick (1966) or Gerschenkron (1962). The evidence based on cointegration analysis shows bi-directional causality between bank credit and real per capita GDP for most of the time from 1834 to 1991 and supply-leading causality for the 1890 to 1934 period. Al-Yousif (2002) applies a cross-country and a time-series approach (1970-1999) and finds bi-directional causality to be the dominant, yet not the only observable pattern. Rousseau and Wachtel (1998) find evidence for supply-leading finance in USA, Canada, UK, Sweden, and Norway for the 1871 to 1929 period. Arestis and Demetriades (1997) employ the classical Granger causality test together with system exogeneity analysis to investigate the links between the stock market capitalisation and per capita GDP in the USA and Germany based on quarterly data for a shorter period (1979-1991). In the case of Germany, they find evidence for supply-leading finance, and no causality in the USA. Kugler and Neusser (1998) use financial sector GDP of 13 OECD countries from 1960 to 1997 as an explanatory variable, and the GDP of the manufacturing sector as a dependent variable, and find a whole range of possible causality patterns varying across countries and time. The recent study by Shan, Morris and Sun (2001) covers 6 EU countries, three EU accession countries and China, and investigates the dynamic link between real per capita GDP growth and bank credits measured by the loans granted by commercial and other deposit-taking banks. They utilize the Granger causality test within a VAR setting with a conditioning set and find demand-leading, supplyleading, and bi-directional evidence.

11

IEF Working Paper Nr.49

Table 4: Time series studies on the links between banking intermediation, stocks and real growth
Authors/Year of publication Hansson, Jonung 1997 Al-Yousif 2002 Rousseau, Wachtel 1998 Country sample Time span Proxy for economic growth Real per capita GDP Real per capita GDP growth Real per capita GDP growth FV I: banking intermediation Bank credit to nonfinancial sector per capita M1, M2 / GDP Money base; various proxies for intermediation based on bank deposit and credit Germany: M2/GDP; USA: Domestic bank credit/GDP Stock market capitalisation/ GDP; Stock market volatility (16-month standard deviation of share prices) FV II: securities Method: Granger causality test together with Cointegration analysis (investment per capita as conditioning variable) Error correction model/ Panel data approach Granger causality tests within VAR framework Investigated links and causality between Major findings

Sweden 30 developing countries USA, UK, Canada, Sweden, Norway

1834-1991

1970-1999

Mostly unstable relationship between Co-evolution of banking intermediation and output; supply-leading intermediation & real 1890-1934; positive influence of income education on supply-leading pattern Strong evidence for bi-directional Banking intermediation causality; limited evidence for other and economic growth patterns Banking intermediation and economic growth Supply-leading in early phase of economic development Cross-country analysis oversimplifies results; links between financial sectors and growth are different in GER & USA: financial-to-real sector causality in GER; no evidence for unidirectional causality in USA. MGDP & FGDP are co-integrated in 7 countries; MTFP & FGDP are cointegrated more often; financial-to-real sector-causality only for USA, JPN and GER; inverse causality in some other countries; no evidence for causal ties in small countries.

1870-1929

Arestis, USA & Demetriades Germany 1997

1979-1991 (quarterly)

Real per capita GDP

System exogeneity analysis

Financial sectors and economic growth

Kugler, Neusser 1998

13 EU+

1960-1997

GDP of manufacturing industry (MGDP); Total factor productivity of manufacturing industry (MTFP) Real per capita GDP growth Bank credit/GDP (Bank credit = loans by commercial banks and other deposit-taking banks) Currency ratio = currency/M1; Broad money velocity = M2/ nominal GDP

Financial sector GDP (FGDP)

Granger causality tests

Financial sector & growth

10 countries Shan, Morris, (6 EU+ Sun 2001 China) Saudi Arabia, Turkey, UAE

1960-1998 max, 19821998 min. 1964-1993 (annual), UAE:19731993

VAR with conditioning set

Banking intermediation & economic growth

Bi-directional causality in five countries, demand-leading in three, no causality in two. Strong support for the supply-leading hypothesis in Turkey. UAE: short-run support. Saudi Arabia: no short-run relations, bi-directional causality over the long run.

Darrat, 1999

Percentage changes of real GDP in 1990 prices

VECM (Johansen procedure)

Financial deepening (esp. banking development) & growth

12

Bond Markets and Economic Growth

While the vast majority of research is concerned with the impact of banks and stocks on growth, studies on the bond-growth nexus are scarce. The few studies on the link between debt securities and growth can be grouped in four categories:
§ § § §

Theoretical articles, Empirical studies, Descriptive reports on bond market developments, And studies with a regional focus on Central and Eastern Europe (see Table 5). 3.3. Bond-market based forecasts of economic performance

Table 5: The link between debt securities and growth: Overview of the studies on debt securities
Category, Authors, Year of publication Theoretica l articles Focus Market segments and Sample coverage Primarily intermediation in up to 20 countries: 1860-1963. USA output growth: 1953-89. German economic crises: 1969-91. USA data Three policy regimes within the 1955-1996 period. Major findings Financial structurefinancial Goldsmith development1969 real development relationship. Harvey 1989 Harvey 1991 Forecasting the US economic growth via the yield curve. Forecasting German economic crises via the yield curve.

The ratios of financial assets to real wealth in different countries converge in the long run. Yield-curve based measures are able to explain > 30% of the variation in economic growth, while stock market measures account to about 5% only. The IRFM outperforms the forecasts of economic growth of major research institutes and predicts correctly the economic turning points during the surveyed period. Empirical evidence obtained from the Granger causality tests indicates that when the Fed sets policy to counter inflation or output growth movements, the slope of the yield curve loses its forecasting power with respect to the particular variable. Monetary policy is exogenous with respect to the yield curve. A simple univariate model is found to be a filter which transforms accurately term spread changes into turning point predictions, and that the term structure is a good recession indicator. Corporate debt issuance is positively related to M&A and industrial production. Substitution with other sources of finance is related to cost differentials. Timing and size of explanatory factors of corporate debt securities issuance differ across maturity. Corporate bond spreads lag shortterm interest rates and lead real economic growth.

Empirical studies

The information Gamber content of the 1996 yield curve. The predicting power of the term structure as a lead indicator of recessions. Corporate debt securities in the Euro area.

Ahrens 2002

Eight OECD countries: 1970:011996:12.

G. de Bondt, August 2002

Euro area corporate debt securities: 1999:01-2001:06.

13

IEF Working Paper Nr.49

Descriptive bank reports on recent developments

ECB 2001

Theoretical analysis of corporate finance issues and external sources of corporate finance.

Merrill Lynch, 2002

Size and structure of the world bond market 2002

Claes, De Primary market Ceuster, activity in the Polfliet Eurobond market. 2002

Recent trends in Mihaljek, debt markets Scatigna, development in the Villar emerging 2002 economies.

Eastern European Köke, capital markets: Schröder historical 2002 development & legal settings

Haiss, Marin 2002

CEEC bond market developments

Bond and stock The dominant liability in the USA and Europe in 1999 was markets, venture equity debt, in Japan-loans. Debt securities as financing capital market. form was not as important in Euroland as in the US and Banking Japan. External financing increased between 1997-1999. intermediation. Strong increases in debt securities and venture capital after Euro area in 1997. The value of equities increased most. According to comparison with flow data analysis this was mainly due to higher stock USA and Japan: prices and in the Euro area financing in loans exceeded 1997-2000. flows in shares. The world bond market size went up by 5% compared to last year. In spite of widening fiscal deficits, the share of central governments’ bonds continued to fall to 30%, a 12-year low, and was chiefly an English-speaking and Scandinavian countries phenomenon. In Japan and Euroland government markets continued to grow. Agencies were the fastest growing market in 2001, reflecting the strength of the housing market. Eurobonds ranked second due to their increasing popularity. The rating quality of the corporate sector continued to fall. The emerging markets tradable debt universe increased by 8% compared to 2001. Geographical concentration of the Eurobond market with Eurobond market respect to the nationality of issuers. More than 11% issued by (Bondware US-based entities. Introduction of the euro reinforced the Databse): 1980market share of European issues. Majority of Eurobonds are 2000. bearer bonds. Despite considerable growth primarily due to public sector deficits associated with fiscal adjustment and related banking Domestic debt and corporate sector reforms, domestic debt markets remain small compared to industrial countries. Public sector debt markets in Asia, Latin America and issues on average account for 2/3 of domestic debt market Emerging Europe: volume. Banks hold the largest proportion of bonds in 1994-2000. domestic markets; institutional investors have become key holders of domestic debt in Latin America and Central Europe; corporate debt markets remained underdeveloped. Stock markets grew unevenly across countries, with developments in the Czech Republic being particularly disappointing. Bond markets dominated by public sector debt securities (exception: the Czech Republic, where corporate Czech Republic, Hungary, Poland: bonds are significant).. The importance of internal finance for 1995-2000 firms’ investment is higher than in EU countries (exception: Hungary). Bank credit remains the main source of external finance even for listed firms.. Most CEE exchanges are too small to survive. The bond market is still dominated by public issuers and Central and bank-based intermediation. Similar to the EU, corporate bond Eastern European issuance was driven by the telecom and utility sectors. Rising Countries (CEEC) demand from pension funds and limited bank loans volume carry the potential for growth of the corporate bond sector.

(the difference between the long-term and the short-term yield) contain information about future real economic performance (Harvey, 1989, 1991; Bernanke, 1990; Estralla and Hardouvelis, 1991; Gamber, 1996; Ahrens, 2002). Using a standard ordinary least squares (OLS) regression linking real growth as a response variables with the yield spread (impulse variable), Harvey (1989) shows empirically that yield-curve based measures are able to explain more than 30% of the variation in economic growth in the USA over the 1953-89 period, while stock market measures account to about 5% only. Moreover, yieldcurve based forecasts are found to compare favorably with leading econometric models’ forecasts,

Studies dealing with Central and Eastern European debt markets

Over the last ten years, several studies have shown that the changes in the slope of the yield curve

14

Bond Markets and Economic Growth

while forecasts from stock market models do not (Harvey, 1989, p. 38). Harvey (1991) uses the interest rate based forecasting model (IRFM) to forecast the German economic crises during the 19691991 period. He utilizes the annual GNP growth in 1980 DM, the money-market rate and the government bond yield from 1969 till 1991. The first component of the IRFM is the yield curve, and the second is a measure of the average propensity to hedge. The intercept is interpreted as the expected level of economic growth when the long-term rate equals the short-term rate, and the coefficient of the term structure is the average level of risk tolerance in the economy. He shows that the IRFM ex post outperforms the forecasts of economic growth of major research institutes and that it has correctly predicted the economic turning points in the period 1969-1991. Gamber (1996) uses the difference between 10-year and 3-month government bond yields over the 1955-1996 period to investigate with Granger causality test whether the information content of the yield curve is only due to its relationship to monetary policy or whether it is an independent indicator of future inflation and output growth. He finds that the slope of the yield curve loses its forecasting power when the Federal Reserve Bank sets policy measures to counter movements in either inflation or output growth. Therefore, it seems that monetary policy is exogenous with respect to the yield curve. Ahrens (2002) uses Markov-switching models to estimate the predicting power of the term structure as a lead indicator of recessions in eight OECD countries between January 1970 and December 1996. He finds that a simple univariate model is a filter which accurately transforms term spread changes into turning point predictions, and that the term structure is a good recession indicator. Recently, G. De Bondt (2002) provided first evidence on the structure and dynamics of the corporate bond market in Euroland after the introduction of the euro. The time horizon of his study is rather short (1999:01-2001:06). The paper reviews three theoretical frameworks to model corporate bond issuance: the first one models simultaneously all corporate financial liabilities in a portfolio modeling framework, the second one models supply and demand for corporate bonds simultaneously, and the third one is to specify a supply function of corporate bonds. Due to data limitations the last approach is adopted. The financing needs and substitution between debt securities and other sources of corporate finance are considered as main explanatory factors for bond supply. The essential findings are: First, since the introduction of the single currency the euro area debt securities markets contributed to finance mergers and acquisitions (M&A), investment and working capital needs. The introduction of the euro triggered (at least partially) corporate restructuring, which in turn resulted in financing needs reflected by M&A. M&A are reflected in short-term bond issuance in the same month and with a one-quarter lag, and in long-term bond issuance after three quarters. A one-to-one relation between corporate debt securities issuance and industrial production is found. Second, the spread between long-term and short-term interest rates is found to be a relevant factor for the mix between long and short-term debt securities. The direct substitution effects are found to be particularly strong for short-term bond issuance. Third, regression results and Granger causality tests show that various

15

IEF Working Paper Nr.49

macroeconomic factors explain corporate bond spread movements and that the latter have forecasting power for real output growth. Our paper advances on earlier work and on the most recently published ECB working paper by De Bondt (2002) in several ways: Our time horizon encompasses 25 to 50 years with several business cycles. We look at the aggregate bond market instead of single segments and thus capture the aggregate effect of gross public and private investment financing. We use changes in stock variables instead of flow variables to account for valuation changes and implicitly for changes in the interest rate level, too. We employ the aggregate bond market capitalization instead of bond issuance or interest rate spreads to achieve comparability with earlier work on stock markets and real growth. Accordingly, our model consists of real GDP growth and real aggregate bond market capitalization. Theoretically, real capital stock growth might be used as an intermediate variable between financial markets and GDP growth. However, there is the usual problem to assemble reliable data on the real capital stock (taking into account depreciation rates and valuation changes over time). 4. Methodology The purpose of the paper is to examine the hypothesized causal relationship between the bond sector development and economic growth in 13 developed economies over the 1950 to 2000 time period. The direction of the causality is of a particular interest as it may help partly explain the driving forces for the emergence and development of financial markets. Thereby it can convey policy implications how to enhance continuing financial growth, which eventually leads to capital accumulation. The direction of the causality cannot be tested via cross-country approaches, as they usually assume a priori that finance influences growth (e.g. Darrat, 1999). The most popular procedure when testing causality between any two variables has become the Granger causality test (Granger, 1969). This is partly due to its simplicity, and due to the fact that this method saves degrees of freedom – this property is a great advantage when relatively small samples are utilized (n<50 observations). Granger (1969) defined causality in terms of predictability: a stationary time series, say X, is said to cause another stationary time series, Y, if the prediction error (measured by the variance) of current X declines by using past values of Y in addition to the past values of X. The question is how much of the current Y can be explained by its own past values and does an adding lagged value of X improve the explanation. Two-way causation, called “feedback”, is frequently the case, that is X Granger-causes Y and Y Granger-causes X. Conversely, Y is not Grangercausal for X, when removing the past values of Y from the model does not change the optimal forecast for X at any forecast horizon. It is important to note that the statement "X Granger-causes Y" does not imply that Y is the effect or the result of X. Granger causality measures precedence and information content, but does not indicate by itself causality in the more common use of the term. Formally, the bivariate autoregressive model looks like this:

16

Bond Markets and Economic Growth

 y t  n α 1,i b  = ∑ β  t  i =1  1,i

α 2,i   y t −i  u yt  + β 2,i   bt −i  u bt     

(1)

where β 1,i = 0, i = 1,2,..., n , the residual series are iid, u t ~ IN [o, σ 2 ] , and the time series variables are stationary, [ y t ,b t ] ~ I (0) . Nonstationarity of the variables incorporated in a model leads to the spurious regression problem discussed by Granger and Newbold (1974), Phillips (1986) and Watson (1994). High adjusted R 2 then may only indicate correlated trends instead of truly captured economic relations. Simple correlation tests including four lead and lag years show, that the GDP and the bond market capitalization are highly correlated. The correlation coefficients decrease from 0.99 in the first to a maximum of 0.5 to 0.7 in the fourth year. As both time series exhibit sustained growth trends over time such correlations are not unexpected. Granger (1986) illustrated that if the non-stationarity is purely due to unit roots, i.e. x t ~ I(1), then the time series can be made stationary by the linear transformation of differencing, i.e. x t − x t −1 = ∆x t ~ I(0). As usual in applied econometric work, this refers to the concept of weak stationarity (constant first moments) as opposed to strong stationarity. Lower-case letters denote natural logarithms of the non-stationary time series in levels. Defined this way, changes can be interpreted as growth rates in the level of real economic activity and bond market capitalization. The order of integration of the time series is determined via the Augmented Dickey and Fuller (ADF) and Phillips-Perron (PP) unit root tests. The ADF unit root test is most commonly used, but it requires homoscedastic and uncorrelated errors in the data generating process. The PP nonparametric test generalizes the ADF test and allows for less restrictive assumptions for the investigated time series (Darrat, 1999). Both tests are run to ensure unbiased inference about the order of integration of the variables of interest. The optimal lag order of the ADF regressions that guarantees uncorrelated residuals is selected via the Akaike’s Information Criterion (AIC), which minimizes the final prediction error (FPE) of the regression. Having determined the order of integration of the variables, equation (1) can be set up. The number of lags to be included in the system is selected by the AIC with an upper bound of 4 years (that is the assumed duration of one business cycle). A greater lag order is not practicable because of the low frequency of the data and would reduce degrees of freedom. The null hypothesis that bt does not Granger-cause y t is rejected if



∑α

2,i

≠ 0 . Whether the estimated coefficients are statistically

significant is checked with the F-test. By analogy, testing the reverse null hypothesis that y t does not Granger-cause bt is equal to testing whether

∑β

1,i

≠ 0.

17

IEF Working Paper Nr.49

However, these standard Granger-causality tests are valid only if the two variables are not cointegrated.4 The economic interpretation of co-integrated systems is that the series may exhibit shortterm dynamics but in the long run they converge to some state of equilibrium. As Engle and Granger (1987) showed, in the presence of co-integration inferences from the above system will be biased as they overlook long-run information. In this case, the Granger’s Representation Theorem postulates that it is better to set up an Error Correction Model (ECM). Engle and Granger (1987) proposed a single-equation based two-step approach to test for co-integration between two series. Alternatively, the system-based Johansen co-integration test can be used. Apparently, when investigating the interaction among three or more time series variables the problem of which test to use becomes redundant in favor of the second approach. Relying on Gonzalo and Lee (1998) and Haug (1996), the preference here is given to the Engle-Granger method. On the whole, Gonzalo and Lee (1998) find that the Engle-Granger two-step approach is more robust than the Johansen likelihood-ratio tests in most of the investigated cases. The main reason is that misspecifications of the variables’ longmemory components are found to affect their correlation structure more than their variances.5 Haug (1996) presents evidence for a trade-off between power and size distortions of the examined cointegration tests. Regarding size distortions, when the regressors are endogenous, the Engle-Granger’s ADF test and the Johansen maximum eigenvalue statistic exhibit the least size distortions. When the regressors are exogenous, the ADF test performs best. Based on these studies, the EG two-step procedure in connection with the Granger test for causality is used to identify the long-run equilibrium relation between the variables under investigation, if any, and to determine the direction of the shortrun causal ties. To determine whether there exists any long-run equilibrium relationship, we follow EngleGranger (1987) and fit the co-integrating regressions:
y t = β 0 + β 1bt + et , y and
bt = α 0 + α 1 yt + et , x ,

(2a) (2b)

where

y t is the logarithmic level of real economic activity measured by the real GDP and bt is the

size of the bond markets in the individual countries proxied by the outstanding amount of debt securities issued by central governments and the private and public sectors. To achieve a

4

If the variables of interest have a common stochastic trend, linear combinations of them are possibly stationary, i.e. there is no unit root in the relation linking Xt and Yt. The variables are said to be co-integrated. Formally: if Xt ∼ I(1) and Yt ∼ I(1), then αXt + βYt ∼ I(0). 5 Gonzalo and Lee (1998) show that the Johansen LR tests tend to find spurious co-integration with probability asymptotically approaching one in two cases: first, when variables with long-memory properties and trend behavior are involved that are not distinguishable from pure I(1) via standard unit root tests, and second, when the variables are non-stationary with a singular or near-singular error covariance matrix. Such a “size pitfall” is more serious than either a simple finite sample problem or a size distortion. When the Johansen procedure is used to analyze co-integrating relations among variables with deterministic linear trends, which are actually not co-integrated, the LR tests statistics diverge and the null is wrongly rejected. The EG test, on the other hand, is not sensitive to incorrect modeling of such components and produces an empirical size very close to the nominal size.

18

Bond Markets and Economic Growth

heteroscedasticity and autocorrelation consistent covariance matrix, Newey and West (1987) corrected standard errors are used instead of the conventional computed standard errors, as in the presence of heteroskedasticity the latter are no longer valid. If the residuals resulting from the above regressions {e t } are not stationary, the series are not co-integrated. The correct model to be specified is a standard VAR(p) in first differences with (p) being the lag order, as given by equation (1). Conversely, the variables are co-integrated in case of stationary residuals. The econometric model to be specified is a dynamic ECM. Essentially, it is a VAR augmented with an error-correction term. The EC-term is the residual series obtained from the associated co-integrating regressions and is the formal representation of the equilibrium. In addition to the independent lags, the error-correction term provides an extra channel for Granger-causality (Granger, 1986). In the transformed regressions, the Granger-causality restriction corresponds to the restriction that the EC-term does not enter the regression. Therefore, bt Granger-causes y t if

∑α

2,i

≠ 0 and/or the coefficient of the EC term is

statistically significant. It has been argued that the significant lags of bt represent short-run Grangercausality, and the coefficient of the error-correction term corresponds to long-run Granger-causality (Darrat, 1999). The empirical analyses should therefore focus on the long-run relationships as the investigated relationship is inherently a long-term one. To end with, it has to be noted, that a statistically significant coefficient of the EC-term does not allow for further inferences on the direction of the long-run causality. This difficulty is due the fact that the EC term results from a regression on two non-stationary variables. 5. Empirical Results The tests are in real terms in order to account for possible inflationary effects. The nominal data are deflated with the GDP deflator. All series are converted to USD using the 1995-dollar exchange rate to reflect real 1995 US dollars in order to achieve a better readability. The 1995 USD exchange rate is preferred to the current USD exchange rate in order to avoid the impact of exchange rate fluctuations on the time series variables. Otherwise the information on the size of the debt markets and the level of economic activity would be distorted. Subsequently, the series are transformed into natural logarithms to produce more homogeneous series. Due to insufficient number of observations, no tests could be performed on Denmark, Luxembourg, Greece, Ireland, Belgium and Norway. Thus, the sample comprises 13 countries. As expected, the ADF and Phillips-Perron unit-root tests performed on the logs of the series in levels show that all series are non-stationary in levels and stationary in first differences.6 Therefore we
6

To save space, the complete output of the unit root tests on the time series in levels and first differences (test specifications include an intercept and a linear trend), the co-integrating regressions and the ADF test on the residuals from the cointegrating regressions (test specifications do not include any exogenous variables due to assumed fluctuations around the zero line) are not reproduced here, but are available from the corresponding author upon request.

19

IEF Working Paper Nr.49

test for co-integration. The decision whether the time series variables are co-integrated or not, is taken according to the ADF test on the resulting residuals. The ADF test rejects the null of non-stationarity in Japan, Italy, Finland and Portugal. The evidence is that in these countries real bond market capitalization and real GDP are co-integrated, i.e. there is common equilibrium convergence. VAR models in first differences with a maximum lag order of four years are applied to seven countries. In the case of Great Britain, only the residuals from the first regression (with bond market capitalization being the dependent left hand-side variable) are found to be stationary, not the residuals from the reverse regression. Therefore, the evidence for co-integration is rather weak and not reliable enough to serve as a basis for an error-correction specification. No intertemporal relationship could be established for France and Sweden (Table 4). According to the AIC (it selects a lag order of zero) and the F-test on the coefficients of the lagged terms when null restrictions are imposed, all the lags are not significantly different from zero, thus filtering out possible intertemporal interrelations and short-run causality patterns. The impossibility to specify a dynamic system may be interpreted as an absence of causality patterns between the bond sector growth and real output growth based on annual data. It cannot be excluded that lags with duration of less than one year may be significant. Table 6: OLS test results for co-integration ordered by fit (R2). Model selection
Model/ Country Long-run equilibrium convergence JPN ITA FIN POR AUT Short-run dynamics No equilibrium convergence USA SUI E UK GER NL Undeter mined FRA SWE Sample size/ Observations 1967-1999/ 33 1963-2000/ 38 1972-2000/ 29 1977-2000/ 24 1960-2000/ 41 1952-2000/ 49 1952-2000/ 49 1962-1999/ 38 1976-2000/ 25 1960-2000/ 41 1951-2000/ 50 1977-2000/ 24 1964-2000/ 37 DV Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP Bond GDP ADF residuals test ADF (6) ADF (6) ADF (12) ADF (12) ADF (12) ADF (12) DF DF DF DF ADF (1) ADF (1) DF DF ADF (1) ADF (1) ADF (10) ADF (8) ADF (1) ADF (1) ADF (1) ADF (1) ADF (1) ADF (1) ADF (1) ADF (1) -3.573 ** -3.627 ** -3.370 * -3.342 * -3.914 ** -6.371 *** -3.419 ** -3.437 ** -2.005 -2.009 -1.471 -1.433 -1.544 -1.434 -3.114 -3.013 -7.554 *** -2.970 -2.493 -2.468 -1.654 -1.521 -1.059 -1.101 -2.609 -2.642 R2 .97834 .93149 .87314 .83382 .99049 .90866 .80547 .73765 .72745 .97503 .61109 .95751 .68060 DW Heterosced asticity Model VECM VECM VECM VECM ∆VAR(1) ∆VAR(1) ∆VAR(1) ∆VAR(1) ∆VAR(1) ∆VAR(2) ∆VAR(2) .227 .0111*** .230 .281 .0181*** .267 .329 .4300 .345 1.321*** .2661 1.122*** .389 .1869 .389 .045 .7994 .045 .181 .1874 .159 .409 .2822 .318 .285 .2339 .221 .339 .8797 .340 .052 .3281 .039 .582 .0406 *** .566 .182 .000006 *** .157

DV: denotes the dependent variable. DW: is the co-integrating regression Durbin-Watson test statistics. Heteroskedasticity: the White’s test for heteroskedasticity on the regression of squared residuals on fitted values (with cross terms). The reported values are p-values for rejection of the null hypothesis of no heteroscedasticity.

20

Bond Markets and Economic Growth

Relevant critical values for the Durbin-Watson statistic are tabulated by Engle & Yoo (1987): 1% (1.00), 5% (0.78), 10% (0.69 for 2 variables and sample size n=50. The ADF test regressions include an intercept and a trend. ADF critical values from MacKinnon (1991). ”***” denotes significance at the 1% level, “**” at 5%, “*” at 10%. The included lags are reported in brackets.

5.1. Evidence for long-run equilibrium convergence and interdependence In Japan, Italy, Finland, and Portugal real GDP and the bond market size follow a common stochastic trend in the long run, which is represented by the error-correction term (EC-term). As outlined in the previous section, a statistically significant coefficient of the EC-term can be interpreted as a long-run equilibrium adjustment and long-run causality. Table 7: Long-run equilibrium convergence VECM models
Country JPN DV
∆bt −1 ∆yt´−1 ∆bt −1 ∆bt −1
∆bt − 2

a -.231 .153 .217 .003 -.275 .042 .047 .314 .156 .818 .246 .00 -

EC-term y = .423b + 8.863

∆bt −3

∆bt − 4

∆yt −1
1.301 .00

∆yt − 2

∆yt −3

∆ yt − 4

const .045 .024 .018 .006 .124 .020 .024 .00 .091 .0002

R2 .762 .767 .784 .585 .579 .951 .526

.406 .00

ITA

∆yt´−1 ∆bt −1

y = .535b + 6.351

.179 .135

FIN

∆yt´−1
∆bt −1

y = .218b + 9.273

-.017 .645 -

FIN’ POR

y = .118b + 10.266

1.576 .0001 -.757 .0002 1.576 .00 -.326 .089 1.220 .008 .045 .582 1.376 .0001

-1.141 .001 .629 .0003 -1.095 .00 .21 .131 -.922 .019 -.026 .686 -1.134 .001

.323 .006 -.192 .0002 .279 .0004 -.049 .192 .265 .098 -.003 .838 .348 .005

1.724 .00

-1.382 .606

-1.816 .014 1.824 .00 -2.497 .0005 1.164 .0004 1.675 .708 1.634 .00 -

1.456 .008 -1.455 .00 1.557 .008 -.666 .016 -.766 .821 -1.214 .00 -

-.369 .013 .366 .00 -.334 .052 .164 .061 .227 .829 .362 .0005 -

-

∆bt −1 ∆yt´−1

Due to near-singular square matrix (i.e. it has a determinant of one and cannot be inverted) the model could not be established.

Notes: yt and bt denote the logarithms of the GDP level and bond market capitalization at time t, respectively; ∆ is the first-difference operator. Growth rates are calculated as ∆yt = ln yt − ln yt −1 . P-values are reported below the corresponding coefficients. DV: denotes dependent variable, a is the adjustment coefficient; EC-term: denotes co-integrating equation.

JPN: GDP growth adjusts to the long-run equilibrium state defined by the EC-term (the a is statistically significant). In the short run, there is evidence for bi-directional causality as all the coefficients on the bond market changes in the regression on GDP growth are statistically significant, and the coefficients on real output changes in the reverse regression are statistically significant also. Although ranked seventh with regard to its bonds/GDP ratio, the empirical evidence for Japan is that real growth follows the developments in the bond markets. ITA: Only bond market growth converges to the long-run equilibrium (a is statistically significant). The lagged GDP changes in the regression on the bond market growth are also significant, which suggests both long-run and short-run causal relationship from GDP to the bond market. Turning to GDP growth, only the second bond market lag is statistically different from zero. These findings can be interpreted as evidence in favor of the demand-leading hypothesis. FIN: The significant adjustment coefficient a in the regression on real GDP growth indicates long-run causality, whereas the bond market changes are not found to influence real GDP growth in the short-run. With bond market growth being the dependent variable, neither the adjustment coefficient nor the coefficients on the lagged changes are statistically significant. Therefore the EC-

21

IEF Working Paper Nr.49

term is dropped from the regression and the model is re-estimated. It can be observed that bond market growth depends on own lags only. POR: The model could not be set up, possibly due to the small number of observations (n = 24). The series on bond market capitalization starts in 1977 and start growing after 1987 (EU accession impact). 5.2. Evidence for short-run dynamics. One-way causality patterns The evidence from the short-run autoregressive models supports the supply-leading hypothesis. Table 8: Interdependence between bond market capitalization and GDP. Short-run VAR estimates
Country AUT DV

∆bt −1
.42*** (2.64) .1*** (2.65) .777*** (7.78) .164* (1.73) .009 (.06) .089*** (2.69) .254 (1.51) -.014 (-1.217) .447** (2.27) .096*** (2.6) .007 (.05) -.033 (-.503) .121 (.81) .019 (.95) .102 (.43) -.013 (-.33) .111 (.55)° .0006 (.034)

∆bt − 2
.048 (.31) .162*** (2.45) .306** (2.06) -.036* (-1.78) -.011 (.08) .055 (.21) .106 (.53)° -.011 (-.661)°

∆y t −1
.04 (.06) .123 (.81) -.187 (-1.21) -.027 (-.18) -.236 (-.43) .524*** (4.35) -.259 (-.15) .663*** (5.77) -1.273 (-1.46) .53*** (3.27) -.488 (-1.30) .356*** (2.19) -.675 (-.62) .469*** (3.16) -.708 (-.48) .395 (1.58)°°° -2.798 (-1.25)°° .516 (2.649)*** -

∆y t − 2

Const .07*** (2.57) .02*** (2.58) .015** (2.12) .026*** (3.73) .044* (1.99) .009* (1.82) .056 (.7) .012** (2.3) .058** (2.15) .006 (1.24) .073*** (3.10) .009 (.93) .063 (1.47) .025*** (4.27) .058 (1.27)°° .014 (1.83)** .075 (1.00)°° .018 (2.72)***

R2
.178 .181 .579 .063 .004 .362 .067 .527 .264 .470 .101 .244 .148 .228 .091 .163 .102 .230

SSR .216 .010 .024 .022 .504 .024 2.318 .010 .123 .004 .115 .022 .734 .014 .079 .002 1.198 .009

∆bt ∆y t ∆bt
USA

∆y t ∆bt

SUI

∆y t ∆bt

E

∆y t ∆bt

UK

∆y t ∆bt

GER

∆y t ∆bt

NL

∆y t ∆bt

FRA

∆y t ∆bt

SWE

∆y t

.608 (1.62) -.068 (-.42) -.233 (-.22) -.281* (-1.97) 1.715 (1.18)°° -.049 (-.20) 1.424 (.62)° -.296 (-1.48)°°°

Notes: yt and bt denote the logarithms of the GDP level and bond market capitalization at time t, respectively; ∆ is the firstdifference operator. Growth rates are calculated as ∆yt = ln yt − ln yt −1 . DV is the dependent-variable column, followed by the lagged differenced terms and the constant. Standard errors are reported in parentheses below the corresponding coefficients. Triple, double and single asterisk denote significance at p=.01 (or .02); .05 and .10, respectively. “°°°”, “°°” and “°” denote significance at p = .20; .40 and .60 respectively.

22

Bond Markets and Economic Growth

AUT: The coefficients of the bond market growth in the VAR system are statistically significant. The bivariate Granger causality test supports the hypothesis that GDP growth is caused by bond market development. USA: The regressions support the hypothesis of supply-leading finance. The high bonds/GDP ratio (USA is ranked fourth) could be the ultimate cause for this observation. SUI: Bond market capitalization changes seem to be independent from own lags or lagged output growth. The low R-squared and the high sum of squared residuals implies that the regression does not explain much of the variance in the growth of the bond market. On the other hand, real bond market capitalization is found to Granger-cause real economic activity. This is supported by the autoregressive representation, where all coefficients are statistically significant. E: More than 50% of the variation in real economic growth can be explained by the VAR (1) model, although the bond market growth has a minimal explanatory power for real output growth. UK: The residuals from the co-integrating regression with real bond market capitalization as response variable are stationary, whereas the residuals from the reverse regression are not. Since the critical values are computed for a sample size of 20 observations, they may not be reliable when only 16 observations are available (the ADF test has been performed with 8 lags). Thus it is ambiguous whether the series are co-integrated. Therefore the Johansen test has been performed on the levels of the series including two time lags as selected by the AIC. Both test statistics fail to reject the null of no co-integration. Based on these findings, the UK real output and the bond market are not co-integrated. A VAR with one lagged annual term on the changes is estimated. There is strong evidence for shortrun causality from bond market growth to real economic growth, obtained both from the VAR(1) and the Granger causality test. GER: There is supply-leading causality from the bond market development to real growth. The reverse causality from the real sector development to the bond market development is weak, as the coefficient on the second lag of the output changes is significant at the 20% level only. The NL: There is only weak evidence for Granger-causality from bond market to real output growth. According to the autoregressive representations, bond market capitalization changes can be explained by own two-year lags, whereas changes in real economic activity are strongly related to own lags and to the second lag of bond market growth. FRA and SWE: The VAR models are not significant. Causal links between bond markets and GDP growth cannot be analyzed with annual data. Table 9: Granger causality tests within the VAR systems
Country Model specification Included observations Austria USA SUI E VAR(1) VAR(1) VAR(1) VAR(1) n = 39 n = 47 n = 34 n = 36 UK VAR(1) n = 27 GER NL VAR(2) VAR(2) n = 21 n = 47 FRA SWE VAR(2) VAR(2) n = 21 n = 34

23

IEF Working Paper Nr.49

Null hypothesis H 0

Probability for null rejection .952 .008 .226 .084 .671 .007 .880 .222 .145 .009 .197 .044 .740 .175 .497 .918 .443 .803

∆y t does not cause ∆bt ∆bt does not cause ∆y t

Note: Bivariate tests for Granger causality within a VAR setting. ∆yt and ∆bt denote the growth rates of the GDP and the bond market capitalization, respectively.

6. Summary and discussion The causal patterns detected by the bivariate Granger causality test are summarized in table 10. Regarding the error-correction models it has to be noted, that although the Granger causality test does not support any causal patterns, in the presence of co-integration, there is causality in at least one direction (Granger, 1969). This applies to Japan, Italy and Finland. Since the Granger causality test is known for its sensitivity to lag length specification, and therefore its results need be interpreted in the context of the coefficients of the vector-autoregressive models. We were able to find a bi-directional causality, or interdependence between the bond sector growth and real growth, in the cases of Japan, Finland and Italy. There is support for supply-leading causality from bond market capitalization change to real growth in USA, Great Britain, Germany, Austria, Switzerland, and to a weaker extent in the Netherlands and Spain. Finally, we find no support for the reverse case, i.e. demand-leading causality from real economic activity to the bond market. Table 10: Causality patterns
∆y t ↔ ∆bt
Interdependence Japan Finland Italy

∆bt → ∆y t
Supply-leading approach Switzerland Austria UK Germany USA

∆bt → ∆y t
Weak causality Netherlands Spain

∆y t → ∆bt
Demand-leading approach No evidence found

Insufficiently determined Portugal France Sweden

The causality patterns differ from country to country due to the heterogeneity of market structures and the different degree of openness and international integration of the capital markets in our country sample. The data available for the individual countries differs substantially across time: the Netherlands has the longest time series (n = 50 observations), whereas for Portugal and France we only dispose of 24 observations (data start in 1977). For these reasons the time series analysis is a more suitable approach than panel data methods, as these would constrict the time sample and lead to generalized assertions, which normally do not allow for country-specific inference. The deficiency of the for some countries relatively short time sample is that it cannot be split to test for structural breaks and linearity in the relationships found. The latter is desirable considering the Portuguese and Spanish financial market development from autarky to more integrated financial sector after the EU entry in 1986. Utilization of higher frequency data instead of the annual data used in this study would offer a possibility to reject our findings or to find even stronger causal ties. To test the

24

Bond Markets and Economic Growth

obtained results for robustness, in a next step the same tests may be performed on the growth rates calculated as percentages instead on the first differences. Our research will also proceed by taking into account possible different financial sector influences on real economic activity. This will be done by splitting the aggregate bond market into corporate and government bonds. To make the picture complete, a further step could be to estimate the same relationship augmented by the influence of stock markets and bank credit on economic growth. As the log levels of the variables under investigation are co-integrated, the examined relationships imply level effects and not permanent growth effects. For the second to be true, the growth rate of real output needs be co-integrated with the level of the bond market capitalization. This case necessitates that the growth rate of output has the same order of integration as the level of real bond market capitalization, which is ruled out by the finding that both variables are difference stationary. The same research for the Central and Eastern European countries would convey very interesting insights for comparative and policy-making purposes. Due to unavailability of sufficiently long time series, this could not be fulfilled here, but this perspective provides also a positive and practical step for future research in international finance. 7. Conclusions This study has presented first evidence on the relationship between bond markets and real economic growth in the EU15, USA, Japan, Switzerland and Norway. The employed econometric techniques take into account stationarity, co-integration, and causality features of the variables of interest. When applied to aggregate domestic bond markets in most of the EU member states and USA, Japan, Switzerland and Norway, the techniques we use deliver well-specified relationships backed by an intuitive economic interpretation. The estimated models support the supply-leading hypothesis implying that real economic activity is significantly influenced by the development of the bonds market.

References
AHRENS, R. (2002), Predicting recessions with interest rate spreads: a multicountry regime-switching analysis, Journal of International Money and Finance 21, pp.519-537. AL-YOUSIF, Y., 2002, Financial development and economic growth: another look at the evidence from developing countries, Review of Financial Economics, Vol. 11, pp. 131-150. ARESTIS, P., DEMETRIADES, P. (1997), Financial Development and Economic Growth: Assessing the Evidence, The Economic Journal 107, pp.783-799. ATJE, R., JOVANOVIC, B. (1993), Stock Market and Development, European Economic Review, Vol. 37, pp.632640.

25

IEF Working Paper Nr.49

BATTEN, J., KIM, Y. (2000), Expanding Long-Term Financing Through Bond Market Development: A Post Crisis Policy Task, Working Paper No: 2000_07, School of Accounting and Finance, Deaking University, Ausralia, http://www.deakin.edu. BERNANKE, B.S. (1990), On the Predictive Power of Interest Rates and Interest Rate Spreads, Federal Reserve Bank of Boston, New England Economic Review, pp.51-68. BERTHÉLEMY, J.C., VAROUDAKIS, A. (1997), Financial Development and Growth Convergence: A Panel Data Approach, in Hausmann, R., H.Reisen (eds.), Promoting Savings in Latin America, Paris: OECD, pp.35-69. BISHOP, G. (2001), Delivering the Benefits of EMU, 4Q2000: Two Years of the Euro Bond Market, Schroder Salomon Smith Barney, London. BLUM, D., FEDERMAIR, K., FINK, G., HAISS, P. (2002), The financial sector nexus: Theory and empirical evidence, Institute for European Affairs, University of Economics and Business Administration, Vienna, Working Paper No. 43. CLAES, A., DE CEUSTER, M.J.K., POLFLIET, R. (2002), Anatomy of the Eurobond Market 1980-2000, European Financial Management, Vol. 8, No. 3, pp.373-386. CLAESSENS S., LAEVEN, L. (2002), Financial Development, Property Rights and Growth, CEPR Discussion Paper 3295. COASE, R. (1956), The Problem of Social Cost, Journal of Law and Economics, Vol. 3, pp.1-44. DARRAT, A.F. (1999), Are financial deepening and economic growth causally related? Another look at the evidence, International Economic Journal, Vol.13, Number 3, pp.19-35. DE BONDT, G. (2002), Euro area corporate debt securities market: First empirical evidence, European Central Bank Working Paper Series, Working Paper No. 164. DEMETRIADES, P.O., HUSSEIN, K.H. (1996), Does Financial Development Cause Economic Growth? Time Series Evidence from 16 Countries, Journal of Development Economics 51, pp.387-411. Economic Commission for Europe (ECE), Economic Survey of Europe 2000 No.1, Geneva. ENDERS, W., GRANGER, C.W.J. (1998), Unit-Root Tests and Asymmetric Adjustment With an Example Using the Term Structure of Interest Rates, Journal of Business & Economic Statistics, Vol. 16, No. 3, pp.304-311. ENGLE, R.F., YOO, B.S. (1987), Forecasting and testing in co-integrated systems, Journal of Econometrics 35, pp.143-159. ESTRALLA, A., HARDOUVELIS, G.A. (1991), The Term Structure as a Predictor of Real Economic Activity, Journal of Finance 46 (June), pp.555-576. European Central Bank, ECB (2001), Characteristics of corporate finance in the euro area, ECB Monthly Bulletin, February, 2001, pp.37-50. European Commission Directorate General for Economic and Financial Affairs (2002), Quarterly Bond Market Note – developments in the euro-denominated bond markets, No. 39, April-June 2002. FINK, G., FENZ, G. (2002), Did the successful emergence of the Euro as an international currency trigger its depreciation against the US dollar?, in El-Agraa, A., The Euro and Britain. Implications of moving into the EMU, Financial Times, Prentice Hall, Part II, pp.191-225. GAMBER, E.N. (1996), The Policy Content of the Yield Curve Slope, Review of Financial Economics, Vol. 5, No. 2, pp. 163-179. GERLACH, S., SMETS, F. (1997), The term structure of Euro-rates: some evidence in support of the expectations hypothesis, Journal of International Money and Finance, Vol. 16, No. 2, pp.305-321. GERSCHENKRON, A. (1962), Economic Backwardness in Historical Perspective: A Book of Essays, Cambridge: The Beltsnap Press of Harvard University Press. GOLDSMITH, R.W. (1969), Financial Structure and Development, New Haven: Yale University Press. GONZALO, J., LEE, T. (1998), Pitfalls in Testing for Long Run Relationships, Journal of Econometrics 86, pp.129-154. GRAFF, M. (2000), Finanzielle Entwicklung und reales Wirtschaftswachstum, Tübingen: J.C.B.Mohr (Siebeck). GRANGER, C.W.J. (1969), Investigating Causal Relations by Econometric Models and Cross-Spectral Methods, Econometrica 37, pp.424-438. GRANGER, C.W.J. (1986), Developments in the Study of Cointegrated Economic Variables, Oxford Bulletin of Economics and Statistics, 48, pp.213-228. GRANGER, C.W.J. (1993), What are we learning about the long-run?, The Economic Journal 103, pp.307-317. GRANGER, C.W.J. (1997), On Modelling the Long Run in Applied Economics, The Economic Journal 107, pp.169-177.

26

Bond Markets and Economic Growth

GRANGER, C.W.J., ENGLE, R.F. (1987), Co-Integration and Error Correction: Representation, Estimation, and Testing, Econometrica, Vol. 55, No. 2, pp.251-276. GRANGER, C.W.J., NEWBOLD, P. (1974), Spurious Regressions in Econometrics, Journal of Econometrics 2, pp.111-120. HAHN, F.R. (2002), The Finance-Growth Nexus Revisited. New Evidence From OECD Countries, Austrian Institute of Economic Research, WIFO Working Papers, No.176. HAISS, P., MARIN, S. (2002), Corporate Bonds als Finanzierungsinstrument in den Mittel- und osteuropäischen EU-Kandidatenländern, Bankarchiv, Vol. 50, No. 12, pp.981-986.. HANNSON, P., JONUNG, L. (1997), Finance an Economic Growth: The Case of Sweden, Research in Economics,Vol. 51, pp.251-301. HARRIS, R.D.F. (1997), Stock Markets and Development: A Re-assessment, European Economic Review, Vol. 41, pp.139-146. HARVEY, C.R. (1989), Forecasts of Economic Growth from the Bond and Stock Markets, Financial Analysts Journal, September-October 1989, pp.38-45. HARVEY, C.R. (1991), Interest Rate Based Forecasts of German Economic Growth, Review of World Economics, Weltwirtschaftliches Archiv, Zeitschrift des Instituts für Weltwirtschaft Kiel, Band 127, Heft 4, pp.701-718. HAUG, A.A. (1996), Tests for Cointegration: A Monte Carlo Comparison, Journal of Econometrics 71, pp.89115. HERRING, R., CHATUSRIPITAK , N.(2000), The case of the missing market: The Bond Market and Why it matters financial development, Asian Development Bank ADB Institute Working Paper (11), July 2000. JOHANSEN, S. (2000), Modelling of cointegration in the vector autoregressive model, Economic Modelling 17, pp.359-373. KING, R.G., LEVINE, R. (1992), Financial Indicators and Growth in a Cross-section of Countries, World Bank Policy Research Working Paper no. 819, Washington D.C. KING, R.G., LEVINE, R. (1993), Finance and Growth: Schumpeter Might Be Right, Quarterly Journal of Economics 108(3), pp.717-737. KÖKE, J., SCHRÖDER, M. (2002), The future of Eastern European capital markets, EIB Papers Vol. 7, No. 1/2002, pp.117-138, European Investment Bank, http://www.eib.org/efs/papers.htm. KUGLER, M., NEUSSER, K. (1998), Manufacturing Growth and Financial Development: Evidence from OECD Countries, The Review of Economics and Statistics, pp.638-46. LA PORTA, LOPEZ-DE-SILANES, SHLEIFER, VISHNY (1998), Law and Finance, Journal of Political Economy, Vol. 106, No. 6, pp.1113-1115. LEAHY, M., SCHICH, S., WEHINGER, G. PELGRIN, F., THORGEIRSSON, T. (2001), Contributions of Financial Systems to Growth in OECD Countries, OECD Economics Department Working Paper no. 280. LEVINE, R., RENELT, D. (1992), A Sensitivity Analysis of Cross-country Growth Regressions, American Economic Review, Vol. 82, pp.942-963. LEVINE, R. (1997), Financial development and economic growth: Views and agenda, Journal of Economic Literature, Vol. XXXV, pp.688-726. LEVINE, R. (1998), The Legal Environment, Bank, and Long-Run Economic Growth, Journal of Money, Credit, and Banking, Vol. 30, No. 3, pp.597-620. LEVINE, R., ZERVOS, S. (1998), Stock markets, bank, and economic growth, American Economic Review, June 1998, pp.537-558. LEVINE, R. (1999), Law, Finance, and Economic Growth, Journal of Financial Intermediation, Vol. 8, No. 1/2, pp.8-35. LEVINE, R., LOAYZA, N., BECK, T. (2000), Financial intermediation and growth: Causality and causes, Journal of Monetary Economics 46, pp.31-77. LUCAS R. (1988), On the Mechanics of Economic Development, Journal of Monetary Economics 22(1), pp.3-42. MACKINNON, J.G. (1991), Critical Values for Cointegration Tests, Chapter 13 in R.F. Engle and C.W.J. Granger (eds.), Long-run Economic Relationships: Readings in Cointegration, Oxford University Press. MC KINNON, R. (1973), Money and Capital in Economic Development, Washington D.C.: The Brookings Institution. MERRILL LYNCH (2002), Size and Structure of the World Bond Market: 2002. MIHALJEK, D., SCATIGNA, M., VILLAR, A. (2002), Recent trends in bond markets, BIS Papers No. 11, http://www.bis.org/publ/bispap11.htm.

27

IEF Working Paper Nr.49

MODIGLIANI, F., MILLER, M. (1958), The Cost of Capital, Corporation Finance and the Theory of Investment, American Economic Review, Vol. 48, pp.261-297. NEAIME, S., PASCHAKIS, J. (2002), The future of the dollar-euro exchange rate, North American Journal of Economics and Finance 13, pp.56-71. NELSON, Ch.R., PLOSSER, Ch.I. (1983), Trends and random walks in macroeconomic time series, Journal of Monetary Economics 10, pp.139-162. NEWEY, WHITNEY, K. WEST (1987A), Hypothesis Testing with Efficient Method of Moments Estimation, International Economic Review, 28, pp.777-787. NEWEY, WHITNEY, K. WEST (1987B), A Simple Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix, Econometrica, 55, pp.703-708. PATRICK, H. T. (1966), Financial Development and Economic Growth in Underdeveloped Countries, Economic Development and Cultural Change 15, pp.174-189. PHILLIPS, P.C.B. (1986), Understanding Spurious Regression in Econometrics, Journal ofEconometrics,33, pp.311-340. RIVAUD-DANSET, D., DUBOCAGE, E, SALAIS, R. (2001), Comparison between the financial structure of SMES and that of large enterprises (LES) using the BACH database, European Commission, Directorate General for Economic and Financial Affairs. ROBINSON, J. (1952), The Generalisation of the General Theory, in The Rate of Interest and Other Essays, 2nd ed., London: MacMillan, pp.1-76. ROUSSEAU, P.L., WACHTEL, P. (1998), Financial Intermediation and Economic Performance: Historical Evidence from Five Industrialized Countries, Journal of Money, Credit and Banking, Vol. 30, pp.658-678. SALA-I-MARTIN, X. (1997), I Just Ran 2 Million Regressions, American Economic Review, Vol. 87, pp.178-183. SCHLITZER, G. (1996), Testing the Null of Stationarity Against the Alternative of a Unit Root: An Application to the Italian Post-War Economy, Applied Economics 28, pp.327-331. SCHUMPETER, J., (1911) 1952, Theorie der wirtschaftlichen Entwicklung, 5. Auflage, Berlin: Duncker und Humbolt. SCHWERT, G. W. (1987), Effects of model specification on tests for unit roots in macroeconomic data, Journal of Monetary Economics 20, pp.73-103. SHAN, J.Z., MORRIS, A.G., SUN, F. (2001), Financial Development and Economic Growth: An Egg-and-Chicken Problem?, Review of International Economics, Vol. 9, No. 3, pp.443-454. SHAW, E.S. (1973), Financial Deepening in Economic Development, New York: Oxford University Press. SIMS, C. A. (1972), Money, Income and Causality, The American Economic Review 62, pp.540-552. WATSON, M.W. (1994), Vector autoregressions and cointegration. In R.F.Engle and D.L.McFadden (eds), Handbook of Econometrics, Amsterdam: Elsevier, Vol. 4, Ch. 47, pp.2843-2915. WU, Jhy-Lin, SUIN, Show-Lin (2001), Mean Reversion of interest rates in the Eurocurrency market, Oxford Bulletin of Economics and Statistics, 63, 4, pp.459-473. WACHTEL, P. (2001), Growth and Finance: What Do We Know and How Do We Know It?, International Finance 4:3, pp.335-362.

28

Bond Markets and Economic Growth

Data Appendix The USA, Japan, Switzerland, Norway and the 15 EU member countries are individually examined. The aggregate bond market capitalization is drawn as a measure for the size and volume of bond markets. The nominal data are deflated by the means of the GDP deflator. The surveys on the impact of the bond markets on GDP growth are conducted in real terms in order to filter out possible inflationary effects. GDP data are from the IFS of the IMF, and the OECD Historical Statistics (19601997, 1998 Paris) and National Accounts of OECD Countries, Volume 1 (Main Aggregates, Paris, diverse issues), and are drawn from the WIFO database (http://www.wifo.at/db/index.html). The nominal GDP time series are from the IFS (99B). The real GDP series until 1969 are converted by transforming the nominal GDP from the IFS with the GDP-deflator from the IFS (99BVP). The real GDP series from 1970 to 1999 are drawn from the OECD National Accounts (FIN"B1_GE_VOB). The OECD Historical Statistics index is used for four countries: Norway (till 1965), Austria (till 1963), Denmark (till 1965), and Portugal (till 1968). The data on bond markets capitalization from 1990 to 2000 comes from the BIS. Data until 1989 is composed as follows:

29

IEF Working Paper Nr.49

Table A1: Sources of bond market capitalization data before 1990
Country Time horizon 1950-1989 USA 1990-2000 1963-1989 JPN 1990-2000 1974-1989 NOR 1974-1989 1990-2000 1950-1959 1960-1989 SUI 1960-1989 1990-2000 1953-1973 1974-1989 AUT 1974-1989 1990-2000 1980-1989 BEL 1990-2000 1972-1989 FIN 1990-2000 1977-1989 FRA 1977-1989 1990-2000 1950-1989 1950-1989 GER Series Source BIS (EDBAUS01: securities outstanding) * All bond types BIS IFS (row 26AB) * All bond types BIS Financial, corporate & foreign bonds IFS * Public sector GFS * All bond types BIS Bonds IFS (row 26) * Financial, corporate & foreign bonds IFS * Public sector GFS (total public debt by type of instrument) * All bond types BIS Bonds IFS (row 26) * Financial & corporate sector IFS * Public sector, foreign bonds GFS (total foreign debt) * ALL BOND TYPES BIS All bond types All bond types Financial & corporate sector Public sector All bond types Financial, corporate & foreign bonds Public sector BIS (EDNABE01: securities outstanding, bonds, total)* BIS (EDNAFI91: securities outstanding, bonds, total)*

IFS * GFS (total domestic debt) * BIS IFS GFS (1950-1974: total government debt; 1975-1989: total public debt) 1990-2000 All bond types BIS 1963-1977 All bond types IFS * 1978-1989 Financial, corporate & foreign bonds IFS * ITA Public sector GFS (domestic government bonds & bills) * 1990-2000 All bond types BIS 1951-1964 All bond types GFS (domestic government securities) * 1965-1989 Financial, corporate & foreign bonds IFS * NL 1965-1989 Public sector GFS * 1990-2000 All bond types BIS 1977-1989 IFS * POR All bond types 1990-2000 BIS 1963-1989 Financial, corporate & foreign bonds IFS * E 1963-1989 Public sector GFS (domestic public debt) * 1990-2000 All bond types BIS 1964-1969 All bond types IFS * 1970 All bond types GFS (domestic public debt) * SWE Financial, corporate & foreign bonds IFS (row 46AB) * 1971-1989 Public sector GFS (domestic public debt) * 1990-2000 All bond types BIS National sources (outstanding amounts of government 1976-1989 UK All bond types bonds) * 1990-2000 BIS Note: Asterisk denotes series that partly consist of original data in value terms. Missing data have been calculated with the help of index time series which were made available mostly by the same reporting institution.

30

Bond Markets and Economic Growth

Bisher erschienene IEF Working Papers
1 2 3 4 Gerhard Fink, A Schedule of Hope for the New Europe, Oktober 1993. Gerhard Fink, Jutta Gumpold, Österreichische Beihilfen im europäischen Wirtschaftsraum (EWR), Oktober 1993. Gerhard Fink, Microeconomic Issues of Integration, November 1993. Fritz Breuss, Herausforderungen für die österreichische Wirtschaftspolitik und die Sozialpartnerschaft in der Wirtschafts- und Währungsunion, November 1993. Gerhard Fink, Alexander Petsche, Central European Economic Policy Issues, July 1994. Gerhard Fink, Alexander Petsche, Antidumping in Österreich vor und nach der Ostöffnung, November 1994. Fritz Breuss, Karl Steininger, Reducing the Greenhouse Effect in Austria: A General Equilibrium Evaluation of CO2-Policy-Options, March 1995. Franz-Lothar Altmann, Wladimir Andreff, Gerhard Fink, Future Expansion of the European Union in Central Europe, April 1995. Gabriele Tondl, Can EU's Cohesion Policy Achieve Convergence?, April 1995. Jutta Gumpold, Nationale bzw. gesamtwirtschaftliche Effekte von Beihilfen insbesondere Exportbeihilfen, April 1995. Gerhard Fink, Martin Oppitz, Kostensenkungspotentiale der Wiener Wirtschaft - Skalenerträge und Kostendruck, August 1995. Alexander Petsche, Die Verfassung Ungarns im Lichte eines EU-Beitritts, September 1995. Michael Sikora, Die Europäische Union im Internet, September 1995. Fritz Breuss, Jean Tesche, A General Equilibrium Analysis of East-West Migration: The Case of Austria-Hungary, January 1996. Alexander Petsche, Integrationsentwicklung und Europaabkommen EU Ungarn, Juli 1996. Jutta Gumpold, Die Ausfuhrförderung in der EU, Juni 1996. Jutta Gumpold, Internationale Rahmenregelungen zur Ausfuhrförderung, Juni 1996. Fritz Breuss, Austria's Approach towards the European Union, April 1996. Gabriele Tondl, Neue Impulse für die österreichische Regionalpolitik durch die EU-Strukturfonds, Mai 1996. Griller, Droutsas, Falkner, Forgó, Klatzer, Mayer, Nentwich, Regierungskonferenz 1996: Ausgangspositionen, Juni 1996. Stefan Griller, Ein Staat ohne Volk? Zur Zukunft der Europäischen Union, Oktober 1996. Michael Sikora, Der „EU-Info-Broker“ – ein datenbankgestütztes Europainformationssystem im World Wide Web über die KMU-Förderprogramme der Europäischen Kommission, November 1996. Katrin Forgó, Differenzierte Integration, November 1996. Alexander Petsche, Die Kosten eines Beitritts Ungarns zur Europäischen Union, Januar 1997.

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

23 24

31

IEF Working Paper Nr.49

25

26

27 28 29 30 31 32

33 34 35

36 37 38

39 40 41

42 43

Stefan Griller, Dimitri Droutsas, Gerda Falkner, Katrin Forgó, Michael Nentwich, Regierungskonferenz 1996: Der Vertragsentwurf der irischen Präsidentschaft, Januar 1997. Dimitri Droutsas, Die Gemeinsame Außen- und Sicherheitspolitik der Europäischen Union. Unter besonderer Berücksichtigung der Neutralität Österreichs, Juli 1997. Griller, Droutsas, Falkner, Forgó, Nentwich, Regierungskonferenz 1996: Der Vertrag von Amsterdam in der Fassung des Gipfels vom Juni 1997, Juli 1997. Michael Nentwich, Gerda Falkner, The Treaty of Amsterdam. Towards a New Institutional Balance, September 1997. Fritz Breuss, Sustainability of the Fiscal Criteria in Stage III of the EMU, August 1998. Gabriele Tondl, What determined the uneven growth of Europe´s Southern regions? An empirical study with panel data, März 1999. Gerhard Fink, New Protectionism in Central Europe - Exchange Rate Adjustment, Customs Tariffs and Non-Tariff Measures, Mai 1999. Gerhard Fink, Peter Haiss, Central European Financial Markets from an EU Perspective. Review of the Commission (1998) Progress Report on Enlargement, Juni 1999. Fritz Breuss, Costs and Benefits of EU Enlargement in Model Simulations, Juni 1999. Gerhard Fink, Peter R. Haiss, Central European Financial Markets from an EU Perspective. Theoretical aspects and statitical analyses, August 1999. Fritz Breuss, Mikulas Luptacik, Bernhard Mahlberg, How far away are the CEECs from the EU economic standards? A Data Envelopement Analysis of the economic performance of the CEECs, Oktober 2000. Katrin Forgó, Die Internationale Energieagentur. Grundlagen und aktuelle Fragen, Dezember 2000 Harald Badinger, The Demand for International Reserves in the Eurosystem, Implications of the Changeover to the Third Stage of EMU, Dezember 2000. Harald Badinger, Fritz Breuss, Bernhard Mahlberg, Welfare Implications of the EU’s Common Organsiation of the Market in Bananas for EU Member States, April 2001 Fritz Breuss, WTO Dispute Settlement from an Economic Perspective – More Failure than Success, Oktober 2001. Harald Badinger, Growth Effects of Economic Integration – The Case of the EU Member States, Dezember 2001. Gerhard Fink, Wolfgang Koller, Die Kreditwürdigkeit von Unternehmen im Hinblick auf die Wirtschafts- und Währungsunion - Wien im österreichischen Vergleich, Dezember 2001. Harald Badinger, Gabriele Tondl, Trade, Human Capital and Innovation: The Engines of European Regional Growth in the 1990s, Januar 2002. David Blum, Klaus Federmair, Gerhard Fink, Peter Haiss, The Financial-Real Sector Nexus: Theory and Empirical Evidence, September 2002.

32

Bond Markets and Economic Growth

44 45 46

47 48

Harald Badinger, Barbara Dutzler, Excess Reserves in the Euroystem: An Economic and Legal Analysis, September 2002. Gerhard Fink, Nigel Holden, Collective Culture Shock: Constrastive Reactions to Radical Systemic Change, Oktober 2002. Harald Badinger, Fritz Breuss, Do small countries of a trade bloc gain more of its enlargement? An empirical test of the Casella effect for the case of the European Community, Oktober 2002. Harald Badinger, Werner Müller, Gabriele Tondl, Regional convergence in the European Union (1985-1999). A spatial dynamic panel analysis, Oktober 2002. Harald Badinger / Fritz Breuss, What Has Determined the Rapid Post-War Growth of Intra-EU Trade?, February 2003.

33

IEF Working Paper Nr.49

Bisher erschienene Bände der Schriftenreihe des Forschungsinstituts für Europafragen
(Zu beziehen über den Buchhandel)

1 2 3

4

5

6

7 8

9

10

11

12 13 14

Österreichisches Wirtschaftsrecht und das Recht der EG. Hrsg von Karl Korinek/Heinz Peter Rill. Wien 1990, Verlag Orac. XXIV und 416 Seiten. (öS 1.290,-) Österreichisches Arbeitsrecht und das Recht der EG. Hrsg von Ulrich Runggaldier. Wien 1990, Verlag Orac. XIII und 492 Seiten. (öS 1.290,-) Europäische Integration aus österreichischer Sicht. Wirtschafts-, sozial und rechtswissenschaftliche Aspekte. Hrsg von Stefan Griller/Eva Lavric/Reinhard Neck. Wien 1991, Verlag Orac. XXIX und 477 Seiten. (öS 796,-) Europäischer Binnenmarkt und österreichisches Wirtschaftsverwaltungsrecht. Hrsg von Heinz Peter Rill/Stefan Griller. Wien 1991, Verlag Orac. XXIX und 455 Seiten. (öS 760,-) Binnenmarkteffekte. Stand und Defizite der österreichischen Integrationsforschung. Von Stefan Griller/Alexander Egger/Martina Huber/Gabriele Tondl. Wien 1991, Verlag Orac. XXII und 477 Seiten. (öS 796,-) Nationale Vermarktungsregelungen und freier Warenverkehr. Untersuchung der Art. 30, 36 EWG-Vertrag mit einem Vergleich zu den Art. 13, 20 Freihandelsabkommen EWG - Österreich. Von Florian Gibitz. Wien 1991, Verlag Orac. XIV und 333 Seiten. (öS 550,-) Banken im Binnenmarkt. Hrsg von Stefan Griller. Wien 1992, Service Fachverlag. XLII und 1634 Seiten. (öS 1.680,-) Auf dem Weg zur europäischen Wirtschafts- und Währungsunion? Das Für und Wider der Vereinbarungen von Maastricht. Hrsg von Stefan Griller. Wien 1993, Service Fachverlag. XVII und 269 Seiten. (öS 440,-) Die Kulturpolitik der EG. Welche Spielräume bleiben für die nationale, insbesondere die österreichische Kulturpolitik? Von Stefan Griller. Wien 1995, Service Fachverlag. Das Lebensmittelrecht der Europäischen Union. Entstehung, Rechtsprechung, Sekundärrecht, nationale Handlungsspielräume. Von Michael Nentwich. Wien 1994, Service Fachverlag. XII und 403 Seiten. (öS 593,-) Privatrechtsverhältnisse und EU-Recht. Die horizontale Wirkung nicht umgesetzten EU-Rechts. Von Andreas Zahradnik. Wien 1995, Service Fachverlag. (öS 345,-) The World Economy after the Uruguay Round. Hrsg von Fritz Breuss. Wien 1995, Service Fachverlag. XVII und 415 Seiten. (öS 540,-) European Union: Democratic Perspectives after 1996. Von Gerda Falkner/ Michael Nentwich. Wien 1995, Service Fachverlag. XII und 153 Seiten. (öS 385,-) Rechtsfragen der Wirtschafts- und Währungsunion. Hrsg von Heinz Peter Rill und Stefan Griller. Wien 1997, Springer Verlag Wien/New York, 197 Seiten.

34

Bond Markets and Economic Growth

15

16

17 18

19

20

23

The Treaty of Amsterdam – Facts Analysis, Prospects. Von Stefan Griller, Dimitri P. Droutsas, Gerda Falkner, Katrin Forgó, Michael Nentwich. Wien 2000, Springer Verlag Wien/New York, 643 Seiten. Europäisches Umweltzeichen und Welthandel. Grundlagen, Entscheidungsprozesse, rechtliche Fragen. Von Katrin Forgó. Wien 1999, Springer Verlag Wien/New York 1999, 312 Seiten. Interkulturelles Management. Österreichische Perspektiven. Von Gerhard Fink, Sylvia Meierewert (Hrsg.), Springer Verlag Wien/New York, 2001, 346 Seiten Staatshaftung wegen Gemeinschaftsrechtsverletzung: Anspruchsgrundlage und materielle Voraussetzungen. Zugleich ein Beitrag zur Gemeinschaftshaftung, Von Birgit Schoißwohl, Springer Verlag Wien/New York, 2002, 512 Seiten. The Bananana Dispute: A Comprehensive Legal Analysis supplemented by an Economic Analysis of Welfare Effects. Von Fritz Breuss, Stefan Griller, Erich Vranes (Hrsg.), ca 300 Seiten (erscheint demnächst, 2002). External Economic Relations and Foreign Policy in the European Union, Von Stefan Griller, Birgit Weidel (Hrsg.), Springer Verlag Wien/New York, 2002, 500 Seiten. Institutional, Legal and Economic Aspects of the EMU, Von Fritz Breuss, Gerhard Fink, Stefan Griller (Hrsg.), Springer Verlag Wien / New York 2003, 346 Seiten.

35

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close