University of Groningen
Faculty of Economics and Business
Stock Markets, Monetary Policy and
Financial Integration in the Eurozone
2
Abstract
Keywords: stock market returns, monetary policy and financial integration.
3
I.
Introduction
In response to the financial crisis of 2008 and European Sovereign Debt crisis, central banks around the world have aggressively intervened to prevent a collapse of the financial system and to counter the recessionary spill-over caused by these crises on the real economy. The interventions produced interest rates that are near the zero-lower bound. Moreover, central banks have expanded their balance sheet with the use of unconventional monetary policy such as quantitative easing programmes (QE). These actions appeared to pay off as economies that where affected by these crises have recovered or are recovering. Stock markets around the world for instance, generally taken to be an indicator of the state of the economy, have reached all-time highs during the past year.
Even though a seeming economic recovery is taking place in major developed countries, current monetary policy receives lots of criticism. Numerous influential economist, investors and politicians such as Carl Icahn (CNBC 2015), Donald Trump (CNBC 2015), George Soros (CNBC 2015) Hans W. Sinn (Financial Times 2015), Peter Schiff (CNBC 2015) and Robert Shiller (Financial Times 2015) have casted their concerns about current monetary policy. They fear that exceptionally low interest rates and QE programs lead to excessive risk taking behaviour and creation of new asset bubbles. Freshly created credit that is injected into economies by central banks, does not necessarily have to pour into the real economy (Block, 2012). It may well flow into the financial sector, triggering an artificial increase in asset prices and a decoupling between market prices of stocks and their underlying macroeconomic fundamentals.
4 line with underlying macroeconomic fundamentals, as predicted by traditional finance theory, or if stock markets are severely influenced by recent aggressive monetary policy.
Expanding this into optimum currency area perspective, such as the European Economic and
Monetary Union (EMU), is even more interesting. Inspired by Mundell’s (1961) theory of
optimal currency area (OCA) several European countries formed the EMU, a currency area with one central bank and hence, common monetary policy. When business cycles across members are not synchronized, a single monetary policy is not advantageous. For economies in recession an expansionary monetary policy is favourable. Contrary, growing economies risk overheating and asset bubble creation (de Grauwe 2013). Merely, the usual focus of research is on business cycles and the financial cycle is not looked upon. Because forming an OCA, entails financial integration (e.g., Ingram 1969; Mundell 1973; Rose and Engel 2002; Mongelli 2005), monetary policy should have similar effects on stock markets of the EMU-members, if there are any in the first place. This would imply a synchronized financial cycle and a high degree of financial integration in the EMU (de Grauwe 2012).
To the best of my knowledge, research has not investigated whether recent stock market returns in the EMU are in line with macroeconomic fundamentals or that these returns are driven by monetary policy. Thus far, numerous studies investigated whether or not stock market returns reflect macroeconomic fundamentals, as proxied by industrial production and inflation (e.g. Chen et al. 1987; Velinov and Chen 2015). In addition, these studies have only sparsely, included monetary policy (e.g. Asprem 1987). Further, focus was either on separate countries such as the US (e.g. Fama 1981; Chen et al. 1986) or on a group of big industrial countries (e.g. Mandelker and Tandon 1985; Asprem 1987; Laopadis 2006, 2011; Velinov and Chen 2015). This study aims to fill this gap in literature and attempts to provide an answer to the following two research questions. First:: “Are stock market returns in line with
macroeconomic fundamentals, or are they severely influenced by monetary policy?”. Second,
to examine the presence of a symmetric financial cycle the following question is answered:
“Is there an equal response of EMU stock markets to changing financial conditions and monetary policy?”.
5 SUR is that instead of running independent OLS regressions for each country, SUR runs several linear regression equations at the same time. Further, it allows the error term to be correlated among independent regressions, creating more efficient estimations than under OLS.
This paper makes several contributions to the literature as it finds that economic fundamentals such as industrial production and inflation appear to be insignificant in explaining stock market returns across the EMU. Furthermore, growth in narrow and broad money supply tends to have an opposite effect on stock market returns. At the country level inconclusive evidence is obtained for the impact of growth in money supply on stock market returns. Stock market returns across twelve initial EMU-countries mainly respond to changes in short-term and hence, ECB’s key interest rates. Long term-interest changes are significant at the aggregate level, though different effects are found before and during the crisis. Finally, the financial cycle moved quite synchronized across the EMU. However, the Euro crisis appears to have caused an asymmetric movement of the financial cycle across the EMU. Hence, financial integration is reduced, which lessens the effectiveness of monetary policy.
Policymakers should be aware of the fact that stock market returns (i.e. assets prices) are severely influenced by movements in short-term interest rates and hence ECB’s key interest rate. These effects have become more intense during the crisis. Macro-economic
fundamentals appear to have insignificant impact on stock market returns. Therefore
movements in interest rates might inflate stock market returns and asset prices. Inflating asset prices stimulates boom and bust dynamics in the economy (Terry Jones 2014). Further, policy makers should address the asymmetric response of financial markets to movements in short-term interest rates and ECB’s key interest rate. The euro crisis seems to have broken down the symmetric movement of the financial cycle and turned it into an asymmetric movement. Asymmetric financial cycles reduces the effectiveness of monetary policy in an OCA, hence restoring the symmetric financial cycles is of importance for the robustness of the EMU (Draghi 2014).
6 section VI. Finally, section VII provides a brief summary of main findings and subsequently concludes.
II.
Literature review
This section discusses past findings of literature on channels through which stock markets are influenced. Subsequently, after each subsection related hypotheses are presented.
Macro-economic variables influencing stock market returns
The fundamental value of a firms share is computed as the firm’s future stream of discounted cash flows (dividends) (Fama and French 1988; Laopadis 2011). Therefore, changing macroeconomic conditions affect the amount of cash available for distribution of dividends Thus, in rational markets; the fundamental value of a stock should be priced according to macroeconomic conditions (Laopadis 2011). In markets where stocks are under-priced relative to their fundamental value, returns tend to be high in the period thereafter while the opposite holds when stocks are overpriced (Campbell and Shiller 1988). In efficient securities markets, the efficient market hypothesis (EMH) states that assets are valued on all available information and hence, to their performance and expected future earnings (Asprem 1987). Stock markets are perceived to be forward looking, and present an indication of future economic activity. Hence, in states of changing macroeconomic conditions, stock prices should immediately adjust to embody the new economic environment and future prospects of the firm. In valuation theories such as Arbitrage Pricing Theory (APT) and intertemporal Capital Asset Pricing Model (CAPM), macroeconomic variables can be incorporated as factors determining assets prices (Chen et al. 1986; Asprem 1989).
According to Fama (1990) and Schwert (1990) there are at least three explanations for the strong relationship between macroeconomic fundamentals and stock movements. First, stock prices are a leading indicator as information about real future activity is thought to be priced into shares well before it occurs. Second, demand for consumption and investment goods can be affected by changes in stock prices. Changes in stock prices are changes in wealth. Third, although stock prices and real investments might be affected similarly by changing discount rates, the output from the real investments made cannot instantly be obtained.
7 1990). The most common proxies for economic activity used in literature are growth rates in industrial production and inflation. Results obtained for impact of growth in industrial production on stock market returns are mixed. While some studies find that current growth of industrial production increases stock returns (e.g. Chen et al. 1986; Nasseh and Strauss 2000; Humpe and Macmillan 2009), others (e.g. Lee 1995; Chung and Lee 1998) were not able to find any significant evidence that growth in industrial production has an impact on stock prices. Several scholars (e.g. Carlson and Sargent 1997; Binswanger 2004; Shiller, 2005; Groeneworld 2006; Laopadis 2006) argue that during the past two decades increases in stock market prices in Australia, Japan, several European countries and the US, could not merely be attributed to more positive future economic fundamentals. Hence, it is suggested that there is an exogenous effect that inflated the stock markets causing a decoupling between the market share price and their underlying fundamentals. Laopadis (2011) argues that the introduction of the euro triggered a decoupling between domestic macroeconomic fundamentals and stock market returns. Nonetheless, Velinov and Chen (2015) found that since 2008, stock prices in France, Germany and Italy do fall in line with their fundamentals. Furthermore, the steep rise in equity prices in the mid-90 was due to structural undervaluation in prior periods, and that prior to the financial crisis of 2008 stock prices were slightly overvalued.
Following the line of literature, macro-economic activity shall be proxied by industrial production. Therefore:
Hypothesis 1a:
An increase in industrial production (as a proxy for macro-economic activity) leads to an increase in stock market returns.
Inflation is another widely investigated macro-economic variable. An increase in inflation affects the level of real production and thus levels of real cash flows, which in turn affect the real value of the stock price (Chen et al. 1986). Several authors found a negative impact of inflation on stock market returns (e.g. Fama 1981 and 1990; Asprem 1989; Wongbangpo and Sharma 2001). In addition, higher levels of inflation directly affect market interest rates (Laopadis 2011).
Hypothesis 1b:
8 Financial variables influencing stock market returns
Financial variables such as short-term (discussed in next section) and long-term interest rates found in financial markets are also influencing stock market returns. Current share prices are calculated as the firm’s future stream of discounted dividends (Fama and French 1988). The discount factor depends both on long-term interest rates i.e. the risk free rate, and on a certain risk premium (Chen et al. 1986). Therefore, an increase in either one of them reduces the time value of future cash flows and thus current stock prices. Increasing long-term interest rates have a negative impact on stock returns (e.g. Nasseh and Strauss 2000; Abugri 2006; Humpe and Macmillan 2009). Lee (1995) and Chung and Lee (1998) on the other hand, did not find any significant evidence that the changes in the long-term interest rate had an impact on stock market returns.
Hypothesis 2:
An increase in long-term interest rates will lead to a decrease in stock market returns.
Monetary policy variables influencing stock market returns
The third channel that is expected to influence returns on stock markets is the monetary policy channel. The primary objective of monetary policy pursued by large central banks, such as the FED and the ECB, is achieving price stability. Economic growth and employment are only subordinate targets. Central banks react to economic fluctuations in the business cycle by adjusting their monetary policy instruments, which in turn affects inflation. To be precise, central banks adjust their key interest rates at which banks can borrow and deposit money from or at the central bank.
9 Hypothesis 3a:
An increase in short-term interest rates corresponds to lower stock market returns.
Similar, though expected at a different magnitude
Hypothesis 3b:
An increase in the central bank’s key interest rate causes a decline in stock market returns.
Apart from conventional monetary policy, central banks can employ unconventional monetary policy to increase money supply. That is, quantitative easing is a policy measure that central banks use when conventional policy no longer has the desired effects in reaching their inflation and/or economic growth targets. Some scholars argue that increased supply of credit and too low interest rates, merely increases investments in non-productive goods, thereby inflating asset prices (e.g. Murphy 2011; Block 2012)
Money supply can be measured by monetary aggregates: M1, M2 and M31. In a detailed investigation on the three measures of money supply, Asprem (1987) finds that there is an insignificant relationship between narrow money supply (M1) and the stock market. The broader measures of money supply have a positive and significant relation with stock returns. This result is of particular interest in the setting of this topic. More interestingly there exists a liquidity effect, where the increase in money supply, thus liquidity, causes an increased demand for financial assets (e.g. Asprem 1987; Dhakal et al. 1993; Mookerjee and Yu 1998; and Humpe and Macmillan 2009).
Hypothesis 3c:
An increase in money supply, M1 or M3, leads to increasing stock market returns.
Common Monetary Policy in EMU
In his famous work, Mundell (1961) describes the characteristics of an Optimal Currency Area (OCA) and addresses the question: “on which kinds of economic criteria could the
decision by various regions of the world to adopt a common currency be based?” (Swoboda 1999). Mundell (1961) argued that benefits of an OCA include increased liquidity of the
currency from which financial markets benefit and a reduction in the transaction costs. On the
1 Monetary aggregates: narrow (M1) includes the physical money in circulation and overnight deposits. M2
10 other hand, costs are that a country can no longer use independent monetary policy to counter economic shocks.
The criteria that Mundell (1961) identifies as important are for instance: absence of asymmetric shocks among potential members, and factor mobility. For this research one criteria of the OCA identified by Mongelli (2008) is of particular interest, which is financial
market integration. Converging member countries economies means co-movement of outputs,
earnings and equity markets (e.g. Canova and DeNicolo 1995). Financial market integration allows portfolio risk diversification and lets economic agents benefit from improved allocative efficiency of capital (Draghi 2014). Financial integration makes it easier for countries short of funds to obtain funds from countries with a surplus of funds. Furthermore, financial integration ensures that funds flow to their most productive use and reduces differences in long-term interest rates (Mongelli 2008). On the other hand, financial integration could have destabilising effects such as risk-taking and contagion (Draghi 2014).
Though this paper does not address design failures and its consequences (see for example De Grauwe 2013), the characteristics of the EMU and its weaknesses are deemed to be important to understand the effects of Central Bank policy on European stock markets. Inherent to capitalism are its boom and bust dynamics of the business cycle (de Grauwe 2013). For members to benefit from an OCA, these booms and busts dynamics need to be synchronized across its member states (Frankel and Rose 1998). Or put it differently, symmetry across countries in the business cycle is required in order for monetary policy to be effective.
11 becomes problematic. As a result, synchronization of business cycles in a currency area is of great importance. Nonetheless, Rose and Engel (2002) argue that causality also runs in the other direction: “members of a common union should experience more synchronized business
cycles since they do not experience national monetary policy shocks”(Rose and Engel 2002,
p: 17). Although theory commonly agrees on the necessity of synchronized business cycles, not a lot of attention is paid to the financial cycle even though financial integration is a condition for an OCA (Mundell 1973; Rose and Engel 2002; Mongelli 2005).
A particular interesting paper, related to this study, is Capiello et al. (2006). They find co-movements among stock markets and government bond markets have increased tremendously after the introduction of the EMU. The effect appears to be larger for countries with larger economies and more important stock markets. Given the fact that most studies have found some sort of relationship between macroeconomic fundamentals and stock markets, one would expect that this co-movement of stock markets reflects a symmetric movement of the business cycles across the Eurozone. Furthermore, the high degree of financial integration and capital mobility across the EMU would make money and monetary policy an influential factor across EMU-members. Therefore, one can expect that the effects of monetary policy on financial assets and markets will be transmitted more or less symmetrical across EMU-member countries.
In an optimal currency area, one would expect a symmetric movement of the business cycle across its member countries, which would be expected to be reflected in a co-movement in the stock market returns of the individual member countries. Therefore, in an optimal currency area which has highly integrated financial markets, one would expect a symmetric response to changing monetary policy across EMU stock markets.
Hypothesis 4a:
An increase in short-term interest rate causes a symmetric decline in EMU stock markets.
Similarly,
Hypothesis 4b:
An increase in ECB’s key interest rate causes a symmetric decline in EMU stock markets.
Hypothesis 4c:
12 By providing answers to the above described hypotheses, it should be possible to provide an answer to the main research questions of this paper. First, “Are stock market returns in line
with macroeconomic fundamentals, or is it influenced by monetary policy?” Second: “Is there an equal response of EMU stock markets to changing financial conditions and monetary policy?”
III.
Methodology
To investigate the research questions, this study employs a variety of econometric techniques and analyses. Standard statistical inferences are used to exam if this empirical setting satisfies conditions of Classical Linear Regression Model. These include for instance tests for heteroskedasticity, serial autocorrelation and normality. An initial analysis has been conducted in the form of a panel data regression. For the basic model, the following regression equation is estimated:
RTRIit = β0 + γi + β1IPit + β2INFit + εit (1)
Where i stands for the country investigated and t stands for the time period and ε represents the error term that is assumed to follow a normal, stationary distribution. To control for any non-normal properties of the error term regressions are executed with heteroskedasticity and serial correlation robust standard errors. Furthermore, RTRI is the return on the Total Return Index, IP is growth in industrial production and INF is inflation and γi stands for country
effects to control for heterogeneity across EMU-members.
The basic model will be extended to include variables that belong to the financial channel. These are the short-term and long-term interest rates, and leads to the following equation:
RTRIit = β0 + γi + β1IPit + β2INFit + β3STRit + β4LTRit + εit (2)
The monetary channel will include measures of growth in money supply, M1 and M3 and the marginal lending rate of the ECB, ECB. The final regression equation is presented as:
RTRIit = β0 + γi + β1IPit + β2INFit + β3STRit + β4LTRit + β5M1it + β6M3it + β7ECBit + εit (3)
To provide a more in-depth analysis at the country level a “Seemingly unrelated regressions
13 individual linear regressions equations, a SUR runs several linear regression equations, independently at the same time. In addition, it allows the error term to be correlated among independent regressions, leading to more efficient estimates than under OLS. Therefore, there is no necessity of adjusted standard errors. Estimating effects of selected dependent variables on Total Return Indices (TRI) for each individual country would require 12 independent OLS analyses. Besides, it is possible that effects of one market spill over to other markets. As mentioned above, SUR analysis is therefore a well-suited technique to perform the analysis. The regression equation of the SUR is presented in equation (2) below:
yit = Xitβit + εit i = 1,…,m, t = 1,…,n. (4)
y (MN × 1) = ( 𝑦1 𝑦2 … 𝑦𝑀 ), (MN × 1) = (ε 𝜀1 𝜀2 … 𝜀𝑀 ), (K × 1)β = ( 𝛽1 𝛽2 … 𝛽𝑀 ) With X (MN × 1) = ( X1 0 … 0 0 X2 … … … … … 0 0 … XM ) And K ≡ ∑ Kj M j=1
where yit is a MN× 1 vector indicating the returns on TRIs of EMU members. The matrix Xit
is a MN*Kit matrix and consist in the basic model of growth in industrial production, IPit; inflation, INFit. Again this is expanded to include growth rate in narrow and broad money
supply, M1it and M3it; ECB’s key interest rate, ECBit; long-term interest rate, LTRit; and
short-term interest rate, STRit. Subscript i represents the equation number, subscript t the time
period. βit is a Kit*1 vector representing the coefficients and the error term εit is a MN*1 vector
that the error term follows a white noise distribution.
IV.
Data and descriptive statistics
14 The data
Monthly data on selected variables spanning the months January 2001 to June are used. The specific data sources are located in table X in the appendix. The motivation for this time setting is that this study is focussed on the EMU that has been introduced in 1999. From January 2001, and onwards, all founding countries were part of the EMU and brought the Euro into physical circulation in January 2002. These countries are: Austria, Belgium, Finland, France, Germany, Greece, Italy, Ireland, the Netherlands, Luxembourg, Portugal and Spain. Hence, the 12 countries fell under single monetary policy of the European Central Bank. Limiting the starting point to January 2001 excludes different monetary policies of individual countries prior to the enactment of the EMU.
This study uses the following variables that are justified by literature. The dependent variable is the returns on countries’ main stock market indices. Instead of using the regular Price Index, this study includes the Total Return Index (TRI). The TRI not only captures regular changes stock market indices but also includes paid-out dividends. Henceforth, provides a more precise measurement for returns. Belgium does not have a TRI while Greece and Portugal have a TRI but are missing observations for the first couple of months (Portugal) or years (Greece). Nonetheless, given the high correlation between the PI and TRI it should not lead to any strong deviations. Table XI in the appendix presents the main stock markets of the 12 countries and their corresponding types of indices. The returns for 12 stock markets indices are calculated as the monthly returns. To be more specific, monthly index returns are the percentage change between closing positions at the first of each month, for each individual stock market index.
This study includes several explanatory variables. In line with literature, the first independent variable is the monthly growth rate of industrial production. This is a widely used proxy for current economic activity and is measured as the percentage change compared to the previous period, at the NACE Rev.2 classification of economic activity. To be precise, it includes NACE Rev2. Codes2: B to E. The data is both seasonally and working days adjusted.
The second independent variable is inflation and is represented as monthly percentage change in the “Harmonised Index of Consumer Prices” or HICP-100. It is a measurement of inflation
2 The NACE Rev.2 classification is a statistical classification of Economic Activity in the European Community.
15 and price stability developed by the ECB. HCIP-100 is a weighted average of price changes for a basket of consumer goods and services for the individual countries (ECB 2015).
Three types of interest rates are analysed. The long-term interest rate is presented as the 10-year government bond. The long-term interest rate is typically used as the discount factor in valuation. The short-term interest rate is defined as the “Day-to-Day money market interest rates” on money markets. The key policy interest rate of the ECB is the marginal lending rate. That is the rate at which credit institutions can obtain overnight liquidity from the central bank, against sufficient eligible collateral (ECB 2015)
Fifth, to study the impact of changes in money supply on stock markets, monthly growth rate (%) of two types of money aggregates, narrow (M1) and broad (M3) are included.
Descriptive statistics
The descriptive statistics of selected variables are presented in table I below. The descriptive statistics at the country level are presented in the Appendix in table XII. From the aggregate descriptive statistics over the 12 Eurozone member countries presented in table 3, several observations can be made.
First, mean monthly return of the total return index is 0.32%, and swings between a minimum of -33.15% and a maximum of 30.51%. Besides wide swings between the minimum and maximum monthly stock returns, there is quite some variation in returns as measured by the standard deviation, which is 6.56%. Henceforth, stock market returns tend to be quite volatile. Similar patterns are found for other variables such as growth in industrial production. These patterns could be driven by outliers in the data and hence estimations based on the data can become inaccurate. Nonetheless, as these are just a few observations in a large sample, outliers are not excluded as they are part of the population.
Table I: Descriptive statistics on selected series at the aggregate level.
Variable Mean Std. Min Max
Stock market returns (%) 0.323 6.558 -33.150 30.514 Industrial production growth (%) 0.041 2.837 -23.800 26.200
Inflation (%) 0.167 0.683 -2.500 3.400
16
Figure 1: Stock market returns Germany versus 3 outlier countries.
17 Figure 1 above presents stock market returns for Germany and three countries that are the most volatile of all twelve EMU-countries. By taking a detailed view at the figure, it appears there that there is one country that stands out with inconceivable volatile returns: Greece. Greece has faced severe economic conditions over the past years and therefore the result is not very surprising. In addition, it seems that up and till 2008 stock market returns tend to move quite tranquil and symmetrical, suggesting a convergence of financial market movements. The financial crisis in 2008 tends to have caused more volatile stock market returns. The subsequent start of the European Debt Crisis tends to have caused a divergence between the stock market returns of EMU-members.
Several additional observations can be made when the sample is split into a period before and after the start of the European Debt Crisis at the end of 2009/beginning of 2010 (table XII). The cut-off is December 2009 which is the date where the euro crisis started as the Greek prime-minister Papandreou announced that Greece’s Debt-to-GDP ratio amounted 113% (BBC 2015).
There seem to be few differences on the aggregate level, before and after December 2009. The monthly minimum and maximum returns on the Total Return Index after December 2009 are more or less the same compared prior December 2009. Though, the monthly mean returns are higher after December 0.53% compared to 0.19% prior December 2009. In addition, the mean monthly growth of industrial production is higher after December 2009, 0.09% vs 0.01% pre-December. The variation in the monthly growth of industrial production is far greater after the crisis erupted. Another prominent difference can be found for different measurements of the growth rate in the money supply. For both M1 and M3 a lower mean growth and variation is observed. Similarly, short-term interest rates have a mean monthly interest rate of 0.31% after December vs 2.89% pre-December 2009.
18 Preliminary tests
It is of importance to check whether assumptions of Classical Linear Regression Model hold. These included a check on normal distribution of residuals, heteroskedasticity, serial autocorrelation and multicollinearity.
First, presence of multicollinearity between independent variables has been investigated. A correlation matrix is presented in the appendix in table XIII. Presence of multicollinearity will not violate the OLS assumptions. Results are still unbiased, but no longer efficient as standard errors tend to be inflated. Further, multicollinearity reduces statistical power of the analysis and can cause coefficients to change signs. Hence, estimates can become inconsistent. Any form of high correlation has a value greater than 0.6. From the correlation matrix it becomes clear that there are no severe issues of multicollinearity. With the exception of a very high correlation of 0.977 between ECB’s marginal lending rate and short-term interest rate. As an additional test for multicollinearity, variance inflation factor test have been computed after each regression.
Second, it has been examined if residuals follow a normal distribution among their mean value. The residual plot generates a normal distribution and thus does not violate the fifth CLMS assumption.
Third, for the panel data analysis several additional tests have been conducted. A Hausman test has been conducted to determine whether a fixed effect or random effect model is appropriate. To be precise, it tests the null hypothesis that the unique errors are uncorrelated with its regressors. The Hausman test rejects the null hypothesis, hence random effects model is the appropriate model. Succeeding, Breusch-Pagan Lagrange multiplier (LM) test has been conducted. This test helped to decide between random effects panel regression and simple pooled OLS regression. The Breusch-Pagan LM tests the null hypothesis that variances across entities is zero (i.e. no panel effect). The test results suggested that simple OLS is appropriate, as the null hypothesis cannot be rejected. Hence, pooled OLS regression is applied to the panel data.
19 Fifth, Lagram-Multiplier test for serial correlation has been conducted to study the presence of serial correlation of the error terms. The null-hypothesis of no first-order autocorrelation has been rejected at the 1% significance level. Hence, there is presence of first-order autocorrelation.
Finally, a heteroskedasticity test has been conducted. In presence of heteroskedasticity, standard errors are no longer efficient. This study employs financial data, hence it is likely that standard errors of regression results will be inflated by heteroskedasticity. A White’s test has been applied and found significant evidence of heteroskedasticity. The null hypothesis has been rejected at the 1% significance level. Concluding, the pooled OLS regression has been performed with cluster-robust Huber/White standard errors at the country level.
V.
Results
Panel data results
Table II below presents the first estimation results from the panel data analyses. The analysis starts with the basic panel data regression model in which the macro-economic channel is included. Thus, growth in industrial production and inflation. The second model includes the financial channel, with its short-term and long-term interest rate. Model three includes the monetary channel with growth in money supply M1 and M3 and without the ECB’s key interest rate. ECB’s key interest rate has been excluded in this model as there is a high degree of multicollinearity between short-term interest rate and ECB’s key interest rate. The fourth model includes the monetary channel with growth in money supply M1, ECB’s key interest rate but excludes the short-term interest rate.
First of all, across all models insignificant3 results are found for the macro-economic channel. The results of the second model are in line with literature. The short-term interest rate is highly significant at 1% level. The coefficient obtained is -0.762, implying a 1% increase in the short-term interest rate decreases the returns on the stock market by 0.762%. This effect becomes even stronger in the third model where the monetary channel is included. The coefficient is now -0.878. The economic explanation for this is quite intuitive. Lower interest rates makes credit cheaper and could thus increase investment and economic activity or inflate stock markets/asset markets. For long-term interest rates negative and significant results are obtained. Again, this is in line with literature, as long-term interest rates are used as discount
20 factors in valuation. This implies: higher long-term interest rates reduces the time value of future discounted dividends, and thus current stock prices. For the monetary channel, significant results are obtained. An increase in narrow money supply has a significant and negative effect on stock market returns. Growth in broad money on the other hand has a positive and significant impact on stock returns. Here a 1% increase in supply of broad money (M3) increases stock returns by 0.676%. This thus confirms Asprem (1987) finding of the liquidity effect that increasing supply of credit indeed inflates stock market returns. However, its significance of this result fades away when ECB’s key interest rate is included. Finally, for ECB’s key interest rate a highly significant and negative value of 0.693% has been found. This implies that a 1% increase in the ECB’s key interest rate leads to 0.693% lower stock market returns.
Table II: Base-line panel data results, for the effects of economic, financial and monetary channel on stock market returns.
Variables (1) (2) (3) (4)
Industrial production growth (%) 0.055 0.035 0.032 0.044 (0.053) (0.049) (0.049) (0.050) Inflation (%) -0.215 -0.132 -0.144 -0.134 (0.159) (0.177) (0.184) (0.186) Short-term interest rate (%) -0.762*** -0.878***
(0.052) (0.062)
Long-term interest rate (%) -0.182*** -0.182*** -0.176*** (0.040) (0.042) (0.035)
ECB’s key interest rate (%) -0.693***
(0.063) Narrow money supply (%) -0.360** -0.297**
(0.137) (0.132)
Broad money supply (%) 0.676*** 0.210
(0.156) (0.162) Constant 0.933*** 3.007*** 3.161*** 3.614***
(0.031) (0.183) (0.235) (0.230)
Observations 2,088 2,088 2,088 2,088
R-squared 0.004 0.042 0.044 0.034
Notes: Robust standard errors in parentheses; *, ** and *** represents the level of significance at the 10% , 5% and 1% level respectively. Column (1) shows the results from estimating equation (1). Column (2) is based on the second equation (2). Columns (3) & (4) are the results from estimating equation (3), where either ECB’s key interest rate is excluded, column (3), or short-term interest rate is excluded, column (4).
21 First of all, across all models growth in industrial production and inflation remains insignificant in explaining stock market returns. The significant results for short-term interest have only increased in sizes during the crisis. Apparently, effects of movements in short-term interest rates have become more pronounced during the crisis than before. Estimation results for term interest rate also provide different results. In line with literature, effects of long-term interest rates on stock market returns are negative and significant. However, during the crisis effects of long-term interest rate turn positive. This is not in line as it suggests that an increase in long-term interest rates (i.e. discount factor) increase stock market returns (e.g. Nasseh and Strauss 2000).
The results of monetary policy on stock market returns have also changed. That is, impact of ECB’s key interest rate on stock market returns is significant prior and during the crisis. Further, effects of increases in key interest rate become more negative during the crisis period than prior to crisis. The impact of narrow money on stock market returns remains negative and significant across all models. Growth in broad money positively influences stock market pre-crisis, though becomes insignificant after the crisis in the third model. In the fourth model where ECB’s key interest rate is included, effects of growth in broad money supply on stock market returns turns negative and significant during the crisis. That implies that an increase in broad money growth causes a decrease in stock market returns. This is contrary to Asprem (1987)’s liquidity effect finding.
22
Table III: Panel data results, for the effects of economic, financial and monetary channel on stock market returns.
(1) (2) (3) (4) (5) (6) (7) (8)
Variables Pre-crisis Crisis Pre-crisis Crisis Pre-crisis Crisis Pre-crisis Crisis Industrial production growth (%) 0.145 -0.099 0.095 -0.108 0.065 -0.108 0.104 -0.101 (0.083) (0.081) (0.065) (0.082) (0.063) (0.074) (0.070) (0.072) Inflation (%) -0.179 -0.282 -0.022 -0.187 -0.016 -0.356 0.028 -0.251 (0.231) (0.167) (0.258) (0.153) (0.254) (0.202) (0.268) (0.185) Short-term interest rate (%) -1.664*** -5.191*** -2.429*** -6.171***
(0.105) (0.631) (0.095) (1.034)
Long-term interest rate (%) -1.169*** 0.301*** -0.767*** 0.260*** -1.267*** 0.209** (0.173) (0.038) (0.169) (0.047) (0.211) (0.095)
ECB’s key interest rate (%) -2.019*** -2.218***
(0.109) (0.498)
Narrow money supply (%) -1.144*** -1.602** -1.056*** -0.790*
(0.161) (0.555) (0.155) (0.407)
Broad money supply (%) 2.385*** -0.439 1.800*** -1.529***
(0.257) (0.287) (0.282) (0.418) Constant 1.262*** 0.430*** 11.010*** 1.347*** 10.90*** 2.687*** 13.990*** 3.530***
(0.039) (0.043) (0.630) (0.209) (0.765) (0.609) (0.814) (0.758)
Observations 1,284 804 1,284 804 1,284 804 1,284 804
R-squared 0.007 0.021 0.117 0.081 0.136 0.099 0.110 0.059
23 SUR analysis results
To distinguish between individual effects at the country level, four SUR analyses have been conducted. Results of the first specification of the macroeconomic effects on stock market returns can be found in table IV. The first SUR results are not very informative. Almost none of the variables are significant at the 5% or 10% level. Only for Portugal has Industrial production a negative and significant effect on its stock market returns. For inflation, a positive effect on Finish stock market returns is found while a negative and significant effect on Greece’s stock market returns is obtained.
Next, the empirical analysis proceeds by incorporating the financial variables into the analysis. The results of the extended SUR-model are in table V. The effects of Industrial production and inflation do not significantly change. Additionally, for France a positive and significant value for inflation is obtained. For short-term interest rate, negative and significant effects on stock markets are obtained for 11 of the 12 EMU-members. Only for Greece appears the effect of short-term interest rate on stock market returns to be insignificant. Furthermore, the magnitudes of an increase or decrease of short-term interest rates on countries’ stock markets is more or less similar across EMU-members. This provides some first evidence for a synchronized financial cycle in the EMU. The results show that stock markets returns are, except for Austria, not influenced by changes in the long-term interest rate. This is contrary to the panel results and literature that suggest that they should be affected, as long-term interest rates are used as discount factors.
24
Table IV: SUR analysis macroeconomic channel
Variables Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Luxembourg Portugal Spain
25
Table V: SUR analysis macroeconomic and financial channel
VARIABLES Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Luxembourg Portugal Spain
Industrial production growth (%) 0.015 0.090 -0.219 -0.113 -0.004 -0.002 -0.012 -0.039 -0.009 0.064 -0.176* -0.145 (-0.161) (-0.078) (-0.184) (-0.085) (-0.121) (-0.151) (-0.045) (-0.125) (-0.057) (-0.075) (-0.101) (-0.159) Inflation (%) 0.884 0.004 1.990* 0.626* -0.098 -0.766* 1.017 0.137 -0.214 0.251 -0.420 0.030
(-0.722) (-0.183) (-1.041) (-0.354) (-0.440) (-0.394) (-0.674) (-0.237) (-0.288) (-0.355) (-0.529) (-0.318) Short-term interest rate (%) -0.988** -0.894*** -0.919 -0.758** -0.666* -0.466 -1.221*** -0.872*** -0.833** -0.880** -0.656** -0.553* (-0.416) (-0.292) (-0.570) (-0.304) (-0.404) (-0.518) (-0.298) (-0.300) (-0.364) (-0.369) (-0.268) (-0.294) Long-term interest rate (%) 0.727* 0.038 -0.082 -0.182 -0.282 0.031 0.065 -0.133 -0.191 0.043 -0.113 -0.157
(-0.435) (-0.293) (-0.631) (-0.275) (-0.366) (-0.105) (-0.164) (-0.241) (-0.340) (-0.338) (-0.122) (-0.259) Constant 0.088 1.783* 2.110 2.303** 2.697*** 0.133 2.337** 2.328* 2.574** 2.005** 1.976** 2.332*
(-1.264) (-0.983) (-1.621) (-0.905) (-0.992) (-1.661) (-1.056) (-1.198) (-1.009) (-0.974) (-0.978) (-1.261)
Observations 174 174 174 174 174 174 174 174 174 174 174 174
R-squared 0.035 0.075 0.031 0.049 0.040 0.009 0.081 0.053 0.062 0.053 0.039 0.017
26
Table VI: SUR analysis macroeconomic, financial and monetary channel, excluding ECB’s key interest rate
Variables Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Luxembourg Portugal Spain
Industrial production growth (%) 0.021 0.082 -0.231 -0.121 -0.048 0.010 -0.007 -0.037 -0.008 0.063 -0.175* -0.162 (-0.160) (-0.079) (-0.188) (-0.085) (-0.121) (-0.151) (-0.044) (-0.125) (-0.057) (-0.075) (-0.101) (-0.157) Inflation (%) 0.986 0.017 1.999* 0.635* -0.145 -0.830** 1.347** 0.126 -0.168 0.245 -0.402 0.092
(-0.723) (-0.185) (-1.045) (-0.351) (-0.439) (-0.395) (-0.674) (-0.238) (-0.289) (-0.361) (-0.530) (-0.318) Short-term interest rate (%) -0.994* -1.033*** -1.166 -0.951** -1.053** -0.448 -1.455*** -0.987*** -0.848* -0.945** -0.828** -0.840**
(-0.523) (-0.368) (-0.718) (-0.378) (-0.490) (-0.639) (-0.371) (-0.377) (-0.452) (-0.468) (-0.332) (-0.368) Long-term interest rate (%) 0.701 0.059 -0.004 -0.148 -0.146 0.022 0.187 -0.150 -0.274 0.008 -0.110 -0.038
(-0.465) (-0.311) (-0.681) (-0.288) (-0.378) (-0.107) (-0.168) (-0.250) (-0.361) (-0.365) (-0.126) (-0.267) Narrow money supply (%) 0.317 -0.230 -0.711 -0.463 -0.894 -0.808 0.683 -0.496 -0.203 -0.166 -0.369 -0.268
(-0.757) (-0.594) (-0.908) (-0.622) (-0.733) (-1.167) (-0.715) (-0.718) (-0.693) (-0.714) (-0.642) (-0.707) Broad money supply (%) 0.124 0.744 1.134 1.001 1.663 -0.153 1.205 0.704 0.397 0.505 1.003 1.544
(-1.394) (-1.086) (-1.697) (-1.131) (-1.333) (-2.085) (-1.280) (-1.292) (-1.265) (-1.315) (-1.138) (-1.269) Constant -0.087 1.789 2.268 2.401** 2.850*** 0.779 1.175 2.632** 2.831** 2.123* 2.075* 1.842
(-1.397) (-1.095) (-1.787) (-1.001) (-1.087) (-1.860) (-1.191) (-1.322) (-1.116) (-1.109) (-1.106) (-1.389)
Observations 174 174 174 174 174 174 174 174 174 174 174 174
R-squared 0.036 0.077 0.036 0.054 0.052 0.013 0.091 0.056 0.061 0.054 0.043 0.023
27
Table VII: SUR analysis macroeconomic, financial and monetary channel, excluding short-term interest rates
Variables Austria Belgium Finland France Germany Greece Ireland Italy Netherlands Luxembourg Portugal Spain
Industrial production growth (%) 0.035 0.086 -0.225 -0.110 -0.031 0.009 -0.008 -0.033 -0.005 0.066 -0.175* -0.160 (-0.159) (-0.079) (-0.185) (-0.086) (-0.119) (-0.150) (-0.044) (-0.124) (-0.057) (-0.075) (-0.101) (-0.157) Inflation (%) 0.963 0.018 2.025* 0.638* -0.117 -0.857** 1.355** 0.116 -0.190 0.243 -0.464 0.089
(-0.723) (-0.185) (-1.053) (-0.358) (-0.441) (-0.392) (-0.670) (-0.238) (-0.290) (-0.363) (-0.531) (-0.319) Long-term interest rate (%) 0.739 0.087 0.074 -0.212 -0.225 0.012 0.243 -0.132 -0.337 -0.001 -0.090 0.001
(-0.523) (-0.348) (-0.744) (-0.315) (-0.396) (-0.102) (-0.167) (-0.255) (-0.384) (-0.396) (-0.123) (-0.276) ECB’s key interest rate (%) -0.859 -0.900** -1.073 -0.742* -0.815 -0.208 -1.329*** -0.820** -0.644 -0.786 -0.590* -0.722* (-0.565) (-0.391) (-0.768) (-0.393) (-0.503) (-0.622) (-0.374) (-0.380) (-0.469) (-0.491) (-0.330) (-0.371) Narrow money supply (%) 0.352 -0.198 -0.694 -0.413 -0.835 -0.697 0.705 -0.449 -0.151 -0.130 -0.267 -0.234
(-0.770) (-0.603) (-0.926) (-0.632) (-0.746) (-1.179) (-0.726) (-0.727) (-0.706) (-0.723) (-0.653) (-0.714) Broad money supply (%) -0.286 0.342 0.738 0.535 1.164 -0.706 0.819 0.261 -0.037 0.095 0.463 1.215
(-1.381) (-1.076) (-1.669) (-1.115) (-1.312) (-2.044) (-1.271) (-1.274) (-1.247) (-1.297) (-1.127) (-1.251) Constant 0.502 2.445** 3.011* 3.111*** 3.603*** 0.774 2.096 3.187** 3.433*** 2.758** 2.267* 2.262
(-1.402) (-1.139) (-1.676) (-1.085) (-1.184) (-2.242) (-1.340) (-1.444) (-1.177) (-1.183) (-1.267) (-1.500)
Observations 174 174 174 174 174 174 174 174 174 174 174 174
R-squared 0.026 0.061 0.027 0.040 0.040 0.011 0.074 0.044 0.051 0.045 0.024 0.014
In the final model in table VII, ECB’s key interest rate is included. Although it has the expected outcome on stock market returns, its effects are only significant for 6 EMU-member countries. This suggest that even though short-term interest rates are directly influenced by the ECB, the effects of key interest rate on EMU stock market returns is of less importance than effects of short-term interest rates.
As a final analysis, the effects of the financial crisis have been investigated. The procedure is similar to the one applied in the panel data analysis. The regression outputs of these analyses are presented in the appendix (tables XIV – XVII).
The results in table XIV show that before the crisis a positive and significant (at 10% level) effect of industrial production on stock market returns is found for Austria , while a negative effect for Portugal is obtained. During the crisis period growth in industrial production has a negative and statistically significant effect on Austria’s stock market returns, while the effect is positive for Belgium. Inflation is only significant and positive pre-crisis for Austria, Finland while negative for Greece (pre-crisis).
In the second set of specifications (table XV), the previous result for Austria, Finland, Belgium and Portugal still hold for the effect of growth industrial production on stock market returns. Only, the coefficients pre-crisis tend to get smaller and during crisis they slightly increase. Inflation is still positive and significant for Austria, while negative and significant for Greece. For the financial channel, interesting results are obtained. Similar to results in table V, effects of short-term interest rate on stock market returns is negative and highly significant across all the EMU-members. However, effects become more negative during the crisis compared to pre-crisis. Their magnitudes tend to be more or less similar, with one exception which is not really surprising: Greece. The effects of short-term interest rates on Greece’s stock market returns are much more severe compared to other countries, probably due to its miserable economic situation. For long-term interest rates, positive and similar effects on stock market returns are obtained during the crisis for Austria and Finland.
29 stock markets are obtained. The effects are solely significant before the crisis and disappears during the crisis (with the exception of Greece) For growth in broad money supply, positive and significant effects on stock market returns are found, but only for four countries, and again only before the crisis. These findings, though small are quite in line with literature e.g. Asprem 1987).
Finally, including ECB’s key interest rate (table XVII) slightly reduces the coefficients that are significant compared to previous models in table XVI. Furthermore, and more interesting, the effects of ECB’s key interest rate on stock market returns is not symmetric across EMU-member’ stock markets. That is, for countries such as Austria, Belgium, Finland, France, Germany, Ireland, the Netherlands, effects either becomes smaller (less negative) or its significance disappears altogether. For Greece, Italy and Spain effects of ECB’s key interest rate became more severe during the crisis. Similar to effects of short-term interest rates across EMU-members’ stock markets, the impact of ECB’s key interest rate tends to vary more during the crisis than before. Before the crisis the coefficients were more or less similar, suggesting a symmetric response of financial markets across EMU stock markets. During the crisis, the response tends to become asymmetric.
For the first hypothesis the results of the SUR are inconclusive. The effects of growth in industrial production on stock market returns appear to be relevant for just a few countries. Equal to Laopadis (2011), evidence is too thin to draw strong conclusions at the country level and therefore it is impossible to accept or reject the hypothesis. For the effects of level of inflation, again thin evidence has been found. Hence, at the country level it is not possible to form clear conclusions. Therefore hypothesis 1b, is only rejected for both pre-crisis and during the crisis for some countries, except for Greece.
Long-term interest rates appear to have a positive effect on three EMU-member stock market returns but only during the crisis. This rejects the hypothesis and conflicts with most literature. Furthermore these results are consistent with the panel data results. However, where the panel data analysis found negative and significant impacts of an increase in long-term interest rates on stock market returns pre-crisis, the SUR could not. Hence, the results at the country level aggregation are too meagre to reject or accept the hypothesis.
30 markets (hypothesis 4a), before the crisis the effects appear to be symmetric and thus in line with the hypothesis. Suggesting that before the outbreak of the Euro crisis in 2009, the financial cycle moved synchronized. Conversely, the results suggest that the start of the Euro crisis has lead to an asymmetric response of EMU stock market returns on increasing short-term interest rates. Similarly, although to a lesser extent, the impact on EMU stock market of increasing ECB’s key interest rates tends to be symmetric before the crisis and asymmetric before crisis. Therefore, hypothesis 4a and 4b are accepted for the period before the crisis and rejected for the crisis period. Hence, it seems that the financial cycle is no longer moving synchronized across the Eurozone and therefore financial integration is reversed.
Finally, without controlling for the crisis insignificant results were found for the impact of growth in narrow and broad money on stock market returns. When controlling for the crisis, significant results were obtained for a few countries. Nevertheless, the results are too sparsely to draw strong conclusions that an increase in the money supply leads to a symmetric increase in EMU stock markets. Hence, it is not possible to accept or reject hypothesis 4c.
To conclude, it appears that both effects of short-term interest rates and ECB’s key interest rate are the only explanatory variables that tend to be significant across most EMU-members while other explanatory variables are not. Subsequently, effects of ECB’s key interest rate and short-term interest rates on stock market returns were quite symmetrical before the crisis. The crisis period appears to have caused an asymmetric response by stock markets on changes in the financial and monetary channels. Hence, before the crisis the financial cycle tended to move symmetrical, suggesting high levels of financial integration. The crisis seems to have contributed to asymmetric movements of the financial cycle, hence causes financial de-integration.
VI.
Robustness
31
Table VIII: Panel data results with one lag
Variables (1) (2) (3) (4)
Industrial production growth (%) 0.042 0.022 0.028 0.035 (0.061) (0.057) (0.059) (0.061) Inflation (%) 0.225 0.310* 0.313* 0.330* (0.165) (0.153) (0.153) (0.151) Short-term interest rate (%) -0.788*** -0.598***
(0.052) (0.075)
Long-term interest rate (%) -0.151*** -0.157*** -0.145*** (0.038) (0.045) (0.044)
ECB’s key interest rate (%) -0.534***
(0.071)
Narrow money supply (%) 0.448** 0.458**
(0.154) (0.157) Broad money supply (%) -1.097*** -1.293***
(0.261) (0.243) Constant 0.833*** 2.859*** 2.715*** 3.140***
(0.031) (0.187) (0.266) (0.316)
Observations 2,076 2,076 2,076 2,076
R-squared 0.004 0.043 0.046 0.044
Notes: Robust standard errors in parentheses; *, ** and *** represents the level of significance at the 10% , 5% and 1% level respectively. Column (1) shows the results from estimating equation (1). Column (2) is based on the second equation (2). Columns (3) & (4) are the results from estimating equation (3), where either ECB’s key interest rate is excluded, column (3), or short-term interest rate is excluded, column (4).
32 As there appear to be some changes in the results, the panel data is again extended. That is, a crisis dummy has been included and results are presented in table IX. Compared to table III, the growth in industrial production has a positive effect on stock market returns during crisis. However, its effect appears to fade away. For inflation, significant effects have been found across almost all models. In addition, the effects of inflation on stock market returns tend to change before and during the crisis. Before crisis, inflation has a positive effect on stock market returns, whereas during the crisis the effect tends to become negative which is in line with literature (Fama 1981; Asprem 1987). The effects of short-term and long-term interest rates and ECB’s marginal lending rate on stock market returns results remain more or less similar to those presented in table III. Only their values decrease.
Again, narrow and broad money growth changes sign. By including a lag variable, the effects of narrow money growth on stock market returns during the crisis becomes positive instead of negative. For growth in broad money supply, before crisis insignificant effects on stock market returns are found while during the crisis negative and significant effects are found.
Thus the robustness check confirms that stock market returns are severely influenced by movements in short-term interest rates and not so much by macroeconomic fundamentals. In addition, the effects increase during the crisis. Furthermore, effects of monetary policy, in terms of growth in money supply, turn out to be positive for narrow money growth during the crisis. Contrary to growth in broad money that seems to have an opposite effect. Nevertheless, reversed causality for the measures of growth in money supply might be the problem as the signs change compared to the non-lagged results. Hence, no hard conclusions can be drawn for the impact of growth in money supply on stock market returns.
Table IX: Panel data results with one lag and crisis dummy
(1) (2) (3) (4) (5) (6) (7) (8)
Variables Pre-crisis Crisis Pre-crisis Crisis Pre-crisis Crisis Pre-crisis Crisis Industrial production growth (%) 0.011 0.071* -0.043 0.063* -0.046 0.051 -0.019 0.054 (0.096) (0.035) (0.079) (0.034) (0.081) (0.032) (0.087) (0.032) Inflation (%) 0.958*** -0.516*** 1.129*** -0.428** 1.053*** -0.199 1.133*** -0.146 (0.275) (0.159) (0.308) (0.175) (0.309) (0.167) (0.311) (0.175) Short-term interest rate (%) -1.814*** -4.812*** -1.966*** -3.028***
(0.175) (0.681) (0.266) (0.576)
Long-term interest rate (%) -1.220*** 0.338*** -1.289*** 0.364*** -1.388*** 0.344*** (0.281) (0.046) (0.370) (0.048) (0.354) (0.081)
ECB’s key interest rate (%) -1.993*** -1.124***
(0.266) (0.343)
Narrow money growth (%) -0.869*** 3.617*** -0.935*** 4.004***
(0.075) (0.467) (0.078) (0.488)
Broad money growth (%) 0.283 -3.912*** 0.293 -4.457***
(0.386) (0.773) (0.386) (0.869) Constant 1.062*** 0.431*** 11.510*** 1.164*** 12.720*** -0.671*** 15.080*** -0.214 (0.043) (0.032) (0.704) (0.136) (1.022) (0.186) (0.729) (0.232)
Observations 1,272 804 1,272 804 1,272 804 1,272 804
R-squared 0.010 0.023 0.135 0.074 0.144 0.130 0.139 0.121
VII.
Conclusion
This paper has attempted to investigate the relationship between stock market returns, monetary policy and financial integration in the EMU. Using a variety of econometric techniques, it was possible to provide an answer to the following two research questions: “Are
stock market returns in line with macroeconomic fundamentals, or are they severely influenced by monetary policy?” and second: “Is there an equal response of EMU stock markets to changing financial conditions and monetary policy?”.
The results from the econometric analysis do not provide enough evidence to reject or accept all stated hypotheses. From both panel data analysis and SUR analysis generally insignificant results are obtained that current growth in industrial production has a positive relation with stock market returns These results are for instance in line with Lee(1995) and Chung and Lee (1998) who were also not able to find any significant evidence that growth in industrial production has an impact on stock prices. Long-term interest rates appear to be significant in the panel data. In addition, the effects of increases in long-term interest rates on stock market returns tend to be negative before the crisis, which is in line with literature. However, at the country level results are inconclusive. Effects of growth in money supply seem to be struck by causality problems. The robustness check employed with one lag changed the signs of the influence of growth in money supply on stock market returns. Therefore it is not possible to draw strong conclusions on effects of the money supply on stock market returns in the EMU. Finally, movements in short-term and ECB’s key interest rate do show significant and negative relationships with EMU stock markets. In fact, their impact on stock market returns increase during the crisis. This suggests that there is a liquidity effect, in which lower interest rates cause higher valuations of financial assets. These results confirm the findings of
35 Concluding, stock market returns in the EMU are severely influenced by monetary policy, i.e. interest rate changes, and are sparsely affected by changes in macro-economic fundamentals. Furthermore, the start of the euro crisis appears to have caused asymmetric responses of EMU stock markets to changing financial and monetary conditions. This reduces financial market integration in the EMU and consequently has a negative impact on monetary transmission mechanism.
Unfortunately this paper has its limitations. For instance, the SUR analysis was not able to find significant results for effects of money supply on stock market returns whereas the panel data was able to do so. Either, countries are not influenced, or the time period was too short to obtain significant results. Henceforth, future research might include longer time-periods. In addition, the robustness check revealed that there is an issue of causality for effects of growth in money supply on stock market returns. Hence, to be able to draw strong conclusion on this channel other estimation techniques might be applied in the future to rule out reversed
causality. Further, this study employs monthly data. The danger is that monthly data might be noisy, hence estimations results can become inaccurate. Therefore I suggest future research to use quarterly data, to see whether stronger evidence can be found. The drawback is that when one employs quarterly data, the amount of observations significantly decreases and therefore it becomes difficult to make clear and robust estimations.
The results are important for policy makers as it confirms that stock market returns in the EMU are severely influenced by movements in short-term interest rates and ECB’s key interest rate. Furthermore, effects are more prolonged during the crisis. Given that macro-economic fundamentals appear to have an insignificant impact on stock market returns, strong effects of movements in interest rates inflate stock market returns, creating an asset bubble. To prevent the creation of bubbles by monetary policy, explicitly including the financial sector into economic models should make monetary policy more effective (Bezemer, 2010)
36 an OCA Thus restoring the symmetric financial cycles is of importance for the robustness of the EMU.
VIII.
References
Authers J. 2015. Fears Grow over US Stock Market Bubble. Financial Times; [accessed 2015 November 15].
http://www.ft.com/cms/s/0/85d0becc-58c5-11e5-a28b-50226830d644.html#axzz3um3aJSjY.
Abugri BA. 2008. Empirical relationship between macroeconomic volatility and stock returns: Evidence from Latin American markets. International Review of Financial Analysis. 17:396-410.
Asprem M. 1989. Stock Prices Asset Portfolios and Macroeconomic Variables in Ten European Countries. Journal of Banking and Finance. 13:589-612.
Bezemer DJ. Understanding Financial Crisis Through Accounting Models. Accounting, Organizations and Society. 35:676-688.
BBC. 2015. Timeline: The unfolding Eurozone crisis. BBC; [accessed 2015 December 3]. http://www.bbc.com/news/business-13856580.
Binswanger M. 2001. Does the stock market still lead real activity? An Investigation for the G7-Countries. Journal of Financial Market and Portfolio Management. 15:15-29.
Binswanger M. 2004. How Important are Fundamentals? – Evidence from a Structural VAR model for the Stock markets in the US Japan and Europe. Journal of International Financial Market Institutions and Money. 14:185-2001.
Block W. 2012. An Austrian Critique of Mainstream Economics. Mises Media; [accessed 2015 December 8]. https://www.youtube.com/watch?v=TI5fjTz1Rbw.
Campbell JY, Shiller RJ. 1988. Stock Prices Earnings and Expected Dividends. The Journal of Finance. 43(3):661-676.
Canova F, DeNicolo G. 1995. Stock Returns and Real Activity: a Structural Approach. European Economic Review. 39:981-1015.
Cappiello L, Hordahl P, Kadareja A, Manganelli S. 2006. The Impact of the Euro on Financial Markets. ECB. Working paper 598.
Carlson J, Sargent K. 1997. The Recent Ascent of Stock Prices: Can It Be Explained by Earnings Growth or Other Fundamentals?. Federal Reserve Bank of Cleveland Economic Review. 33(2):2-12.
37 Chung H, Lee B-S. 1998. Fundamental and non-fundamental Components in Stock Prices of Pacific-Rim Countries. Pacific-Basin Finance Journal. 6:321-346.
Clinch M. 2015. Soros: ECB QE Means Inequality and Asset Bubbles. CNBC; [accessed 2015 September 20]. http://www.cnbc.com/2015/01/23/soros-ecb-qe-means-inequality-and-asset-bubbles.html.
CNBC. 2015. How to Read the FED Minutes: Schiff. CNBC; [accessed 2015 December 20]. http://video.cnbc.com/gallery/?video=3000456224.
Dhakal D, Kandil M, Sharma SC. (1993). Causality Between the Money Supply and Share Prices: A VAR Investigation. Quarterly Journal of Business Economics. 32(3):52-74. Draghi M. 2014. Speech by Mario Draghi President of the ECB at 20th anniversary of the Establishment of the European Monetary Institute Conference Brussels. 12 February 2014. ECB. 2008. The Role of Banks in the Monetary Policy Transmission Mechanism. ECB Monthly Bulletin. 2008 (8):85-98.
ECB. 2013. Stock Market Developments in the Light of the Current Low-Yield Environment. ECB Monthly Bulletin. 2008 (8):44-48.
ECB. 2015. Measuring Inflation- the Harmonized Index of Consumer Prices (HICP). ECB; [accessed 2015 September 20].
https://www.ecb.europa.eu/stats/prices/hicp/html/index.en.html.
ECB. 2015. Monetary Policy Instruments: Standing facilities. ECB; [accessed 2015 December 2]. https://www.ecb.europa.eu/mopo/implement/sf/html/index.en.html.
ECB 2015. Monetary Aggregates. ECB; [accessed 2015 October 12]. https://www.ecb.europa.eu/stats/money/aggregates/aggr/html/index.en.html.
Egan M. 2015. Trump Warns of a Stock Market Bubble: is he right? CNBC; [accessed 2015 October 30]. http://money.cnn.com/2015/10/15/investing/donald-trump-stock-market-bubble/. Fama EF. 1981. Stock Returns Expected Return and Real Activity. The Journal of Finance. 45(4):1089-1108.
Fama EF, French KR. 1988. Permanent and Temporary Components of Stock Prices. Journal of Political Economy. 96 (2):246-273.
Frankel JA. and Rose AK. 1998. The Endogeneity of the Optimum Currency Area Criteria. Economic Journal. 108:1009-1025.
Gali J. Gambetti L. 2014. The Effects of Monetary Policy on Stock Market Bubbles: Some Evidence. National Bureau of Economic Research. wp: 19981.
