Note: Summarize the article in your own wording with a professional format.

MACROECONOMIC VARIABLES AND STOCK MARKETS: AN

INTERNATIONAL STUDY

Francisco JAREO

Ana ESCRIBANO

Alberto CUENCA

Abstract. This paper studies the potential correlation between the stock market of six relevant countries (Germany, Italy, Spain, France, UK and US) and some important macroeconomic factors, such as the gross domestic product (GDP), the consumer price index (CPI), the industrial production index (IPI) and the unemployment (UNEMP). GDP and UNEMP show statistically significant correlation with these international stock markets, mainly in the crisis sub-period, finding, in addition, the expected signs.

Keywords: International Stock Market; Macroeconomic Factors; Correlation Analysis; US; European Countries

1. Introduction and literature review.

A large part of the financial literature agrees that the globalization process begins to develop at the beginning of the 21st century, fundamentally the globalization of financial systems, which is the focus of this work. Financial globalization occurs mainly due to the liberalization of national financial systems, which causes a greater connection between international financial systems. Thus, this would be one of the main reasons for the rapid and general spread of the global financial crisis of 2008 that affected the world economy.

The equity markets experienced a generalized growth during the beginning of the century, showing the economic moment of growth that extends, approximately, until the year 2007. At the end of this year, a recession begins in the United States that mainly affects to the stock markets. Financial globalization, therefore, is what makes the US recession begin to move to markets throughout Europe and the rest of the world at the beginning of 2008. This year there has been a generalized fall in yields in the international equity markets, which has continued for several consecutive quarters, reaching even 2009, as shown by the data on the evolution of stock prices in that period.

According to Chen et al. (1986), Humpe and Macmillan (2009), and Jareo and Negrut (2016), among others, the aim is to analyze the possible relationship between international stock market returns and a pool of relevant macro-economic variables, largely gathered from the previous studies. Because of the recent sample period, this research may observe whether changes in the economic cycle -before, during and after the recent global financial crisis- affect in some way the relationship studied between the macro variables and the returns of different international stock markets.

Many researches investigate the relationship between stock markets and macroeconomic factors, although they do not find agreement in their conclusions. However, according to Chen et al. (1986), Wasserfallen (1989), Schwert (1990), Peir (1996 and 2016), Humpe and Macmillan (2009) and Jareo and Negrut (2016), the expected signs of the most relevant macroeconomic variables could be those collected in Table 1.

Table 1. Relationship between stock markets and macroeconomic factors: expected signs

GDP

CPI

IPI

Unemployment

Stock Market

Positive

Uncertain

Positive

Negative

Source: Own preparation based on Chen et al. (1986), Wasserfallen (1989), Schwert (1990), Peir (1996 and 2016), Humpe and Macmillan (2009) and Jareo and Negrut (2016)

Thus, the study includes the most used macroeconomic factors in the previous literature: the Consumer Price Index (CPI), the Industrial Price Index (IPI), the Gross Domestic Product (GDP) and unemployment (UNEMP) during a sample period between 2000 and 2014. The impact of these variables on some international stock market indices is analyzed, in concrete, for Germany, Spain, France, Italy, UK and US. In addition, the analysis of the relationship between the selected macro-economic variables and different stock market returns is carried out in a period that includes the recent global financial crisis, because this paper aims to study if this relationship changes according to the phase of the economic cycle, focusing attention on the global financial crisis phase.

The rest of the paper is structured as follows. Section 2 shows the data sample analysed in this paper. Section 3 analyses the time evolution between the stock market and the different macroeconomic variables. Section 4 shows correlation matrices between the stock market price and the various macroeconomic factors. Finally, Section 5 shows the main conclusions of this study.

2.Data

This paper examines the impact of some relevant macroeconomic variables (CPI, GDP, IPI and UNEMP) on international stock market returns (Germany, Spain, France, Italy, UK and USA) from 2000 q1 to 2014 q4.^[1] As previously said, this study breaks the whole sample period into three different sub- periods: pre-crisis (2000-2006), crisis (2007-2010) and post-crisis (2011-2014).

Specifically, we use quarterly data for the 2000-2014 sample period. Furthermore, data on the selected macroeconomic variables were obtained from the Eurostat website

(http://ec.europa.eu/eurostat) and the National Bureau of Economic Research (http://www.nber.org/). Data from differente international stock markets were obtained from the Econstats (http://www.econstats.com/).

For comparison reasons, this research studies six different international stock markets, such as Germany, Spain, France, Italy, UK and USA. Thus, we analise the following stock market indices: DAX30 (Germany), IBEX35 (Spain), CAC40 (France), MIB30 (Italy), FTSE100 (UK) and S&P500 (US). International market indices have been incorporated into the analysis through the yields of the quarterly closing quotations.

Finally, the explanatory variables have been incorporated into the analysis as growth rates, which guarantees that the variables included in the analysis are stationary variables. Thus, the four macroeconomic variables used in this research are defined as follows: (1) the

Gross Domestic Product (GDP) represents the value of all goods and services produced in the United States; in concrete, this study uses two different measures of GDP: GDP in real terms, and the growth rate in percentage; in addition, these measures are seasonally adjusted; (2) the Consumer Price Index (CPI) is the original data used to obtain the US inflation rate; in particular, we have considered this factor as an index and as an inflation rate; (3) the Industrial Production Index (IPI) according to the National Statistics Institute is a cyclical indicator that measures the productive activity of the industrial sector (excluding construction); this factor has been considered in the analysis as an index (in levels), and the growth rate (seasonally adjusted); finally, (4) the unemployment (UNEMP) represents the total number of individuals who are not working but are actively seeking employment.

Table 2. Market indices and macroeconomic variables

International Stock Market Indices

Macro Variables

DAX30 (Germany)

Consumer Price Index (CPI)

IBEX35 (Spain)

Industrial Production Index (IPI)

CAC40 (France)

Gross Domestic Product (GDP)

MIB30 (Italy)

Unemployment

S&P500 (US)

FTSE100 (UK)

According to Table 1, based on Jareo and Negrut (2016), among others, a positive relationship between the stock market and both GDP and IPI may be expected. Thus, higher prices in the stock market are associated with higher values for both variables (GDP and IPI), and their behavior proceeds according to the stock market cycle: good news in the financial economy also means good news in the real economy and vice versa. By contrast, the unemployment and interest rates are negatively related to the stock market; that is, higher prices on the DJ index are associated with lower values for these macroeconomic factors, showing anti-cyclical behavior. Again, good news in the financial economy produces good news in the real economy (because these factors, in principle, are better when the values are lower). Moreover, the relationship between the inflation rate and the stock market is uncertain because it can fluctuate according to the needs of the economy.

3.Analysis of the time evolution between the stock market and the different macroeconomic variables

This section collects the graphs that show the evolution of the explanatory variables and the respective stock market index of each country. Figure 1 shows CPI, IPI, and GDP, since previous literature hypothesizes a positive relationship with market returns. Figure 2 exhibits the relationship between unemployment and international stock market indices, assuming an inverse relationship.^[2] The explanatory variables collected in Figure 1 and 2 are shown in levels, although next section includes growth rates. In addition, these graphs show a shaded area that refers to the global financial crisis period in 2008. In particular, shaded areas in Figure 1 indicate recession periods based on the NBER dating.^[3] Thus, the beginning of the aforementioned economic recession was in 2008q1, with the end of it being dated in 2009q2. This concrete period is the one that has been highlighted in the following graphs with a shaded area.

Figure 1. Graphs of the combined evolution of the stock market and macroeconomic factors (CPI,

IPI, GDP): Germany, Spain and France, Italy, USA, UK

Note: Shaded areas in this figure indicate recession periods based on the NBER dating.

According to Figure 1, which shows the historical evolution of the different market indices and the macro-economic variables (in levels), in Germany -which serves as an example of the evolution of data in Europe-, it is observed how until the year 2002 certain variables suffer decreases and changes of tendency to, from that moment, begin a joint evolution that leads to the beginning of the crisis. The shaded period, which refers to the global financial crisis, reflects the decreasing trend and changes in the evolution of the variables. The rest of Figure 1 verifies the increasing tendency of the magnitudes in almost all its route, with the exception of certain periods of decrease.

The rest of the countries show a similar evolution until the global financial crisis. As of 2010, according to the economic policies developed and their different impact, for instance, in France and Spain it is found that some indicators do not recover the growth trend after the crisis, keeping their variables horizontal (stability) or even decreasing. Thus, a similar evolution in the GDP is observed, since it follows the evolution of the market index, or even in the IPI, which does not reach its previous growth rate.

Figure 2. Graphs of the combined evolution of the stock market and macroeconomic factors

(UNEMP)

Note: Shaded areas in this figure indicate recession periods based on the NBER dating.

In Spain, at least until 2010, a correspondence is observed in the evolution of all the variables, since the Consumer Price Index continues to increase along with the IPI until the final stretch, where the latter remains constant. In its case, GDP continues to maintain an evolution similar to the market index (although softened), which decreases in the final tranche. The CPI maintains a rhythm of growth that only stops during the recession, reaching a decline in this period. The IPI does maintain a behavior similar to the market index for much of the total period, because until the crisis maintain a common growth. After this, it continues with its previous growth rate, which is only slowed down in the last years analyzed, where it remains constant, moving into a growth phase, as well as the stock market returns (between 2012 and 2014).

Figure 2 contains the temporal evolution of international market indices and unemployment (in levels). In concrete, in France, in most of the period the relationship is inverse, especially in the shaded part, during the crisis, in which the decrease in market prices is accompanied by an increase in unemployment. Previously, the increase in market returns is accompanied by a large decline in unemployment. Therefore, unemployment is expected to have an inverse relationship with the market index, as shown in Table 1. Graphically, Figure 2 seems to confirm this relationship during a large part of the period analyzed. This relationship is more clearly reflected in the case of the US, although it is also observed in the European countries, because in the pre-crisis period we find falling unemployment rates and increasing market returns. Once the period of recession begins, the market index falls sharply and unemployment begins to have a constant growth rate that places it several percentage points above its previous data.

In sum, a visual inspection of the graphs that reflect the temporal evolution of the variables analyzed by country allows us to anticipate the potential existence of a relationship between the different macroeconomic variables and the stock market returns in the different countries. Thus, there may be a direct relationship between three out of four explanatory variables (CPI, IPI and GDP) and the international stock market indices, and inverse in one of the cases (unemployment). However, we find certain quarters in which this relationship is diffuse or does not come into existence. Therefore, we confirm that Unemployment seems to show an inverse relationship with the stock market indices during a large part of the sample period. There is also a direct relationship with GDP, since both variables show a similar evolution during the whole sample period. The CPI, however, has an almost continuous trend of growth, which makes it go away in times of recession in the market index, showing a direct relationship at times of economic growth. As for the IPI, to a lesser extent than the GDP, it also exhibits some direct relationship with the equity market, decreasing at times when the market indices show a negative trend, especially in the crisis period (2008q1 - 2009q2).

4.Relationship between international stock market returns and some macroeconomic factors

For robustness, to study the existence of a relationship between the explanatory variables included and the stock market returns of different countries, we check our preliminary results through a correlation analysis and scatter plots.

For this second analysis, the variables are expressed in growth rates (one quarter compared to the previous quarter) to guarantee the stationarity of the explained and explanatory variables. The analysis is carried out by countries with their respective variables for the entire period (2000-2014) in a first matrix. Later we will proceed to show the matrices for each sub-period mentioned previously: pre-crisis (2000-2006), crisis (2007-2010), and post-crisis (2011-2014). The complete sample period is divided into sub-periods to check if the relationship is greater at certain times and eliminate distortions that may have occurred in certain periods of time, thus affecting the whole sample.

4.1. Scatter plots to show the relationship between international stock market returns and some macroeconomic factors

Figure 3, in the Annex, exhibits the relationship between the international stock market returns and the explanatory variables of each country, collected with scatter plots. Again, a positive relation between stock market returns and the GDP (growth rate), IPI (growth rate) and the inflation rate is observed, and, on the other hand, an inverse relationship between the international market returns and the unemployment rate. This last relation is the clearest, since the cloud of points perfectly shows the inverse relationship that exists with unemployment, since at a lower rate of unemployment the returns are, in general, higher. In addition, the slope of the represented regression line seems to show a greater slope than the rest.

The GDP growth rate, on the other hand, also shows a revealing relationship in the US scatter plot, although in this case with a positive trend line. This result would indicate that at times of higher stock market returns, the GDP growth rate would also be expected to be higher. Furthermore, in US this direct relationship show a higher slope in the case of GDP and IPI. The US inflation rate show a weaker relationship in the whole sample period, because the trend line is almost horizontal. Finally, the unemployment rate shows a clearly inverse relationship with a negative trend line and with a steep slope.

Thus, in cases where the relationship is positive, the GDP growth rate is the one that reflects a relationship with a steeper slope. Therefore, the higher growth rates of this variable correspond to higher stock market returns, although it is true that some distortion of the results is observed at specific moments in the sample period analyzed.

UK shows a reality very similar to what happens in the rest of Europe. However, in this case the inflation rate (CPI growth rate) seems to show a slightly higher correlation with stock market returns. In addition, the unemployment rate in the United Kingdom would again show a markedly steep slope as it happens in the rest of the European countries analyzed, because even though the point cloud is presented in a dispersed way, the trend line indicates the negative relationship between the unemployment rate and stock market returns. As in the rest of the countries, the unemployment rate exhibits a clearer relationship with stock market returns. This overview will contrast with the correlation matrices presented later. As in the previous case, the GDP growth rate shows a clear positive relationship with respect to stock market returns, generally observed in the rest of the countries analyzed. The same happens with the growth rate of the IPI, which although sometimes with a less pronounced trend line, shows scatter plots for all countries with a clear positive trend.

Finally, the relationship found between international stock market returns and the growth rate of the CPI (inflation rate) and the IPI (growth rate of Industrial Production Index) is positive but slightly lower than the rest. In both cases, there is a positive trend line but little pronounced. This would show some positive relationship in this case but with a very scattered cloud of points that makes the relationship observed in the graphics somewhat less clear.

4.2. Correlation matrices between international stock market returns and some macroeconomic factors

To confirm the relationships observed in the previous scatter plots, and in order to improve the perception of the possible relationship between the stock market returns and the growth rates of the different macro-economic variables, we show some correlation matrices by countries that express numerically these relationships.

This analysis shows four correlation matrices since, as previously said, the analysis has splitted the whole sample period into three different sub-periods. First, the total sample period (2000-2014) is analyzed, since it is the period used in the previous section of the dispersion diagrams. In this way, you can confirm the previous results with those obtained through the correlation matrices.

Thus, tables 3-6 present the correlation matrix to indicate whether the relationships between the six international stock market indices and the analyzed macroeconomic variables are statistically significant. To that end, Student's t distribution and the associated probabilities (p-value) are used to confirm the statistically significance.

4.2.1. Correlation matrix: the whole sample period (2000-2014)

Table 3 shows the Pearson correlation coefficients between macro variables and international stock market returns (from Germany, Spain, France, Italy, UK and US) during the whole sample period. When the coefficient shows a positive sign, it means that the relationship is direct, as seen in the GDP growth rate, the inflation rate, and the growth rate of the IPI. In the case of the unemployment growth rate, the coefficient and, therefore, the relationship is negative. Thus, at higher values in the market returns, lower values are found in terms of the unemployment rate.

Table 3. Correlation matrix between stock market returns and some macroeconomic factors: the whole sample period (2000-2014)

GDP

CPI

IPI

UNEMP

Germany

0.3210 *

0.2947 **

0.1081

-0.6793 **

Spain

0.5150 **

0.0832

0.3069 *

-0.3298

France

0.4216 *

0.1468

0.3878

-0.6794 **

Italy

0.5752 **

0.2395

0.5129 **

-0.4329 *

US

0.2726 *

0.1568

0.1869

-0.4443 **

UK

0.1654

0.2268 *

0.2590 *

-0.3252 *

Note: As usual, *, **, *** indicate statistical significance at the 10%, 5% and 1% levels, respectively.

In this way, it is observed that the unemployment rate is the macro-economic variable with the highest correlation (in absolute values) and the greatest statistical significance. This variable is statistically significant at 5% level in half of the countries, and at 10% level in two countries. In a paradoxical manner, only in Spain UNEMP is not statistically significant. In addition, the negative coefficients are around 0.5, which indicates a linear and inverse relationship between both variables, with an average correlation. This value confirms the results obtained in the scatter plots, with the unemployment rate being the variable that exhibits the greatest relationship with respect to stock market returns. Finally, if we compare the results by countries, France and Germany show higher coefficients, with a correlation close to 68% in the case of the unemployment rate.

On the other hand, the GDP growth rate shows a positive and statistically significant relationship with the international stock market returns in most countries, corroborating some previous results observed in the scatter plots. This relationship is statistically significant in most countries, except in the UK. In addition, in the rest of the countries there are coefficients that, as in the case of the unemployment growth rate, hover around 0.5 with a positive sign (although the coefficients, in absolute value, are slightly lower). Furthermore, the countries that show a higher correlation coefficient are, in this order, Italy (0.58) and Spain (0.52).

The growth rate of the IPI shows an insignificant relationshipt in the case of certain countries, situation that was illustrated in the previous scatter plots. The same case is observed in the inflation rate (growth rate of the CPI). Both variables show a relationship that is only statistically significant in the case of certain countries. This may be due to different behaviors depending on the phase of the economic cycle, which are compensated in the entire period, showing an inconclusive result.

4.2.2. Correlation matrix: pre-crisis, crisis and post-crisis sub-periods

A second analysis of correlation matrices linked to the different sub-periods of precrisis, global financial crisis and post-crisis could provide complementary information, mainly in the case of growth rates of the IPI and the inflation rate, with inconclusive results in the full sample period.

Table 4. Correlation matrix between stock market returns and some macroeconomic factors: the pre-crisis sub-period (2000-2006)

GDP

CPI

IPI

UNEMP

Germany

0.5201 **

0.1538

0.2821

-0.7829 ***

Spain

0.5147 **

0.0104

0.2935

-0.3196

France

0.3739 *

-0.0964

0.3739 *

-0.6546 **

Italy

0.5060 **

-0.0092

0.3021

-0.4220 **

US

0.2312

0.0289

-0.0475

-0.2309

UK

-0.2978

0.2013

0.0825

0.0257

Note: As usual, *, **, *** indicate statistical significance at the 10%, 5% and 1% levels, respectively.

Table 4 that shows the results in the pre-crisis sub-period and shows a situation similar to that collected in the analysis of the whole sample period. In this case, however, the existence of statistically significant correlations between the macro-economic variables and the stock market returns is lower. Only statistically significant correlations are observed for the growth rate of GDP and the unemployment rate in a part of the countries analyzed. The unemployment rate in the United Kingdom shows a Pearson correlation coefficient that is virtually zero. The other two variables (IPI growth rate and inflation rate) show little correlation for this period, except in the case of France for the IPI. The rest of correlations show values very close to zero, which implies the non-existence of a linear relationship. Negative coefficients are also found for these variables, since the time series shows that, during the pre-crisis period (2000-2002), the market indices are falling while the explanatory variables (IPI and CPI) exhibit a positive trend. In the sub-period of the global financial crisis, a greater correlation between macroeconomic variables and the respective international stock market returns is observed. Table 5 shows higher Pearson correlation coefficients than in the previous samples, since all the variables find a statistically significant relationship depending on the country. Specifically, the growth rate of the GDP is the variable with the greatest relationship with respect to the stock market returns -with the exception of the UK-.

Table 5. Correlation matrix between stock market returns and some macroeconomic factors: the crisis sub-period (2007-2010)

GDP

CPI

IPI

UNEMP

Germany

0.6049 *

0.7306 **

0.5278 *

-0.3604

Spain

0.7370 **

0.3204

0.4446

-0.5532 *

France

0.5347 *

0.4634 *

0.5651

-0.7875 **

Italy

0.8214 **

0.3187

0.6463 *

-0.3335

US

0.6348 *

0.5240 **

0.5763 **

-0.5785 *

UK

0.3233

0.1081

0.4552

-0.7074 **

Note: As usual, *, **, *** indicate statistical significance at the 10%, 5% and 1% levels, respectively.

The unemployment rate, on the other hand, also shows a relationship with stock market returns in four out of six countries analyzed. This inverse relationship implies that in this stage of crisis, while the stock market returns experience a downward trend, the unemployment rate begins to increase -and vice versa-. The coefficients are close to the value 0.5, which would imply an average level of correlation, although in France and the UK they have a correlation level of 70%.

The inflation rate and the growth rate of the IPI in this case show a stronger Pearson correlation coefficient than in the correlation matrices presented previously. In the crisis subperiod they show a higher level of probability and coefficients with higher values. However, we find some countries in which these variables are not statistically significant. Again, these macro-economic variables exhibit a lower correlation with the stock market returns. In short, the financial crisis sub-period seems to show a greater correlation between the explanatory variables and the international market returns, because this period corresponds to a recession that affects the economy as a whole. This translates into falls and loss of the positive trend of almost all of the variables for the different countries analyzed. Finally, the unemployment rate changes its tendency to decrease in 2007 due to a continuous increase in unemployment. When dealing with an economic crisis that affects most variables, they begin a cycle of depression that translates into linear movements of the macro-magnitudes in a framework of crisis.

Finally, Table 6 shows the Pearson correlation coefficients of the post-crisis period, in which Europe and the United States may show quite different situations. Potentially diverse economic policies could have carried out depending on the geographical zones, which may affect the respective countries differently. This could explain why the variables act differently depending on the country analyzed. Thus, the correlation between macro variables and stock market returns would be affected by this context, oscillating each one differently and even erratically.

Table 6. Correlation matrix between stock market returns and some macroeconomic factors: the post-crisis sub-period (2011-2014)

GDP

CPI

IPI

UNEMP

Germany

0.0502

-0.1971

-0.3888

0.1880

Spain

0.4339

0.0337

0.4887

-0.4292

France

0.3488

0.0210

0.1067

-0.5877 *

Italy

0.2416

-0.3205

-0.2715

-0.3131

US

0.3412

-0.1342

0.0858

-0.5306 *

Note: As usual, *, **, *** indicate statistical significance at the 10%, 5% and 1% levels, respectively.

Thus, the only macro-economic variable that shows a statistically significant Pearson correlation coefficient in the post-crisis sub-period is the unemployment rate, and only in France and the UK, but with relatively low coefficients with respect to samples previously analyzed. One can get to appreciate a positive correlation between UNEMP and German stock market returns, situation that is not expected by the inverse behavior that assumes of both. In the rest of the countries, no statistically significant correlation is observed.

With regard to the correlation between the growth rate of the GDP and the stock market returns, wich shows statistically significance in previous analyses, it does exhibit an insignificant correlation for all countries in the post-crisis sub-period. The other two macro variables (IPI growth rate and inflation rate) show lower Pearson correlation coefficients than other magnitudes, since there is no statistically significant correlation for the countries studied. The coefficients extracted from the correlation matrix in Table 6 show Pearson correlation coefficients close to zero in most cases.

Therefore, in the post-crisis sub-period there is no clear correlations between macroeconomic variables and stock market returns. This situation may be due to the different economic policies adopted in each country to face the situation after a period of economic recession. The instruments used and the policies implemented may affect the variables and markets in different ways depending on the situation of each country, which would cause a multitude of trends in the different variables and countries, distorting or masking the correlation that, a priori, is expected. In summary, as expected, positive Pearson correlation coefficients are found between stock marke returns and the macro variables IPI growth rate, inflation rate and GDP growth rate. On the contrary, a negative and statistically significant correlation is observed between market returns and the unemployment rate. In addition, these last two varaibles (GDP and unemployment) are the most intense, with more pronounced trend lines. Furthermore, the analysis by sub-periods shows how these correlations are more relevant in the crisis sub-period, but reflect minimum values (both in terms of correlation coefficients and statistical significance) in the post-crisis stage.Moreover, the Pearson correlation coefficients between stock market returns and the growth rate of GDP and the unemployment rate show the highest values and significance levels. On the other hand, the inflation rate and the growth rate of the IPI do not show a clear correlation with stock market returns. Thus, they show scatter plots with a horizontal trend line in many cases. Second, the correlation matrices show Pearson correlation coefficients very close to zero and statistically insignificant.

5.Summary and concluding remarks

The aim of this research is to analyze if the expected relationships between a set of relevant macro-economic variables (Consumer Price Index: CPI, Industrial Production Index: IPI, Gross Domestic Product: GDP and Unemployment: UNEMP) and six international stock markets are verified: DAX30 (Germany), IBEX35 (Spain), CAC40 (France), MIB30 (Italy), FTSE100 (United Kingdom) and S & P500 (United States).

A priori, the relationship between stock market returns and the CPI variable would be uncertain, positive for GDP and IPI, and negative for UNEMP. The results obtained in the previous studies reviewed vary in many cases depending on the different periods analyzed, the economic cycle, the sample size, or even the methodology used.

The time evolution of the macro-economic variables and the benchmark indices of the stock markets analyzed show that the UNEMP and GDP variables are those that, apparently, show a fairly clear relationship with the market performance, the first inversely and the second directly. The other two variables, CPI and IPI, show a seemingly less intense and clear relationship.

The correlation analysis between the macro variables and the international stock market indices shows greater and statistically significant correlations during the crisis subperiod. In addition, this analysis corroborates that the UNEMP variable shows an intense and inverse correlation with the market performance, and the GDP variable a strong but positive correlation, confirming that these variables seem to be the most correlated with the stock market returns. On the other hand, the IPI and IPC variables show a lower correlation, which is only remarkable at certain moments of time and with less statistical significance.

The Pearson correlation coefficients by country show statistical significance in the crisis period for all the macroeconomic variables at least in some of the countries analyzed, and in all countries there is at least one statistically significant correlation. In the pre-crisis period, there is a smaller number of statistically significant correlations (only for GDP and UNEMP) for countries such as Germany, France, Italy and Spain. After the crisis, the variables obtain less correlation with the market, possibly due to the different effects of the recession. Only statistically significant correlations are observed in the post-crisis period for UNEMP in countries such as the UK and France.

In general, UNEMP and GDP are the magnitudes that show a clearer result, since in most of the tests carried out there is a statistically significant relationship with the stock market returns. In addition, this relationship appears with a positive sign for GDP, which indicates movements in the same direction in stock market returns and GDP. In the case of UNEMP, the sign is negative, that is, there would be movements in the opposite direction in the stock market returns and UNEMP, as expected. Both correlations are stronger in times of crisis and in countries where policies accompany these joint movements.

On the other hand, the other two macro-magnitudes, CPI and IPI, show, in general, a lower correlation with stock market returns, which is only remarkable at certain moments in time for any specific country. So the evolution of these two macro variables does not seem to be linked to that of the stock markets, but rather acts more independently.

These results may allow us to make certain predictions of the future movements that the international stock market returns may experience to changes in the macro-economic variables GDP and unemployment. Statistically significant correlations mean that increases in GDP and UNEMP decreases would allow improving stock market returns, while movements in the opposite direction indicated in the macro-economic variables would lead us to a deterioration of the international stock market returns in the countries analyzed (with the particularities of each of the countries included in the analysis).

[1] The sample period ends in 2009 in the case of Italy, due to a lower availability of data from this country for the most recent dates.

[2] They are included separately to show more clearly their different evolution, since while the first three have a positive relationship, the fourth is negative.

[3] This page offers information on the start and end dates of the different phases of the economic cycle.