• No results found

Extreme-day return as a measure of stock market volatility : comparative study developed vs. emerging capital markets of the world

N/A
N/A
Protected

Academic year: 2021

Share "Extreme-day return as a measure of stock market volatility : comparative study developed vs. emerging capital markets of the world"

Copied!
45
0
0

Loading.... (view fulltext now)

Full text

(1)

Extreme-day return as a measure of Stock Market Volatility

Comparative study developed vs. emerging capital markets of the world

Authors: Muashab Kabir and Naeem Ahmed Subject: Master Thesis in Business Administration 15 ECTS Program: Master of International Management Gotland University Spring semester 2010

(2)

Abstract

This paper uses a new measure of volatility based on extreme day return occurrences and examines the relative prevailing volatility among worldwide stock markets during 1997-2009. Using several global stock market indexes of countries categorized as an emerging and developed capital markets are utilized. Additionally this study investigates well known anomalies namely Monday effect and January effect. Further using correlation analysis of co movement and extent of integration highlights the opportunities for international diversification among those markets. Evidences during this time period suggest volatility is not the only phenomena of emerging capital markets. Emerging markets offer opportunities of higher returns during volatility. Cross correlation analysis depicts markets have become more integrated during this time frame; still opportunities for higher returns prevail through global portfolio diversification.

(3)

Acknowledgements

We are really thankful to our supervisor Adri de Ridder whose worthwhile and technical supervision gave us the direction to complete our Master Thesis. With his valuable knowledge and technical expertise, he helped us on each step while heading towards the final stage of our Master thesis. Further the kind and helpful comments from Mathias Cöster which furthermore helped us to make this piece of work an academic product.

At the same time we are also grateful to Gotland university administration for making us available the technical facilities and access to vast academic literature.

Visby, June 2010

(4)

Table of Contents

1 Introduction ... 1

1.1 Problem definition ... 3

1.2 Hypothesis Development ... 3

2 Literature review ... 5

3 Data & Methodology ... 7

3.1 Data ... 7

3.2 Methodology ... 9

3.2.1 Extreme day returns definition ... 9

3.2.2 Frequency of extreme day return ... 10

3.2.3 Predictive power ... 11

3.2.4 Co-movement among frequencies of extreme day returns and level of integration among the capital markets ... 12

4 Results ... 15

5 Conclusion and recommendations for future research ... 22

References………..23

(5)

1

1 Introduction

Stock market volatility is a well thought out topic and has vital importance for investor’s decision making, thus has considerable influence on investor behavior in the market, as it gives just around the corner about risk reward trade off premises. With technological advancements financial markets have become global and easily reached to whole world for investment. This led to interdependence among worldwide stock markets. Such interdependence has implications for international diversification and even more during worldwide volatility. Eventually volatility in international capital markets has also become the key issue for international investment strategy. In this paper we compare volatility and level of integration among a combination of developed and emerging capital markets located in Europe, Latin America, and Far East. Motivating factor behind this study is emergence of financial markets’ internationalization, which has become more central in academic world due to the fact that international diversification with combination of emerging capital markets is considered optimum investment strategy.

Several methods have been in use for measuring asset return volatility, but time after time standard deviation had been mostly used. Standard deviation is considered useful at the same time not reliable as a measure of risk. Longin (1996) while analyzing risk and extreme values of the US stock prices from 1885 to 1990 argued that tail analysis is superior to standard deviation. Standard deviation gives misleading results in the daily stock returns due to absence of normal distribution and non stationary trend with positive and negative autocorrelation overtime (Jones et al, 2004). On the same lines Wander and Vari (2003) put forward standard deviation contains some flaws as a risk measure but it gives some useful insights. Nelken (1997) while describing the stock prices series of Motorola also rejected standard deviation as a reliable measure of volatility, due to the fact that standard deviation is deviation from mean, which does not remain constant throughout the series. These limitations put question mark on standard deviation as a reliable measure of volatility (risk). Eventually extreme value analysis is visible in literature. Jondeau & Rockinger (2003) analyzed the extreme return tails for 20 mature and emerging markets of the world in investor perspective. Similarly Assaf (2008) used extreme observations and portrayed it better risk measure and management approach analyzing the equity markets of Middle East and North Africa (MENA) region.

(6)

2 An alternative measure of volatility, which has been in use by researchers taking into consideration, is the frequency of extreme-day returns (positive and negative) proposed by Jones et al. (2004). Volatility using this measure is simple and independent of statistical distribution, measuring volatility using extreme-day return can be categorized as an alternative measure of risk (Jones et al, 2004). Further argued this measure is comparatively sound as compared to standard deviation1 and classification of positive, negative extreme day returns becomes even more explanatory about risk and explanation of investor behavior and asset management industry.

We are adopting the methodology proposed by Jones et al. (2004) in this study. As this methodology is easier to understand, more widespread in nature and provides more informative insights for better investment decisions taking into consideration the risk more explicitly. Thus this will be supportive and give valuable inputs for international investment strategy that can be utilized by investment analysts, institutional and individual investors as well. The purpose of this study is to compare volatility among developed and emerging capital markets, secondly to point out the opportunities of higher returns during volatility, thirdly to determine the financial and volatility contagion among developed and emerging capital markets.

This study will contribute to literature in two ways first by applying an alternative of traditional volatility measures and evaluation for looking into different types of markets from all over the world for providing useful information about international diversification strategy. Secondly analysis of emerging markets through this technique as to date; no emerging market has been studied by this methodology.

Application of this measure is not at large till now in literature, use of this approach on the data of USA by analyzing frequency of extreme days and in case of Sweden Ridder & Djehiche (2007) are found but not as high-flying as other various traditional measures are visible in literature. To our best knowledge, we cannot find application of this volatility measure in European markets other than Sweden and markets in Asia Pacific, Latin America so far in literature.

1

Standard deviation does not give explicit explanation about volatility as dispersion from mean are squared so direction of dispersion positive or negative is overlooked.

(7)

3

1.1 Problem definition

Previous studies shed light extensively on volatility and internationalization of capital markets but how risk (volatility) to be measured comprehensively in international scenario has been exposed using traditional measures. We are trying to do the same with more comprehensively by utilizing the approach of extreme day return in this comparative study. More specifically in terms of extreme day return pattern in emerging and developed markets, what are the differences that can be highlighted? Further how those differences can be exploited to measure and control the risk in the presence of correlation among capital markets. That can lead to earn higher returns in emerging markets as compared to developed capital markets.

1.2 Hypothesis Development

There are socio economic factors which also contribute to volatility as Xing (2004) finds that education level of investors, industry concentration, relative size of the stock market and number of listed firms on stock exchange, these factors effect volatility, as a rule emerging and developed markets differ on these attributes. Usually it is view that emerging markets are more volatile, returns are more predictable and non normality in return series as argued by Harvey and Bekaert and Harvey (1997) analyzing the twenty emerging markets for the time period January 1976 to December 1992. Whereas Ridder & Djehiche (2007) for Sweden and Jones et al. (2004) on USA data using a frequency of extreme days also confirm the non normality in case of developed markets as well. This is agreed upon and proved that stock return series are not normally distributed irrespective of market type developed or emerging. As categorization of developed and emerging markets is on the basis of comparative characteristics of economies, financial markets and investors which determine the behavior of return series. So our first hypothesis in this regard is

H1: Return series in case of emerging markets show more non normality as compared to

developed ones so more daily extreme returns.

While referring to predictive power (variation in frequency explained by past frequency) of frequency of extreme days, there is evidence of some predictability in the frequency of extreme days in case of developed markets USA argued by Jones et al. (2004), and also shown by Ridder & Djehiche (2007) in case of Sweden. While Chang et al. (2004) studying developed economies

(8)

4 USA, Japan and emerging economies from Latin America and Asia Pacific rejected random walk hypothesis in emerging economies and argued about return predictability in emerging markets as compared to Developed one. Similarly many others documented the evidence about predictability like (Achour et al. 1999), (Barry et al., 2002). Thus Past studies give evidence about the more tendency of predictability of returns in case of emerging stock markets in comparison to developed ones. And also extreme returns pattern is related to market predictability Muchnik et al. (2009). So our second hypothesis in this study is as follow.

H2: Frequency of extreme day returns have more predictive power in emerging markets as

compared to developed markets.

The rest of the paper is organized as follow in section 2 we discuss literature review relevant to our research area and problem definition. Section 3 contains the data description and methodology, in section 4 we describe results and empirical findings. In section 5 we conclude and identify issues for future research.

(9)

5

2 Literature review

Literature on volatility is extensive, as volatility is the main consideration for individual, institutional and also international investors to make any investment decision. Du and Budescu (2007) while analyzing the investor conduct regarding volatility pointed out less confidence to high volatility and past volatility as a risk measure, so sound measures of volatility can provide insights to investors as volatility in the stock market quantitatively measures the risk which is key determinant in the assessment of cost of capital, investment and leverage decision (Chukwuogor & Freidan, 2007). In the modern era where world is called global village, investors have opportunities for portfolio diversification internationally. Main objective behind the international investment is the hope of earning higher returns as compared to investing in single or local market, In this regard emerging markets are considered attractive. There is huge capital inflow to emerging markets as an alternative to diversification and emerging markets by and large offer higher return with high associated risk (Chang et al., 2004). Further equity risk premium is more in emerging economies as compared to developed one but extent changes from time to time (Salomons & Grootveld, 2003). Similarly Stevenson (2001) argued that emerging markets are attractive on the basis of downside risk due to non normality of market returns. While emerging markets only offer higher returns if less integrated with developed capital markets. Usually emerging markets are less correlated with developed markets due to shortage of functional competencies in comparison to well developed capital markets, as put forward by Gupta & Donleavy (2009) that emerging markets like Chile, Greece, India, Korea, Malaysia, and the Philippines have less correlation with Australian equity market returns. Therefore emerging markets are of fundamental importance for international diversification; inclusion of emerging market can lead to optimum gains. Another crucial issue is that return co-variances change over time (King et al. 1994), similarly Makridakis and Wheelwright (1974) and Bennett and Kelleher (1988) have documented about the changing correlations over time. So in this state of affairs, volatility trends in different international stock markets including emerging markets can make the clear picture (risk and return) of the international investment scenario that in due course becomes a torch bearer for international investors to make investment decisions in the different regions of the world. As number of studies like Harvey and Bekaert and Harvey (1997) & Hassan et al. (2003) have pointed out stock market volatility in emerging capital markets in different perspectives. Those in some way differentiate developed equity markets from emerging

(10)

6 markets. Some authors have analyzed in term of comparative stock market efficiency between emerging and developed markets like Cajueiro and Tabak (2004), Kim and Shamsuddin (2008), Lim (2007) and so many others. All such studies are trying to make available insights for the determining the investment strategy in international scenario broadly. Instead of so many studies still there are number of specific questions arise in the mind of international investors in terms of what is the right time to invest, or stop in volatile international diverse set of stock markets that is not possible without understanding and evaluating the volatility closely in different stock markets.

(11)

7

3 Data and methodology

3.1 Data

Data to be used in our study is secondary, closing prices of all major indexes2 of the countries chosen as sample from the different regions of the world namely Europe, Latin America, and Asia Pacific. Countries chosen as sample from three different regions are developed and emerging markets, our selection of developed and emerging markets3 was according to latest classification laid down by FTSE group although this classification is updated every year after the evaluation of markets. FTSE is an independent body owned by The Financial Times and the

London Stock Exchange. According to ground rules set by FTSE, a country’s classification

principally depends upon the following determinants given below.

Table 1: Determinants of Country’s classification

1 Wealth (GNI per capita)

2 Total stock market capitalization

3 Breadth and depth of market

4 Any restrictions on foreign investment

5 Free flow of foreign exchange

6 Reliable and transparent price discovery

7 Efficient market infrastructure(trading, reporting and settlement systems, derivatives market etc)

8 Oversight by independent regulator Source: FTSE International Limited, 20094

Morgan Stanley Capital International (MSCI) International Equity Indices also categorizes countries into three categories namely developed, emerging and frontier (www.mscibarra.com),

2 Major indexes description collected mainly from on line resources see appendix

3 FTSE categorizes into four groups namely developed, advanced emerging, secondary emerging and frontier, we have chosen as sample the countries from developed and two groups of emerging markets to distinguish between developed and emerging markets in this study.

4

http://www.ftse.com/Indices/FTSE_Global_Equity_Index_Series/Downloads/FTSE_Global_Equity_Index_Series_I ndex_Rules.pdf

(12)

8 our selection of emerging and developed markets is also consistent with the classification laid down by MSCI.

In Table 2, we have shown Countries, their stock index, and their capital market type which constitute our sample. For information about each index description, see the appendix.

Table 2

Country & Index used Capital Market Type

Denmark (OMX Copenhagen 20 Index) Developed

France (CAC 40 Index) Developed

Germany (DAX 30 Index) Developed

Spain (IBEX 35 Index) Developed

Sweden (Stockholm OMX 30 Index) Developed

Turkey (ISE 100 Index) Second Emerging

China (Shanghai Composite Index) Second Emerging Hong Kong (Hang Seng Index - HSI) Developed

Japan (Nikkei 225 Index) Developed

India (Bombay SENSEX Index) Second Emerging

Malaysia (KLCI Index) Second Emerging

Pakistan (KSE 100 Index) Second Emerging

Brazil (Bovespa Index) Advanced Emerging

Mexico (IPC all share index) Advanced Emerging

Venezuela Second Emerging

Source: FTSE The Index

Company, 20095

We are using daily closing prices change throughout in our study, excluding cash dividends of every global index. Time period covered in this study is starting from January 1997 to December 2009 which contains more than 3000 thousand daily observations for each index. Time horizon

5

(13)

9 is not so long as it should have been due to limitations of time and financial resources. With these limitations we had to depend upon mainly on the online resources of data like www.econstats.com, www.finance.yahoo.com and www.marketwatch.com for collection of daily major index closing prices from different countries. We were restricted to only January 1997 to December 2009 due to non availability of long series of historical data of different countries from online resources especially in case of emerging markets. This period 1997-2009 constitutes the world largest financial crisis and periods of boom and recession regional economic integration, and also major financial liberalization took place in emerging markets. Notably collapse of many Asian economies in 1997, Brazilian currency crisis in 1999, and major developments regarding economic and financial integration took place in Europe. Also previous studies like Chang et al. (2004) about USA and Ridder & Djehiche (2007) about Sweden using extreme day return technique, observed at maximum frequency of extreme days during the time period late nineteen’s and initial years of new century.

3.2 Methodology

It is well known about the non-symmetric6 property of arithmetic mean of returns, geometric mean is symmetric as it takes into account the effect of continuous compounding. As a consequence just like so many other studies calculating the share price volatility in short run, we also use the standard measure daily prices/rate of change of indexes continuously compounded as follow

Ri,t%= 100*ln (Pt/Pt-1)

Here Ri,t is the continuous compounding return, Pt is the current day price and Pt-1 is the price of

previous trading day.

3.2.1 Extreme day returns definition

There is no wide-ranging definition of extreme day in literature so far, notably an extreme day Jones et al. (2004) defined as a trading day with an absolute daily logarithmic percentage change greater than or equal to 1.5% using the sample from USA the same by Ridder & Djehiche (2007) using the sample of Sweden. On the other hand, Pactwa and Prakash (2004) have used 2%, 3%, and 4% absolute change as an extreme day return. It is evident from past studies that definition

6

(14)

10 of extreme days varies from lower end 1.5% to upper end 4% absolute daily logarithmic percentage change, Considering our sample size which is short as compared to previous studies who used extreme day measure, it is appropriate to use lower limit 1.5% to include maximum number of extreme days in our analysis. As % change value defined for extreme day will be higher less extreme days will appear and vice versa.

So in this study like earlier period studies, an extreme day will be regarded as a trading day with an absolute daily logarithmic percentage change greater than or equal to 1.5%. If absolute percentage change is +1.5% or above increase that will be regarded as positive extreme day other way around decrease of - 1.5% or lower will be considered as negative extreme day. Further Positive and negative extreme day returns will be placed in upper and lower tails respectively. This decomposition will contribute to additional highly structured analysis. Decomposition will take place as follow

Upper tail: Absolute logarithmic percentage change on market returns ≥ + 1.5% Lower tail: Absolute logarithmic percentage change on market returns ≤ - 1.5%

3.2.2 Frequency of extreme day return

Frequency of extreme day return is relative frequency7 of extreme day returns defined as total number of extreme days divided by total number of trading days within specified time period. This definition will be applicable for all countries under analysis.

Frequency (extreme days) = (total number of extreme days/ total number of trading days).

For comparative study and to underline the basic features of daily logritmatic return series from dissimilar capital markets, descriptive statistics to be used measures of central tendency mean, median. Whereas measures of dispersion standard deviation, skewness and kurtosis of logritmetic return series to have an overview of insidious and predisposition about normality. It will make available the broad-spectrum outlook about the proportional volatility patterns in terms of extreme days in poles apart situated countries of our sample and ranking of capital markets on the basis of volatility. First hypothesis will also be tested on the starting point of

7 Relative frequency portraying the probability of extreme day returns occurrence (repeatedly) that can also give roughly moment view about comparatively prevailing risk in diverse markets of our sample.

(15)

11 descriptive statistics results. Frequency of extreme day return is the measure which explicitly will give the explanation of volatility in our analysis part. For the better explanation of volatility; frequency of extreme days will be arrayed according to yearly time periods, calendar months, and week days, from which also some well known anomalies like Monday effect, January effect can be highlighted in case of every country.

3.2.3 Predictive power

To measure the predictive power of frequency of extreme days also simple OLS regression model is to be employed taking into account the frequency of extreme day return in time yeart as

a dependent variable while independent variable is the same but of the previous year yeart-1. Frequencyext day(t) = a + β(Frequencyext day(t-1)) + ε………...(1)

Fabozzi and Francis (1977) using the sample single market index 700 NYSE analyzed about the empirical results of regression risk and return model and concluded not significant difference on results in two types of market conditions Bull and Bear. On the other hand in terms frequency of extreme days, it is interesting to see whether different market conditions contribute to the occurrence of extreme day return frequency in comparison of developed and emerging markets of the world. Recent study on Sweden in this regard by Ridder & Djehiche (2007) measuring the relationship between large negative changes in market(bear) and frequency of extreme days concluded frequency of extreme days is not influenced significantly by previous year market conditions. To find this evidence comparatively in developed and emerging markets OLS model 2 is put by introducing a dummy variable bear market in the previous year in model 1. If previous year is regarded as bear variable will contain value 1 and zero otherwise.

Frequencyext day(t) = a + b1(Frequencyext day(t-1) ) + b2(Bear markett-1) + ε……….(2)

Bear market is the condition when sharp downward pressure on prices is observable, although there is no agreed upon definition of bear market in terms of % decrease in prices, different researchers defined ranging from – 15% to -10% on overall market returns as evident from two notable studies using extreme day approach by Jones et al. (2004) Ridder & Djehiche (2007) respectively. As our study constitutes larger proportion of emerging markets, we are categorizing

(16)

12 bear market condition if -15% and -20% or less change is observed on overall market returns during a particular year.

3.2.4 Co-movement among frequencies of extreme day returns and level of integration among the capital markets

We measure the level of integration of different capital markets to get an over view for optimal international diversification decision in the selected sample of our study. In the modern era where we call world a global village, investors go for international diversification. International diversification is beneficial if worldwide stock markets are least correlated. Greater the integration of international stock markets and correlated stock price volatility decreases the prospects for international portfolio diversification (Bekaert and Harvey, 2003). It is the capital market integration, which leads to correlation among stock prices. Well documented factors behind the capital market integration discussed in academia are mainly financial liberalization, economic corporation and integration (increase in cross border trade by curtailing the tariff barriers). In recent two decades there were intense developments about economic integration and financial liberalization all over the world. But there are evidences of volatility spillover among capital markets even in the absence of real and financial ties. How this is happening? Answer lies in the question how international diversification has become possible? International diversification became possible due to advancements in Information and communication technologies (ICT) mainly, at the same time that also became the free source of economic and relevant financial information flow. In emerging capital markets ICT advancement contributed to financial linkages, growth and volatility during 1990s (www.unescap.org). It is proved empirically that free information flows considerably become the source of contagion among the capital markets internationally. As King and Wadhwani (1990) put forward the cause of volatility spillover, information about price changes in other markets incorporated by rational agents. Similarly evidence of information revealance in volatility was proved empirically by Kyle (1985). Most recently by Mukherjee & Mishra (2010) documented the return and volatility correlation caused by transmission of market information among India and its major corresponding markets in Asia. In such state of affairs it is very rare for any capital market of the world to be isolated and independent functioning. There are number of studies which give evidence about the prevailing interdependence/integration among the world wide capital markets

(17)

13 including our sample like Diamandis (2008), Lin et al. (2008), Ko (2008) Ndu (2007) , Hyde et al. (2007), Kearney(2001), Ayuso & Blanco (1999).

While it is well documented that during volatility return series of worldwide markets correlate more and benefits of international diversification fade away. As Butler and Joaquin (2001) concluded that cross-market correlations rise in periods of high volatility, reducing the benefits of portfolio diversification at the times when they are required intensively and some other studies give similar conclusions. Ramchand & Susmel(1998) empirically found that during high variance(volatility) in US, distant markets correlate with Us markets and argue it as allusion for international portfolio diversification. We address this issue in our study by conducting cross-sectional correlation analysis among all market’s yearly frequencies of extreme days using the Pearson Correlation statistic as shown below.

Where xi and yi are the yearly frequencies of extreme days for particular markets.

Findings of this analysis would provide idea, to how much extent selected capital markets of our study are associated in terms of frequency of extreme days. In other words how much volatility correlation prevails among the markets of our sample, as extreme day returns is being used as measure of volatility in this study. Cross section correlation has been utilized in the field of sciences and social sciences to measure the association at glance association among different variables of interest. Similarly applied among the return series of worldwide capital markets for the determination of co movement among capital markets. While correlation among frequencies of extreme day return is more comprehensive as it takes into account the return and volatility at the same time. That is prime concern to be addressed for diversification strategy.

Correlation coefficient is unit free quantification of linear association between two variables. In this study we are calculating correlation among the yearly frequencies of extreme day returns in capital markets of our sample. At the same time it should be kept in mind, that correlation does not say everything about causality (specification of relationship in terms of independent and dependent variables). As correlation value ranges from -1 to +1 portraying perfectly negative and

(18)

14 positive association respectively, more specifically together movement in opposite direction and in the same direction. Combination of capital markets with stronger correlation (coefficient close to +1) among yearly frequency of extreme day returns during the specified time period would exhibit fewer or no chances of international diversification gains. On the other end low (coefficient somewhere close to zero) or negatively correlated combination, would provide reasonably high chances of preferred gains. The impetus behind the correlation analysis is to endow with insight about interdependence and co movements into the international capital markets. That can provide prospects to minimize diversifiable risk to make certain higher returns through international diversifications.

(19)

15

4 Results

We present summarized statistics for all capital markets during 1997-2009 in Table 3. Statistics are based on logarithmic percentage returns, no of observations in all markets is not the same in the specified time period. At the same time as all these markets are economically, regionally, religiously and culturally diversified, trading days being different thus number of return observations are not the same. Some remarkable points can be observed from these statistics, as reported in Table 3 mean and median of logarithmic percentage return series for all capital markets do not coincide. Thus return series of any capital market does not support the general idea about normally distributed around mean, eventually same is reinforced by the values of skewness and kurtosis. Value of skewness shows a discrepancy across the all markets, with nine of them twisted to left and rest of the six to the right indicating non symmetric data sets. while Mean and median values differ sharply in case of three emerging markets Mexico, India and Pakistan. Kurtosis values indicate that all series have a thicker tail and a higher peak relative to normal distribution, intensity is higher among emerging markets. Beginning with China having the maximum value of kurtosis 91.55 containing the fattest tails among the all markets of sample, on March 1st 1999 lowest in upper tail -41.5% and highest in upper tail highest 40,06 on February 10 1999, highest values of extreme day returns in over all sample. Also observed minimum and maximum values of extreme days, most extreme days ranging +17% and above in upper tail and less than -17% and less in lower tail are found in China, Malaysia & Turkey. Moreover, most of the developed markets in the respect of minimum and max extreme values seems to more stable than of developing ones. In terms of kurtosis next to china place occupied by Malaysia 43,4 ,then stand Venezuela and Brazil with kurtosis value 20,2 and 11,95 respectively not surprisingly all four are emerging capital markets. Other way around least value of kurtosis is observable among developed capital markets, In case of Sweden 3,103 which almost near normality, next Germany 3,63 and then Spain , France 4,093, 4,42 respectively. Thus largest values of extreme day returns (fat tails) occurred in emerging markets and on the other end associated with developed capital markets in the sample during the specified time period. This is demonstration of more deviation from normality in the data sets of emerging capital market supporting our first hypothesis in this study.

In Table 4 we rank volatility of capital markets measured by standard deviation and percentage of extreme days. As reported in Table 4 Turkey with standard deviation 2,833% ranked first

(20)

16 having highest volatility in terms of standard deviation among the all capital markets in sample. Next to Turkey stands Brazil ranked second with standard deviation 2,351% then China 2,145%. Top three volatile capital markets measured by standard deviation are emerging markets and from all three regions Europe, Latin America and Far East. On the other end according to volatility ranked lowest are three developed capital markets of Europe Denmark lowest ranked on 15th with lowest standard deviation 1,345%, Spain 14th then France 13th with standard deviations 1,514% and 1,546% respectively.

Volatility ranking with percentage of extreme day returns places again highest levels in ranks Turkey and Brazil with 48,429% and 41,612% extreme day returns respectively, in the same rank array granted by standard deviation. Other way around lowest place in rank is also an emerging market Malaysia with 16,116% extreme days. Hong Kong and India occupy the same rank 4th and 5th respectively in terms of frequency of extreme days. Pakistan and Sweden move up to next higher rank 6th and 7th respectively relative to the ranking 7th and 8th using standard deviation. The ranking of Japan, Germany, china Mexico, France, Spain, Venezuela and Denmark granted by extreme day return are different as per standard deviation. These observations8 do not support our first hypothesis at large. Generalizing these findings, evidence is found to conclude about occurrence of more extreme day returns in emerging capital markets in comparison to developed capital markets. At the same time it cannot be concluded explicitly that volatility in stock returns measured by extreme day return occurrence is the only phenomenon of emerging markets during the selected time period at least. As earlier mentioned this time period contained many booms and recessions including global crisis, Brazilian crisis in 1999, and Asian currency crisis 1997.

Another well acknowledged feature is of higher returns associated with higher volatility in emerging capital markets relative to developed capital markets. As reported in Figure 1 % of extreme days in upper and lower tails during 1997-2009 for all capital markets. Upper tail is heavier than lower tail in all emerging capital markets, magnitude of frequencies in upper tail is more than lower tail in the mentioned time period. More specifically occurrence of positive extreme day returns is more than occurrence of negative extreme day returns in emerging capital

8Ranking was also done defining extreme day return ±2% absolute logrithmetic change, Turkey, Brazil ranked in same order highest one in volatility. While third among highest in ranking stood India, Pakistan moved to 5th China 7th and Mexico 10th highest in volatility ranking, Malaysia at last exceptionally.

(21)

17 markets. Yearly frequencies also show the same pattern at most in figure 2. These observations give evidence of higher return opportunities in emerging markets all through volatility during 1997-2009. This is sustaining a general view about higher returns in emerging markets as pointed out by many prior researches. Despite the fact that vice versa with reference to some developed capital markets not observed like Japan, Hong Kong and Denmark, whose upper tails are heavier than lower tails in overall time period. Contrary to view of higher returns in emerging markets, evidence of higher returns are also found in developed capital markets as well. Similar findings were put forward by Ndu (2007) during the time period 1997-2004 analyzing the daily returns and volatilities of returns of 40 developed and emerging worldwide stock markets. Similarly we are also powerless to conclude overtly that in terms of extreme day return analysis during 1997-2009, only emerging markets exhibit the opportunities of higher returns.

Next we investigate about well known anomaly January effect among all the markets. As reported in Table 5 among most countries observed extreme days in the month of January are more in number as compared to other months. By analyzing the number of positive extreme days, it has been found that in almost all countries positive extreme days are higher in the months of January as compared to December which in some way verify the famous anomaly “The January Effect” in which market usually goes up during the initial days of January. These finding also in some way make a sense that history repeat itself (i.e. future returns can be predicted based on passed returns) which to some extent conflicts with efficient market hypothesis.

While the analysis of minimum and maximum values of extreme days in different countries most extreme values are found in emerging e.g. China, Malaysia, and Turkey where extreme values occurred more than 17% and less than -17% returns which in some way shows that these market are highly volatile and can give higher rates of returns with the higher level of risk. Most Developed markets had shown more stable minimum and maximum values (if we compared these month by month) as compared to most of the developing ones which show greater variation in their minimum and maximum values from one month to another.

Reportings in Table 6 suggest the more occurrences of extreme days on Monday in most of the markets, also pointed out by Ridder & Djehiche (2007) analyzing the extreme days for Sweden.

(22)

18 With the analysis of negative extreme days, one of the famous anomalies Monday effect (which states that market usually goes down on Monday) can also be observed to some extent as in our case, in most of the countries, negative extreme days on Monday are more in number as compared to Friday.

Next we see the predictive power of extreme days to predict the extreme days among different capital markets of our sample. We use simple OLS regression model, in which dependent variable is ratio of total extreme days to total trading days in a year t and independent variable is the same but of the previous year t-1.

Frequencyext day(t) = a + β (Frequencyext day(t-1) ) + ε

As reported in Table 7 all developed capital markets estimated slope coefficient β is positive for the overall sample period 1997-2009, but statistically significant at 10% level only in case of Sweden, Germany, and France, while statistically significant at 5% level in case of Hong Kong. Coefficients put forward the positive trends in volatility in terms of frequency of extreme days on lagged volatility. Estimates give evidence about the correlation of current year extreme day returns with previous year. Thus give the evidence of predictive power, but extent varies among the different developed capital markets, as indicated by the value of R2 0,49 for Hong Kong, 0,286 in case of Germany, 0,28 and 0,25 for Sweden and France respectively. On the other end Lowest predictive power is reported for Denmark and Japan with R2 values 0,073 and 0,097 respectively but not statistically significant. While with reference to emerging capital as reported estimated coefficients for over all time period 1997-2009 also show positive trend in frequency of extreme day returns, correlation and eventually the predictive power. And similar to developed capital markets extent of predictive power varies among all emerging capital markets as shown by estimated value of R2. Highest on the basis of predictive power as per estimates among emerging capital markets stood turkey with R2 value of 0,49, secondly Malaysia and then china with R2 values of 0,47and 0,43 respectively statistically significant at 5% level. In case of Mexico reported value of R2 is 0,32 but statistically significant at 10% level. Correlation of extreme day returns frequencies with prior year is found to be stronger in case of emerging markets. Thus more predictive power of frequency of extreme dates exists to predict frequency of extreme dates. These observations support our second hypothesis. While In case of India, Pakistan and Venezuela reported R2 is low at the same time not statistically significant. Possible

(23)

19 explanation in case of India and Pakistan is the regional and domestic uncertainty during the 1997-2009, series of domestic and regional tensions prevailed reflected in return series as well. Most of negative days in these markets are associated with incidents took place during the tenure. And after 9/11 uncertainty was even more after the start of afghan war. Due to these factors markets had not operated normally, thus return series are not portraying the reasonable results. Similarly Venezuela market is found showing return series9 patterns which does not lay down proper statistical findings.

More predictive power among emerging capital markets can be mere explanation of lacking in stock market efficiency (infrastructure, reporting), lack of independent regulatory authority. This is general view about emerging markets well documented in literature. Vice versa can be the case of developed capital markets where reportedly low predictive power. These findings justify the criteria laid down by FTSE.

Frequencyext day(t) = a + b1(Frequencyext day(t-1) ) + b2(Bear markett-1) + ε

Next we re-estimate the initial model by introducing dummy variable bear market condition in the previous year. Unfortunately none of the result appears statistically significant. Perhaps it was mainly due to short data series only two years in every market were reported bear market condition by defining -15% and -20% as well. During the selected time period year 2008 highest bear market condition (in all markets -50% down turn) very due to global recession. While during initial years of the new century simultaneously some markets appeared to be in bull market condition. Although coefficients for bear market are not statistically significant, as reported in Table 8 but coefficients of bear market condition are found to be negative and positive across all the markets with the reasonable value of R2.

Next we look into the level of integration and co movement of frequencies of extreme days during the 1997-2009.

9 There is no explicit explanation for this pattern, as we report in the correlation analysis part Venezuela is found to be isolated just like Pakistan and India.

(24)

20 We report in Table 9 the correlation coefficients in the sub period 1997-2000, 2001-2005 and 2006-2009 in European capital markets, Latin American and Far East Asian capital markets. Developed capital markets seem to be well integrated as these countries have economic and financial ties, emerging market of turkey relatively low. Reported values advocate correlation at rise with the passage of time. During 2001-2005 some decreasing co movement is observable at the turn of new century. Possible explanation for this is the crisis faced by European economies in the initial years of the century.

While in case Latin American markets trend of integration was not consist, as all are emerging economies secondly Brazilian currency crisis also took place during 1999. Venezuela appeared to be isolated within Latin markets with respect of rest of the markets as well; such isolation was stated in emerging market correlation report (www.bradynet.com). Similar pattern was reported among the Asian capital markets. Developed markets like Japan and Hong Kong substantially showed more co movements with other developed markets in the sample. China appeared to be less integrated, while Pakistan and India remain isolated between each other due to ongoing tensions. Particularly Pakistan has been the much isolated among all other markets possibly due to prevailing domestic political, law and order situations particularly after the 9/11 which led to Afghan war.

In Table 10 we report the estimated cross correlation during the sub periods 1997-2000, 2001-2005 and 2006-2009 and overall sample period 1997-2009 among all markets. In all the capital markets under analysis, estimates depict the integration at rise with the passage of time, after the turnaround of new century sharp increase in correlation coefficients depict the impact of intense liberalization especially in emerging markets. While on the other hand, most of emerging markets got well connected with world markets due to abundant developments in ICT. At the same time major developments regarding economic liberalization also took place in Europe. Into the new sub period most recent 2006-2009 co movements further enhanced except Venezuela.

At large correlation coefficients are positive and of substantial quantity, suggesting the presence of co movement among capital markets thus rarely any market functioning in isolation as we previously argued. While correlation has increased with the passage of time, as reported in sub periods 1997-2000 and onward till 2006-2009. As reported in the overall 1997-2007 panel Mostly Highest coefficients are observed among all the developed capital markets of our sample

(25)

21 like Sweden, Spain, Germany, France, Spain, Japan and Hong Kong. Even higher among the developed capital markets located within European Union (EU) namely Sweden, Spain, Germany, France indicating the strong real and financial ties that is of course due to the fact that EU is an economically integrated unit. An additional barrier to perfect financial integration is different currencies (risk of exchange rates). Markets located in euro zone without exchange rate risk like France, Germany, and Spain, highest correlation coefficient reported among all estimates is between France and Spain (0,963) very near to 1.As portfolio diversification theory suggests where there is presence of highly positive correlation among the markets, higher returns cannot be expected by including stocks from such markets. Accordingly is the case in all the developed capital markets of the EU and rest of the developed markets located in Asia like Japan and Hong Kong. Thus portfolio diversification merely including all the developed capital markets of our sample cannot yield higher returns. This highlights the need to include some emerging market least correlated for higher returns.

Looking into the emerging capital markets correlation coefficients of frequencies of extreme day returns depict the presence of much financial and volatility contagion among each other and with developed markets as well. Like Turkey which has been reported highest in volatility in this study is found to be well integrated with developed markets as well possible reasons its geographical location and intense financial liberalization and real financial ties with European countries. Its correlation coefficient with European markets like Sweden, Spain, Denmark, Germany and France is 0.67, 0.71, 0.37, 0.63 and 0.65 respectively while with Asian developed markets namely Japan and Hong Kong is 0.5353 and 0.5911 respectively. Next Brazil fairly well integrated with developed markets of Europe. While also strong correlation coefficients appear with turkey (0.81), Mexico (0.67), Venezuela (0.55). Similarly Malaysia has been reported to have co movement with Hong Kong, turkey with Latin American markets but not considerably with European developed capital markets. China and Pakistan also appear some isolated from European developed capital markets and Japan.

These observations suggest Malaysia, Venezuela, China and Pakistan are the potential emerging markets to be included for diversification with developed capital markets but by taking into consideration the other risk factors of international diversification like political, inflation, exchange rate and liquidity risks as this study does not cover such factors.

(26)

22

5 Conclusion and recommendations for future research

During 1997-2009 all the emerging capital markets showed heavy upper tails, an indicator of higher return opportunity during volatility. In terms of extreme day return occurrence most volatile emerging capital market appeared Turkey, while least volatile Malaysia emerging as well. Volatility ranking measured by extreme day return approach contradict the ranking by standard deviation, perhaps it is due to the flaws with standard deviation to measure volatility comprehensively, argued by researchers using extreme value observations. At the same time estimates direct not to associate volatility to emerging markets merely especially during the time when booms and recessions are taking place. Yearly frequency of extreme day show more predictive power relative to developed markets. Co movement among the capital markets has increased over time reducing the possibility of global diversification gains.

We have tried to dig into the very vast and interesting issue, which requires more elaborated analysis. Due to the limitation of time and resources we were restricted to conclude general findings with limited data set. Future research can be extended using extreme day return approach utilizing the more sophisticated techniques of operational research and statistics particularly for measuring the co movement and global market integration with frequency of extreme days with long series of historical data. Further increasing co movement among the capital markets highlights the need for some international governance body, which is not being discussed in academia much.

(27)

23

References

Achour, C.R., Harvey, G. H. and Lang C., 1999. Stocknext Selection in Malaysia, Emerging Markets Quarterly, (Spring), pp. 54–91.

Assaf, A., 2009. Extreme observations and risk assessment in the equity markets of MENA region: Tail measures and Value-at-Risk. International Review of Financial Analysis, 18(3), 109-116.

Ayuso, J. & Blanco, R., 1999. Has Financial Market Integration Increased during the Nineties?, Banco de España Working Papers 9923, Banco de España. Available at: http://ideas.repec.org/p/bde/wpaper/9923.html [Accessed May 11, 2010].

Barry, C.B., Goldreyer E., Lockwood L. J. and Rodriguez M., 2002. Robustness of size and value effects in previous equity previous markets. 1985-2000 (May 2001). Texas Christian University Center for Financial Studies Working Paper. Available at SSRN:

http://ssrn.com/abstract=270226 or doi:10.2139/ssrn.270226.

Bekaert, G. & Harvey C. R., 1997. Emerging Equity Market Volatility. September, research paper, Duke University, (online SSRN).

Bekaert, G. & Harvey C.R., 2003. Emerging Markets Finance. Journal of Empirical Finance, 10, pp. 3–55.

Bennett, P. and Kelleher, J., 1988. The international transmission of stock price disruption in October 1987. Federal Reserve Bank of NY Q. Rev. 12, pp. 17–33.

Butler, K.C. & Joaquin, D.C., 2001. Are the Gains from International Portfolio

Diversification Exaggerated? The Influence of Downside Risk in Bear Markets, EFMA 2002 London Meetings, (Online SSRN).

BradyNet. 2001. Emerging Market Correlation Study, [Online] Available at: http://www.bradynet.com/acrobat/Correlations.pdf [Accessed 12 May 2010]

Cajueiro, D.O. & Tabak, B.M., 2004. Ranking efficiency for emerging markets. Chaos,

Solitons & Fractals, 22(2), pp. 349-352.

Chang, E.J., Lima, E.J.A. & Tabak, B.M., 2004. Testing for predictability in emerging equity markets. Emerging Markets Review, 5(3), pp. 295-316.

Chukwuogor, C. & Feridun, M., 2007. Recent Emerging and Developed European Stock Markets Volatility of Returns. European Journal of Finance and Banking Research, 1 (1), pp. 28-46, (online SSRN).

Diamandis, P.F., 2008. Financial liberalization and changes in the dynamic behaviour of emerging market volatility: Evidence from four Latin American equity markets. Research in

(28)

24 Du, N. & Budescu, D.V., 2007. Does past volatility affect investors' price forecasts and confidence judgements?. International Journal of Forecasting, 23(3), pp. 497-511.

EconStats International. Daily Data about indices. [Online] Available at: http://www.econstats.com/eqty/eq_d_la_6.htm [Accessed March 24 2010].

Fabozzi, F.J. & Francis, J.C., 1977. Stability Tests for Alphas and Betas Over Bull and Bear Market Conditions. The Journal of Finance, 32(4), pp. 1093-1099.

FTSE International limited, 2009. FTSE Global Equity Index Series, Country Classification, [Online] FTSE International limited. Available at:

http://www.ftse.com/Indices/Country_Classification/Downloads/FTSE_Country_Classificati on_Sept_09_update.pdf [Accessed March 24 2010].

FTSE International Limited, 2009. Ground Rules for the Management of the FTSE Global

Equity Index Series, [Online] FTSE International limited. Available at:

http://www.ftse.com/Indices/FTSE_Global_Equity_Index_Series/Downloads/FTSE_Global_ Equity_Index_Series_Index_Rules.pdf. [Accessed March 24 2010].

Hassan, M. K., Islam A. M., and Basher S., 2003. Time-Varying Volatility and Equity Returns in Bangladesh Stock Market. Finance 0310015, (online EconWPA).

Hyde, S., Bredin, D. & Nguyen, N.T., 2007. Correlation Dynamics Between Asia-Pacific, EU and US Stock Returns. SSRN eLibrary. Available at:

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1090524 [Accessed May 11, 2010]. Gupta, R. & Donleavy, G., 2009. Benefits of diversifying investments into emerging markets with time-varying correlations: An Australian perspective. Journal of Multinational

Financial Management, 19(2), pp. 160-177.

Jondeau, E. & Rockinger, M., 2003. Testing for differences in the tails of stock-market returns. Journal of Empirical Finance, 10(5), 559-581.

Jones, C. P., Walker M. D. and Wilson J. W., 2004. Analyzing Stock Market Volatility Using Extreme-Day Measures. Journal of Financial Research, 27(4), pp. 585-601.

Kim, J.H. & Shamsuddin, A., 2008. Are Asian stock markets efficient? Evidence from new multiple variance ratio tests. Journal of Empirical Finance, 15(3), pp. 518-532.

King, M.A. & Wadhwani, S., 1990. Transmission of Volatility between Stock Markets. The

Review of Financial Studies, 3(1), 5-33.

King, M., Sentana, E. & Wadhwani, S., 1994. Volatility and Links between National Stock Markets. Econometrica, 62(4), 901-933.

Ko, K.W., 2008. Financial integration, information and communication technology, and macroeconomic volatility: Evidence from ten Asian economies. Research in International

(29)

25

Business and Finance, 22(2), 124-144.

Kyle, A.S., 1985. Continuous Auctions and Insider Trading. Econometrica, 53(6), 1315-1335 Lim, K., 2007. Ranking market efficiency for stock markets: A nonlinear perspective.

Physica A: Statistical Mechanics and its Applications, 376, pp. 445-454.

Lin, K., Menkveld, A.J. & Yang, Z., 2009. Chinese and world equity markets: A review of the volatilities and correlations in the first fifteen years. China Economic Review, 20(1), 29-45.

Longin, F. M., 1996, The asymptotic distribution of extreme stock returns, Journal of

Business, 69, 383–408.

Makridakis, S.G. and Wheelwright, S.C., 1974. An analysis of the interrelationships among the major world stock exchanges. J. Bus. Fin. Accounting 1, pp. 195–216

MSCI Barra. MSCI Frontier Market Indices. [Online] Available at:

http://www.mscibarra.com/products/indices/international_equity_indices/ [Accessed March 28 2010].

Muchnik, L., Bunde, A. & Havlin, S., 2009. Long term memory in extreme returns of financial time series. Physica A: Statistical Mechanics and its Applications, 388(19), 4145-4150.

Mukherjee, K.N. & Mishra, R.K., 2010. Stock market integration and volatility spillover: India and its major Asian counterparts. Research in International Business and Finance, 24(2), 235-251.

Ndu, C., 2007. Stock Markets Returns and Volatilities: A Global Comparison. Global

Jounrnal of Finance & Banking Issues, 1(1), pp. 1-21. (online SSRN).

Nelken, I., 1997. Volatility in the Capital Markets: State Of-The-Art Techniques for

Modeling, Managing and Trading Volatility, illustrated March edn, Eric Dobby Publishing.

Pactwa, T. & Prakash, A., 2004. Using Extreme Value Theory to Value Stock Market Returns, Working Paper. Florida International University, Miami.

Ridder, A. D. & Djehiche B., 2008. Extreme Day Returns on Stocks and Investment Behavior: Evidence from Sweden. In: FMA, International Conference, United States, October 2008.

Salomons, R. & Grootveld H., 2003. The equity risk premium: emerging vs. developed markets. Emerging Markets Review, 4(2), pp. 121-144.

Stevenson, S., 2001. Emerging markets, downside risk and the asset allocation decision.

Emerging Markets Review, 2(1), pp. 50-66.

(30)

26

Social Survay of Asia & the Pacific. [Online] (updated 20 Dec 1999) Available at:

http://www.unescap.org/drpad/publication/survey1999/svy7b.htm [Accessed March 12, 2010]

Wander, B. H. & Vari R. D., 2003. The Limitations of Standard Deviation as a Measure of Bond Portfolio Risk. The Journal of Wealth Management 6, 3(1), pp. 35-38.

World Federation of Exchanges. Domestic Market Capitalization. [Online] Avaiable at: http://www.world-exchanges.org/statistics/ytd-monthly [Accessed April 5 2010].

Xing, X., (2004). Why Does Stock Market Volatility Differ Across Countries? Evidence from Thirty-Seven International Markets. International Journal of Business, 9(1), pp. 83-102, (online SSRN).

(31)

27

Appendix

Capital markets Total daily observation Mean Median Skewness Kurtosis Standard Deviation

Sweden(devel oped) 3261 0.022 0.063 0.134 3.103 1.685 Spa i n(devel oped) 3267 0.027 0.090 -0.123 4.093 1.514 Germa ny(devel oped) 3295 0.022 0.094 -0.031 3.625 1.652 Denma rk(devel oped) 3246 0.028 0.063 -0.296 5.381 1.345 Fra nce(devel oped) 3305 0.017 0.046 -0.025 4.419 1.546 Turkey(emergi ng) 3183 0.125 0.118 -0.032 4.939 2.833 Bra zi l (emergi ng) 3213 0.071 0.146 0.330 11.811 2.351 Mexi co(emergi ng) 3256 0.069 0.112 0.032 6.155 1.659 Venezuel a (emergi ng) 3131 0.067 0.011 0.238 20.171 1.786 Hong kong(devel oped) 3228 0.016 0.043 0.142 8.897 1.883 Ja pa n(devel oped) 3190 -0.019 0.001 -0.207 5.286 1.630 Ma l a ys i a (emergi ng) 3208 0.001 0.016 0.429 43.374 1.619 Chi na (emergi ng) 3347 0.030 0.008 -0.413 91.548 2.145 Indi a (emergi ng) 3206 0.052 0.118 -0.122 5.002 1.790 Pa ki s tan(emergi ng) 3150 0.061 0.123 -0.348 4.741 1.780

Statistics are based on observations of daily changes from 1997 to 2009 for major stock indices (known as representative of whole market endorsed by international institutions). The basic formula for the logarithmic percent change is Rit% = 100*ln (Pt/Pt-1). Number of total daily observations are not same

among all different markets(emerging)

(32)

28 Rank Capital markets Standard Deviation Capital markets

% total extreme days Rank 1 Turkey(emergi ng) 2.83% Turkey(emergi ng) 48.429 1 2 Bra zi l (emergi ng) 2.351 Bra zi l (emergi ng) 41.612 2 3 Chi na (emergi ng) 2.145 Ja pa n(devel oped) 31.242 3 4 Hong Kong(Devel oped) 1.883 Hong Kong(Devel oped) 30.4 4 5 Indi a (emergi ng) 1.79 Indi a (emergi ng) 29.884 5 6 Venezuel a (emergi ng) 1.786 Pa ki s tan(emergi ng) 28.762 6 7 Pa ki s tan(emergi ng) 1.78 Sweden(devel oped) 28.309 7 8 Sweden(devel oped) 1.685 Germa ny(devel oped) 27.478 8 9 Mexi co(emergi ng) 1.659 Chi na (emergi ng) 27.121 9 10 Germa ny(devel oped) 1.652 Mexi co(emergi ng) 26.106 10 11 Ja pa n(devel oped) 1.63 Fra nce(devel oped) 24.429 11 12 Ma l a ys i a (emergi ng) 1.619 Spa i n(devel oped) 24.287 12 13 Fra nce(devel oped) 1.546 Va nzuel a (emergi ng) 22.134 13 14 Spa i n(devel oped) 1.514 Denma rk(devel oped) 19.613 14 15 Denma rk(devel oped) 1.345 Ma l a ys i a (emergi ng) 16.116 15

Rank of volatility as measured by standard deviation and percentage of total extreme day occurrence during the time period 1997 to 2009 in selected emerging and developed capital markets of the sample. Percentage of total extreme days is calculated total extreme day (percentage logritmetic change greater than or equal to +1.5% and

less than or equal to -1.5%) observations divided by total observations.

Table 4 - Ranking

Ranked as per percentage pf total extreme days Ranked as per standard deviation

(33)

29

Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max

Jan 44 47 34.47 -7.70 7.12 Jan 42 41 30.40 -7.49 6.16 Jan 44 39 31.56 -20.66 8.40 Feb 43 30 28.74 -5.24 4.71 Feb 47 33 30.77 -9.26 40.06 Feb 30 23 22.18 -4.85 9.91 Mar 47 48 35.98 -6.22 8.59 Mar 38 40 27.37 -41.50 5.94 Mar 21 22 15.93 -4.25 8.33 Apr 48 37 34.14 -8.62 6.99 Apr 43 35 28.16 -23.24 21.94 Apr 27 34 24.30 -8.49 13.16 May 42 43 31.25 -11.81 15.99 May 38 30 24.29 -6.72 5.31 May 35 25 22.90 -4.91 6.45 Jun 42 45 31.29 -5.98 7.31 Jun 46 45 32.73 -8.62 8.85 Jun 30 19 19.52 -6.78 5.23 Jul 49 39 30.56 -6.37 5.77 Jul 42 43 29.41 -7.92 6.26 Jul 27 27 20.30 -4.63 5.17 Aug 34 34 25.09 -4.38 3.62 Aug 31 41 25.17 -8.73 7.35 Aug 21 39 21.74 -9.45 4.10 Sep 33 38 26.10 -6.03 5.31 Sep 34 43 27.90 -7.04 9.03 Sep 47 23 25.18 -7.83 7.54 Oct 41 47 33.85 -11.60 7.90 Oct 37 44 28.93 -17.80 18.82 Oct 26 31 20.58 -8.71 4.53 Nov 43 29 27.69 -6.84 5.58 Nov 33 34 24.19 -6.51 7.02 Nov 27 24 18.82 -4.64 9.90 Dec 42 27 25.18 -3.92 5.37 Dec 27 25 18.18 -4.66 4.12 Dec 35 17 22.91 -10.80 20.06

Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max

Jan 42 34 30.40 -9.99 5.87 Jan 27 22 19.37 -4.50 8.70 Jan 67 64 48.52 -10.50 28.83 Feb 44 30 31.76 -5.13 5.14 Feb 23 16 17.18 -6.03 20.82 Feb 58 39 40.42 -6.86 8.43 Mar 51 41 34.98 -4.52 5.30 Mar 19 19 13.62 -9.98 3.52 Mar 70 55 45.13 -5.44 8.41 Apr 30 31 23.37 -4.70 5.80 Apr 24 18 15.44 -6.34 4.80 Apr 51 43 36.15 -5.89 6.14 May 35 57 34.07 -7.74 8.51 May 23 22 17.05 -5.01 4.50 May 58 52 40.59 -6.64 6.38 Jun 51 47 35.64 -13.21 12.76 Jun 17 22 14.18 -4.35 4.54 Jun 57 53 40.74 -5.48 6.77 Jul 46 31 27.02 -6.93 7.16 Jul 25 19 15.33 -4.85 3.50 Jul 52 50 36.69 -8.90 8.44 Aug 45 41 31.97 -6.20 5.49 Aug 16 32 17.39 -8.00 8.44 Aug 51 58 38.25 -10.48 5.03 Sep 28 25 19.85 -4.60 4.16 Sep 21 41 22.71 -24.15 20.26 Sep 59 59 43.70 -17.21 17.13 Oct 40 36 27.74 -9.93 6.45 Oct 22 21 15.47 -6.88 6.52 Oct 64 67 46.79 -16.21 13.68 Nov 33 35 26.77 -4.67 10.00 Nov 21 16 14.29 -11.74 7.54 Nov 69 45 44.53 -10.76 9.26 Dec 23 30 21.29 -6.25 4.71 Dec 24 15 14.72 -7.70 10.77 Dec 56 32 34.38 -9.20 7.98

Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max

Jan 40 53 35.36 -9.10 10.18 Jan 40 43 33.74 -5.81 5.93 Jan 34 42 28.25 -5.32 6.60 Feb 34 31 26.64 -5.55 13.40 Feb 22 26 19.20 -4.82 4.58 Feb 27 33 22.90 -5.08 4.21 Mar 39 46 30.25 -5.32 6.23 Mar 56 44 36.23 -5.17 7.22 Mar 42 42 30.11 -7.20 5.54 Apr 40 29 27.38 -8.94 7.15 Apr 39 29 25.37 -7.23 4.30 Apr 47 28 28.85 -4.89 5.73 May 33 37 25.64 -5.40 5.39 May 34 31 25.29 -4.97 4.45 May 28 38 25.19 -4.88 5.35 Jun 38 35 26.94 -5.89 6.20 Jun 39 31 25.09 -4.23 4.30 Jun 27 38 25.29 -4.65 4.63 Jul 41 33 26.91 -3.89 4.71 Jul 32 38 25.27 -3.47 3.43 Jul 50 46 32.99 -5.26 8.84 Aug 39 48 30.63 -7.35 8.13 Aug 34 40 25.87 -4.06 3.77 Aug 37 43 27.97 -4.40 4.77 Sep 42 46 32.35 -9.29 9.18 Sep 37 50 34.12 -6.86 5.18 Sep 38 47 30.47 -8.53 8.60 Oct 50 53 38.29 -14.73 17.25 Oct 39 52 32.97 -12.11 13.23 Oct 53 49 35.29 -7.51 11.02 Nov 51 38 32.13 -7.34 5.77 Nov 43 39 32.28 -7.14 7.66 Nov 44 37 29.24 -6.54 9.87 Dec 38 34 26.97 -5.63 8.30 Dec 36 39 28.20 -6.56 5.08 Dec 33 33 26.40 -5.96 9.10

Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max

Jan 41 45 30.94 -6.17 7.49 Jan 35 31 24.35 -7.84 6.72 Jan 30 34 23.02 -7.43 7.09 Feb 24 20 17.67 -5.98 4.80 Feb 32 31 24.14 -5.33 3.91 Feb 26 43 26.54 -4.87 4.56 Mar 46 36 30.48 -5.03 5.86 Mar 38 35 26.16 -4.71 4.81 Mar 44 39 29.64 -6.34 6.64 Apr 35 28 24.14 -8.27 6.18 Apr 34 23 22.09 -3.52 4.59 Apr 37 26 24.14 -4.16 5.90 May 34 35 25.46 -4.16 5.78 May 26 30 20.36 -3.78 3.42 May 32 34 24.35 -4.74 3.98 Jun 41 39 28.99 -4.75 6.51 Jun 29 34 22.66 -4.04 3.19 Jun 31 31 22.46 -3.79 4.00 Jul 34 38 25.00 -5.02 5.94 Jul 32 36 23.61 -5.20 5.69 Jul 44 36 27.68 -5.43 7.55 Aug 27 49 26.86 -6.30 6.15 Aug 29 36 23.13 -6.03 4.88 Aug 31 46 26.92 -6.10 6.84 Sep 35 31 25.00 -10.34 12.15 Sep 36 43 28.42 -7.33 8.35 Sep 38 58 34.53 -6.65 6.43 Oct 45 45 31.14 -14.31 11.05 Oct 49 47 33.80 -9.59 10.12 Oct 62 52 39.72 -7.34 10.80 Nov 37 39 29.12 -5.73 6.77 Nov 31 31 22.71 -6.47 7.82 Nov 37 36 26.35 -7.08 9.84 Dec 33 13 17.23 -5.12 5.27 Dec 22 25 19.50 -4.59 6.22 Dec 38 28 26.19 -6.06 7.36

Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max Month + Ext. Days

- Ext.

Days Total % Min Max

Jan 28 21 17.63 -5.58 5.70 Jan 28 30 20.86 -7.08 5.83 Jan 78 74 58.91 -11.43 14.10 Feb 22 23 17.37 -3.27 3.99 Feb 29 33 23.75 -4.34 3.35 Feb 74 68 57.03 -19.98 10.64 Mar 28 31 21.22 -4.41 3.64 Mar 37 32 24.64 -5.83 7.00 Mar 71 71 51.64 -13.82 11.38 Apr 33 26 23.98 -4.66 5.11 Apr 34 25 22.69 -4.04 5.23 Apr 77 48 48.26 -9.32 12.69 May 24 18 16.15 -3.99 4.28 May 25 35 22.06 -4.35 2.80 May 56 68 48.25 -8.67 6.62 Jun 26 22 18.46 -3.81 3.31 Jun 26 30 20.29 -3.89 4.26 Jun 50 62 43.58 -6.61 7.36 Jul 23 24 16.26 -5.59 4.97 Jul 41 36 27.02 -5.55 6.80 Jul 70 48 42.29 -9.44 9.65 Aug 36 28 22.70 -4.39 3.75 Aug 34 43 27.02 -5.25 5.30 Aug 58 58 42.80 -14.07 5.06 Sep 25 38 22.83 -6.26 6.59 Sep 30 44 26.52 -7.68 8.87 Sep 61 51 40.88 -12.03 15.64 Oct 39 43 28.37 -11.72 9.50 Oct 52 43 32.87 -9.47 10.59 Oct 79 59 50.74 -11.87 10.46 Nov 24 25 17.69 -6.62 7.60 Nov 32 34 24.18 -6.59 9.62 Nov 82 74 58.87 -16.17 11.79 Dec 20 15 13.89 -3.54 6.29 Dec 30 22 19.55 -5.75 8.33 Dec 76 38 42.70 -10.37 17.77

Germany

Denmark France Turkey

INDIA China Venezuela

Pakistan Malaysia Brazil

Table 5 - Frequencies of extreme days by calender month

In this table, frequencies of extreme days with respect to month are given for the period 1997-2009, whereas Ext. stands for Extreme and total percentage is calulated by total extreme days in each month divided by total trading

days in each month.

Hongkong Japan Sweden

Figure

Table 1: Determinants of Country’s classification
Table 3 - Descriptive Statistics Jan, 1997 - december 2009
Table 4 - Ranking
Table 5 - Frequencies of extreme days by calender month
+6

References

Related documents

He gives in his thesis more precise and accurate information on the Kyoto mechanisms, which have been really helpful, but mostly on the financial instruments used on the Carbon

46 Konkreta exempel skulle kunna vara främjandeinsatser för affärsänglar/affärsängelnätverk, skapa arenor där aktörer från utbuds- och efterfrågesidan kan mötas eller

The vote share of the Sweden Democrats is collected at the municipality level for the general national election, the county elections and for the municipal elections 5.. The vote

Therefore, the objective of this paper is to contribute to the literature on equity market integration through examining a relationship between the leading stock

The results are consistent at 5% signicance level that OMX Monday return variance is dierent from other weekdays, while Friday and Thursday return variances are considered dierent

The reconstruction showed an overall better performance than many of the previous attempts to use tree-ring data from the region to infer past moisture variability, implying

Results based on Fama-French three-factor model in the study of Penman and Reggiani (2013) show that excess returns for portfolios increases along with a higher B/P ratio, holding

OLS, Quantile and FE Regression on Ln(Return) against Debt/Equity, Lagged Volatility and Interaction This table documents the regression models (as specified in equation 3)