School of Management Blekinge Institute of Technology
Stock Market’s Short-Term Reactions and Volatility as a Result of Political Instability in Serbia
Short-Term Reactions to Political Changes and Crisis - GARCH Model Analysis
Vladimir Stamenovic 670612-P214
Thesis for the Master’s degree in Business Administration Autumn 2011
I would like to thank my mentor and the
BTH staff for their support and patience.
Plans are nothing; planning is everything.
(Dwight D. Eisenhower, 1944)
C O N T E N T S
1. Chapter one – Introduction 1
1.1 Background 1
1.2 Volatility and Market Strategies 2
1.3 Causes of Volatility and Investors’ Behavior 3
1.4 Research Objective 5
1.5 Methodology 5
1.6 Demarcations 6
1.7 Outline of the Study 7
2. Chapter two – Terminology and Literature 8
2.1 Volatility and financial time-series characteristics 8
2.1.1 Volatility clustering 9
2.1.2 Leptokurtosis 10
2.1.3 Heteroscedasticity 10
2.2 Information Impact on Market Behavior 11 2.3 Political instability, country risk and 12 economic outcomes
2.4 Evaluating Political Risk and Country Risk Indicators 13
3. Chapter three – GARCH Analysis 16
3.1 Introduction 16
3.2 Belgrade Stock Exchange 17
3.3 GARCH Analysis 18
4. Chapter four – Events Analysis and Overview 24 Of Significant Events
4.1 Shock Dates Overview 24
4.2 Serbian Stock Market 29
4.3 Events Analysis 32
4.3.1 World Financial Crisis Effect 35
5. Chapter five – Conclusions 38
CHAPTER ONE
INTRODUCTION
“October. This is one of the peculiarly dangerous months to speculate in stocks in. The others are July, January, September, April, November, May, March, June, December, August and February."
(Mark Twain 1894) This chapter provides insight into the research area, beginning with some observations about the eventful months of autumn 2008 on stock markets around the world as perhaps the biggest financial crisis in the modern history developed. Than follows the explanation of the idea for this study and the methodology. At the end of the chapter the outline of the study is presented.
1.1 Background
Stocks worldwide lost 42 percent of their value in 20081, erasing more than 29 trillion USD in value and all of the gains since 2003. The Dow Jones Euro Stoxx 600 Index, a broad measure of European market, finished the year down 46%, the MSCI Asia-Pacific Index fell 43% and in the USA Dow Jones fell 33.8%, while S&P500 lost 38.5%.
If the news from the developed world were bad, it was even worse for many emerging markets, Russia’s Micex loosing 67.2%, the Shanghai Composite Index down 65.4%, and the Sensex 30 in Mumbai down 52.4%, not to mention Iceland where free fall of its OMX Index (ended the year down 94.4%), among other things, brought the country to the verge of bankruptcy.
Safe places for investments were few and scattered – in Bangladesh, main Dhaka Stock Index fell only 7.4% in 2008, similar case was Venezuela, while the odd bird was Ghana where its Databank Stock Index [DSI] ended the year with astonishing result – up 40.6%. There was some light at the end of a tunnel, at least for investors in Ghana.
The markets also became extremely volatile; from September 2008 till January 2009 the S&P500 has moved more than 5 percent in either direction on 18 days, compared to only 17 such days in previous, well, 53 years. October was the most volatile month ever for the
American stock market, yet it was one of the calmer markets in the world. In normal times, markets in developed countries might experience years without even one “big day”2 (none from 2003 through 2007, three in the 1950s, and two in the 1960s). October 2008 saw nine such days. During the Great Depression in the USA, from 1929 through 1933, there were 14 months with at least five such days but the prices then were so low that even small moves looked large on percentage scales. But, even during that period the record number of big days in one month was eight, in September 1932. Furthermore, two of the days in October 2008 were such that market moved more than 9 percent (only nine such days in the history).
Even worse was the autumn of 2008 on other world markets, Russia leading with 17
“big days” for its index, and who knows where it could have ended had the government not temporarily closed the market for trading on October 10. Hong Kong had 13 days of big changes, Argentina and Brazil 12 each, Canada, Britain and Japan 10, and so on.
Diversification did not pay off for investors in 2008 either, casting doubt over this cornerstone of modern investing. The attractiveness of spreading the investments was fading fast in the short term, as strengthening correlations of different asset classes were aggravating losses even in a most diversified portfolios on all markets around the world.
1.2 Volatility and Market Strategies
Critics of market–timing strategies argue that it is hard, if not impossible, for investors to consistently beat the traditional ”buy and hold” approach by jumping back and forth between stocks and cash. But it is also known that markets tend to produce bulk of their gains in just a few explosive sessions. For instance, a hypothetical investor who fully invested in stocks over the last decade, except for the market’s 20 best days during which he opted for cash, would have ended up losing 36 percent instead of gaining 48 percent had he held the stocks for the entire period, 20 best days included3. Even better example is the opposite type of investor who held cash during the 20 worst days of the last decade. He would have gained 242 percent, almost 200 percent more than what the “invest – and – wait” strategy would have brought him. But the third investor, the one who missed both the 20 best and 20 worst market days of the last decade during which he held cash would have made a return of 48 percent, exactly the same as the return from buying stocks and forgetting about them for ten years.
What does this suggest? That there seems to be a rough equivalence of big up and down days during periods of high volatility and moving away from the market during these periods would give one’s portfolio the returns roughly the same as if nothing has been done to
2 Index moving in either way at least 4 percent.
3 All data according to Dow Jones 5000 Wilshire Index, a benchmark representing combined value of all domestic U.S. stocks.
“protect” the investment (or destroy it, depending on investor’s luck, or skills?). But, this would be achieved with much less risk as the volatility would have been much lower than the market’s overall as the third investor’s portfolio shows, being out of the game during the days with extreme volatility. Also, it has been recorded that market sessions with particularly good or bad returns tend to cluster together. By mid September 2008, the VIX Index4 had climbed to its highest level in five years, predicting the likelihood that the following period would be highly volatile and, sure enough, during the next six weeks 5 of the 20 biggest daily losses, as well as 5 of the biggest daily gains of the past decade occurred.
Volatility-avoidance strategies require much closer attention to the market and produce higher costs for the potential investor, but if such successful strategy existed it might attract the attention of investors willing to devote more of their time and energy to better understand the extraordinary market movements, witnessed not only during the 2008. With the increase in the use of new information technologies, computers’ capacities and speed, availability of information on Internet, better modeling software and services it might be worth keeping a closer look on the market behavior on a day–to–day or even intra-day basis, at not so high costs. That, in turn, could benefit investors by protecting their investments in stocks during times of high volatility and unexpected market movements like those seen during the autumn 2008.
1.3 Causes of Volatility and Investors’ Behavior
Investors, big and small alike, have tried to outsmart the market ever since the introduction of stocks. That dream, and search for magic formula which would enable them to
“beat the system” could be compared with the mankind’s dream of eternal life and the search for the Spring of Youth, or with alchemists’ attempts to turn lead into gold. The number of theories, methods and algorithms designed for such purpose matches the number of perpetuum mobile inventions, more or less equally successful.
Yet, a number of theories have proved correct, at least to some extent, in explaining a posteriori market movements, investors’ reactions and relations between them.
There is no much debate among economists on what stock market volatility is (from Latin volare,”to fly”)5, but certainly far less agreement exist on the causes of it. Volatility might be described as a statistical measure of the tendency of a market (or security) to rise or fall sharply within a short period of time. It is typically measured by the standard deviation of the return of an investment, which in turn is a statistical concept that denotes the amount of
4 VIX – Chicago Board Options Exchange Volatility Index
5 Sometime in the 17th century the meaning of volatile became ”birdlike, capable of flying” and that gave rise to
variation or deviation that might be expected. For example, the S&P500 has a standard deviation of about 15%, while a guaranteed investment, such as a bank account, has a standard deviation of zero because its return never varies.
The causes of volatility are harder to define. Some researchers see the causes of volatility in the arrival of new, unanticipated information that could alter expected return of the stock (Engle and Ng, 1993). It means that changes in market volatility would precisely reflect changes in the local or global economic environment. Other claim that volatility is caused by changes in trading volumes, practices or patterns, which are driven by factors such as modifications in macroeconomic policies, shifts in investor risk tolerance and increased uncertainty.6 Also, publication of company news, or changes in policies, launch (or just an announcement) of a new product, a recommendation from a well-known analyst, a popular initial public offering (IPO) or unexpected earnings results, all cause volatility in the markets. Or, one could blame the day traders, short sellers and institutional investors.
Yet another explanation is that investors’ reactions are caused by psychological forces, contradicting the Efficient Market Hypothesis (EMH), which states that market prices are correct and adjust instantly to reflect all information available7. This behavioral approach8 claims that substantial price changes (volatility) result from a collective change of mind by the investing public.
There is no clear definition of the causes of volatility but investors nevertheless need to develop ways to deal with it. Market volatility is inevitable but if and when it is politically induced investors start rethinking their strategies. This is perhaps even more important for small, emerging economies attempting to develop their stock markets often characterized by high volatility. Unlike well developed and well monitored stock markets of advanced economies, markets of less developed countries began to develop rapidly only in the last two decades and are sensitive to changes on the internal political stage, international economic environment and changes in the levels of economic activities. High volatility of emerging stock markets is characterized by frequent and sudden changes in variance, and the periods with high volatility are often associated with important developments in the country, rather than global events.
6 “Measuring Stock Market Volatility in an Emerging Economy”, R. Mala, M. Reddy,
7 Efficient Market Hypothesis is a cornerstone of modern financial theory but often disputed. It states that it is impossible to "beat the market" because stock market efficiency causes existing share prices to always incorporate and reflect all relevant information. According to the EMH, stocks always trade at their fair value on stock exchanges, making it impossible for investors to either purchase undervalued stocks or sell stocks at inflated prices. The only way investor could possibly obtain higher returns is by purchasing riskier investments.
8 A field of finance that explains stock market anomalies through psychology-based theories. Within behavioural finance, it is assumed that the information structure and the characteristics of market participants systematically influence individuals' investment decisions as well as market outcomes. There have been many documented long-term historical phenomena in securities markets that contradict the EMH and which cannot be captured plausibly in models based on perfect investor rationality. Behavioural finance attempts to fill this void.
1.4 Research Objective
Serbia is one of the countries in Eastern and Central Europe that has been experiencing slow economic recovery and development through painful process of transition to market oriented and free trade based economy. Considering its size, central geographical position in the region of Western Balkans, relatively well-developed industrial base and educated workforce, combined with solid natural resources and transport infrastructure, Serbia has many comparative advantages that should have enabled the country to develop faster. That Serbia has not fully utilized all its potentials might be attributed to many factors, one of them being that even after the democratic changes of October 2000 it has been undergoing through constant political crisis with brief periods of political stability and positive business and investment climate.
Although political risk in the country and instability are more commonly associated with the long-term economic development as they affect negatively the growth rate, direct foreign investments, exchange rate, etc, it has also been witnessed that major political events in a country have dramatic overnight effects on its stock market. Some of the changes and political crisis in Serbia during the observation period have caused extreme reactions of the Belgrade Stock Exchange, with daily changes of several percent at a time.
The objective of this study will be to examine whether Serbian stock market shows the characteristics of other emerging markets described in the literature and, if so, try to show whether the unexpected events of political or other nature (either negative and thus causing instability, or positive causing elevated expectations from investors) in Serbia influence its stock market volatility significantly on a short-term, i.e. daily basis.
It will also be checked if any specific category or groups of events/changes on the Serbian internal political arena influence its stock market more severely than other types of changes or events. The starting premise is that political events and changes would have significantly stronger impact on the stock market than changes in other areas such as fiscal, monetary or general economic policy; and that negative events will have more profound influence on the stock market than positive developments
1.5 Methodology
During volatile times, investors start to question their investment strategies and are tempted to pull out of the market and wait on the sidelines until they feel safe to dive back in.
This has been obvious in Serbia during the decline of its stock market which started in 2008, when foreign investors withdrew and stopped their activities almost completely and domestic
capital could not compensate the trade volume, which eventually lead to the fall of both indices far below their initial values. Serbia can serve as an ideal case study because its market is characterized by high volatility, with frequent and large jumps, which should provide a high degree of confidence and accuracy in associating jump market movements with political events. Serbia’s stock market is relatively young and shallow and very perceptible to shock news, while its political situation was fragile during the bigger part of the observation period, with frequent changes and crisis, which could be divided into several categories.
Belgrade Stock Market publishes two indices: BELEXLINE and BELEX15. Analysis will be conducted using GARCH (1, 1)9 model to examine whether BELEX15 daily returns data show the characteristic behavior of financial data series – leptokurtosis, heteroscedasticity and volatility clustering. Jump dates will be identified as the date with significant change of the return, i.e. 4% or more in any direction. Events of the previous day(s) to the day when the extreme jump of the market was observed will be examined to see if any significant changes (in politics, economy, security, regional, etc.) had been recorded. This will be done by checking the corresponding daily issues of three top Serbian daily newspapers10. These events will be categorized according to their nature, using the broad list of widely used indicators used by various agencies providing country risk ratings. The correlation between these categories and changes in market returns (with regard to their value and sign) will be examined in order to check for interdependence of political events and stock market extreme returns. Conclusions about the existence of correlation between unusual market returns and political, or changes of other kind, will be drawn from these results.
1.6 Demarcations
This thesis does not purport to be a political analysis of any kind. It will try to examine if political changes in Serbia have an effect on its stocks market through identifying the jump dates and trying to connect them to the arrival of unexpected news about changes in the political arena.
The statistical analysis applied in this thesis will be limited to GARCH (1, 1) model to show that financial time-series data from the Belgrade Stock Exchange do show the characteristic of financial data series.
9 Courtesy of the Faculty of Mechanical Engineering, Kraljevo, University of Kragujevac, Serbia
10 ”Politika” (www.politika.co.rs), ”Blic” (www.blic.co.rs), ”Danas” (www.danas.co.rs)
1.7 Outline of the Study
This study is organized into five chapters. The first chapter has already been presented, and the outline of the study is shown below.
Chapter 1:
Introduction
Chapter 2: Terminology and Literature Review Chapter 3: GARCH Analysis
Chapter 4: News Analysis and Overview of Significant Events
Chapter 5: Findings and Conclusions
Fig. 1 Outline of the Study
The second chapter provides an overview of literature related to the research area and the explanation of the terminology used in this study. Chapter three presents the results of GARCH (1, 1) analysis, and Chapter four provides the information about the significant events that might have caused unusual behavior of Serbian stock market. Chapter five presents the interpretation of findings and conclusions.
CHAPTER TWO
TERMINOLOGY AND LITERATURE
“The Catterpiller toward the End of Summer waxeth Volatile, and turneth to a Butterflie”
(Francis Bacon, 1626) 2.1 Volatility and financial time-series characteristics
Volatility of returns is an essential parameter in investment decisions and portfolio diversification strategies. Because returns are highly influenced by volatility, most models and risk measurements rely on volatility forecasts as one of critical parameters for financial planning. As stocks prices sometimes move a few percentage points within the period of couple of days, often as a result of news about the political events, new government regulations, company statements, etc, good volatility forecasts become more and more crucial.
Volatility of a stock is a measure of the uncertainty of the returns from that specific stock, expressed as the standard deviation of the return. But, volatility itself is very problematic to measure because it is time-varying and virtually unobservable, thus making measurement difficult, if not impossible. So, the only observable measures of volatility are the actual realized returns.
Econometrics has developed various quantitative or statistical methods used in combination with economic theory to analyze and test economic relationships. These methods use standard statistical models, but some special features of economic data distinguish econometrics from other branches of statistics. Econometrics data are mostly observational, rather than being derived from controlled experiments, so work in econometrics focused on time-series.
It was discovered early that financial time series data show some characteristics distinguishing them from other experimental data, namely: volatility clustering, leptokurtosis and heteroscedasticity.
2.1.1 Volatility clustering
One of the most striking characteristics of financial data series is volatility clustering, which means that extreme returns tend to be followed by more extreme returns or simply - they cluster together. In stock markets, periods of high risk seem to be followed by periods of high risk and the same applies for low risk periods.
This phenomenon is known as volatility clustering. When a shock occurs on a stock market, the volatility is high for the subsequent period. In financial modeling, it has been constantly pointed out that volatility clustering and conditional non-normality induced leptokurtosis are observed in high frequency data. Almost fifty years ago Mandelbrot (1963), followed by Fama (1965), noted that "…large changes tend to be followed by large changes, of either sign, and small changes tend to be followed by small changes."
While returns themselves are uncorrelated, absolute returns or their squares display a positive, significant and slowly decaying autocorrelation function:
corr (|rt|, |rt+ |) > 0, ranging from a few minutes to a several weeks.
This characteristic behavior has long been overlooked, or at least not considered and used in a satisfactory way, until Engle in 1982 described the idea of Autoregressive Conditional Heteroskedasticity (ARCH), inspiring the interest in the research of volatility. His model was further developed by Bollerslev (1986) with Generalized ARCH model (GARCH), which almost immediately became a standard procedure for estimating volatility due to its good performance. Many authors followed with works in the early 90’s, which led to a large and diverse time-series literature on volatility modeling.
This all led to the use of GARCH models in financial forecasting and derivatives pricing.
The ARCH and GARCH models aimed to describe the phenomenon of volatility clustering and related effects such as kurtosis. The main idea behind these two widely-used models was that the variance of returns in the current period depends on either the returns in the earlier periods (ARCH), or the returns and the variance in the earlier periods (GARCH). The variance in each period is hence conditional on the history of the time series. This is a formulation of the intuition that asset volatility tends to revert to some mean value rather than remain constant or move in monotonic fashion over time. The simplest GARCH model still remains the most widely used, due to its adequacy and ease of use in most applications.
Other phenomenon in equity markets is that volatility tends to increase more when the return decreases, than when the return increases. This is referred to as the leverage effect. As a company's equity reduces in value the leverage for the company increases. When the
company’s stock price drops, its equity looses value. The company has become more risky. On the other hand, when the company's stock price increases, the leverage decreases. The risk of equity is therefore dependent on the sign of the return of the stock. (Hull 2000).
2.1.2 Leptokurtosis
Another characteristic of financial time series is called leptokurtosis, which means that returns data tend to have fatter tails compared to standard normal distribution. Kurtosis (from Greek kyrtosis, “convexivity”) is a measure of relative concentration (flatness or peakedness) of data values in the center versus in the tails of a frequency distribution when compared with normal distribution (which has a curtosis of 3). If a distribution has higher kurtosis it will show fatter tails or more extreme values, which is called leptokurtosis (from Greek lepto,
“slim”, referring to the central part of the distribution)11, and distributions with lower kurtosis show fatter middle sections or fewer extreme values, which is than called platykurtosis.
2.1.3 Heteroscedasticity
Heteroscedasticity (from Greek hetero,”different” and skedasis, “dispersion”) is yet another characteristic of financial data series, describing the set of random variables with different variances12. Financial time series data cannot be adequately modeled by normal distribution. Mixture distributions are more likely to capture heteroskedasticity observed in high frequency data.
A number of other models have been developed following the original model of Engle and its generalization by Bolerslev, but it has been pointed out that not one can precisely forecast the market returns. Refinements of ARCH and GARCH models were introduced to deal with the asymmetry and the leptokurtosis found in stock markets time series. Asymmetric volatility means that stock return volatility is asymmetrically related to past returns, with negative unexpected returns having more impact on future volatility than positive unexpected returns.
11 Leptokurtic is an adjective describing a distribution with high kurtosis, meaning the fourth central moment is more than three times the second central moment and this term was used by Bollerslev and Hodrick in 1992 to characterize stock price returns.
12 When using statistical techniques, such as ordinary least squares (OLS), a number of assumptions are typically made. One is that the error term has a constant variance. This will be true if the observations of the error term are assumed to be drawn from identical distributions. Heteroskedasticity is a violation of this assumption.
For example, the error term could vary or increase with each observation, something that is often the case with cross-sectional or time series measurements. Heteroskedasticity is often studied in econometrics, which frequently deals with data exhibiting it. Robert Engle won the 2003 Nobel Memorial Prize for Economics for his early studies on regression analysis in the presence of heteroskedasticity, which led to his formulation of the ARCH.
2.2 Information Impact on Market Behavior
Efficient Market Hypothesis (EMH) states that all the information about the stock price is already incorporated into that stock’s today price. The hypothesis has three levels: weak, semi-weak and strong efficiency. Weak form means that it would be impossible to create a trading strategy that performs better than market, using the information of historical prices of stocks. Therefore, all technical and statistical analysis is fruitless. Semi-weak efficiency means that it is impossible to perform better than average, even when including exogenous variables, such as available public information. Finally, strong efficiency means that private information is also included as exogenous variables, meaning that not even with the inside information would it be possible to perform better than average (LeRoy ,1989).13
According to EMH, it is impossible to predict tomorrow’s price even if we had all the possible information on which that price will be formed. Of course it is never the case, one could never gather and incorporate in a real-time all the information needed to forecast tomorrow’s stock price. In other words, if the theory stands and the information is available, it just cannot be used efficiently to predict tomorrow’s prices.
Stock prices depend on many different factors, but changes in prices are strongly connected to the flow of information. Certain events have strong effects on prices, and when the news of such events reach the market (or sometimes only the rumors about them), prices react more or less immediately. Price of a stock can vary extremely in a short period of time or remain virtually unchanged for a long period, depending on the arrival or absence of relevant information. These include typical information like annual reports, news about changes in macroeconomic policy or economic forecasts, etc. Other factors are trade volume, patterns of trading, actions of major investors/traders on the market, etc. Fluctuations in these parameters sometimes can have an effect on the stock price even without the existence or arrival of actual information that could justify the price change. Bekaret and Campbell (1997) find that in developed markets large changes in prices across securities suggest a greater flow of private information being revealed to the market, and Ross (1989) found that the volatility of prices is directly linked to the rate of information flow in the market.
The effects of information arrival on market volatility were studied in order to describe local factors that have effect on stock market’s return volatility in relation, or with respect to, that country’s specific risk. Many authors found that volatility of emerging markets is driven more by local than by world factors, and it was further examined which local factors have the biggest influence on stock market’s volatility. Credit risk appears to be the dominant proxy for local factors, as well as that national factors and local information have more power in
13 Weak and semi-weak efficiency have their pros and contras, while strong efficiency is highly controversial and
predicting national stock market returns than the world information and global factors, for both developed and emerging markets. A number of studies focused afterwards on discovering these local factors that determine and could help predicting asset returns. The major result of those studies was that the asset returns could be predicted based on the macro-assessments, such as political risk, country-credit worthiness and other macro-economic indicators (Bailey and Chang (1995), Diamonte et al. (1996), Erb et al. (1996a), Patelis (1997), Ferson and Campbell, (1998)).14
2.3 Political instability, country risk and economic outcomes
The relationship between political instability and economic outcomes has been examined since early 80’s. Drazen (2000) described two reasons why political instability affects economic outcomes: it creates uncertainty with respect to future institutions and policymakers, which affects the behavior of individuals and firms; it also changes the incentives of policymakers who either try to extend their term in office or take benefit of the position they already have. Second, the political instability might have a direct effect on productivity, because it disrupts market functioning and economic relations.15
La Porta et al. (1997), Chan and Wei (1996), Willard et al. (1996), established that country’s political, financial end economic risks significantly influence stock volatility and predictability. This is especially the case in emerging markets which are highly volatile, promise higher returns and have low correlation with each other and with global market factors.16 On the other hand, emerging stock markets should be very attractive for investors, not only for the higher returns they offer, but also as an opportunity for portfolio diversification. Significant local factors influence is a result of relative isolation and insulation of these emerging economies from the factors that dictate the behavior of international markets (Bekaret and Campbell, 1995) find that important political events in a country are associated with sudden changes in volatility. Dramatic changes of a political system or changes in government administration affect fiscal and monetary policies, thereby affecting stock markets.
Emergence of capital markets in developing countries has been noted as a major event in financial history. Portfolio flows to emerging countries rose tenfold from 1989 to 1995 (IFC, 1997) and kept rising. Local stock markets responded to the influx of capital and grew
14 “Country risk and stock market volatility, predictability and diversification in the Middle East and Africa”, M.
Kabir Hassan et al
15 “On the Measurement of Political Instability and its Impact on Economic Growth”, R. Jong-A-Pin
16 Goetzamann and Jorion, 1999, found that average dollar return for a sample of emerging markets was 9,1%
compared to 6,9% for developed markets, while the average standard deviation (or volatility) of the dollar return was 34,8% compared with 19,8% for the developed market sample, ibid.
considerably in size. Many developing countries conducted privatization sales through IPO’s on local stock exchanges, which helped to increases the market capitalization. Although privatization is associated with stock market development, the speed and magnitude of market development by far exceed direct impact of successful privatization sales only, which means that the process must have produced significant indirect benefits for local market development. Success in privatization is closely correlated with improvements in perceived political risk and as political risk is resolved over time, equity investment becomes more attractive and leads to further stock market development. Resolution of political risk is strongly associated with stock market capitalization, volume of trading and excess returns as Erb et al. (1996) show that expected returns are related to the magnitude of political risk, i.e.
the lower the level of political risk, the lower are required stock returns.
Financial risk measures contain the most information about future expected returns and political risk contains the least.17 This makes sense when talking about long-term returns, but in the short-run it can be expected that political factors affect the market immediately.
2.4 Evaluating Political Risk and Country Risk Indicators
A clear definition of political risk has not been established. Robock and Simmonds (1973) say that “…political risk in international investment exists when discontinuities occur in the business environment, when they are difficult to anticipate, and when they result from political change”18. The risk can be defined as the probability of politically motivated change that affects the outcome of foreign direct investment. Root (1973) describes a difference between transfer risks (potential restriction on transfer of funds, products, technology and people), operational risks (uncertainty about policies, regulations, governmental administrative procedures which would hinder results and management of operations in the foreign country) and risks on control of capital (discrimination against foreign firms, expropriation, and force local shareholding). It can occur as an explicit event or an ongoing change. Explicit events take the form of legislation or decrees such as expropriations, nationalizations, devaluations, or the form of direct actions such as strikes, boycotts, terrorist acts, and their nature is that they arrive intermittently at discrete intervals and generate an actual losses.
F. Knight (“Risk, Uncertainty and Profit”, 1921), makes a clear distinction between three types of “probability”: (1) a priori probability19, (2) statistical probability20, and (3) estimates.
17 “Political Risk, Economic Risk and Financial Risk”, Erb.C.B., Harvey C.R., Viskanta T.E.
18 “Valuing political risk”, E. Clark, Journal of International Money and Finance
19A priori probability is best described by coin tosses and lottery tickets.
20 Most real-world events are not repeated iterations of one another the way coin tosses and lotteries are; thus
An estimate is different from both types of probability because it is not possible to make a quantitative determination of true probability. If only estimates exist , than there is no
“risk;” rather, there is “uncertainty”, which, according to Knight, is “one of the fundamental facts of life.” The practical difference between risk and uncertainty is that in the former the distribution of the outcome in a group of instances is known (either through calculation a priori or from statistics of past experience), while in the case of uncertainty this is not, because it is impossible to form a group of instances, as the situation dealt with is in a high degree unique.21 Techniques have been developed for computing probabilities and payoffs sufficient to put a price on standard risks. But these techniques are not sufficient to price political risk or politically risky securities) with precision. The techniques help to reduce the uncertainty to some extent, enough that investors are sometimes willing to buy politically risky securities. But the uncertainty is not eliminated; significant uncertainty remains.
Regardless of definition, it is generally accepted that political risk evolves in time as a result of a reaction to countless events on the international, national and personal levels.
Methods for assessing political risk range from the comparative techniques of rating and mapping systems to the analytical techniques of special reports, expert systems and probability determination, to the econometric techniques of model building and analysis.
Widely available assessments of a country’s political, financial and economic risk can serve as good proxies for explaining a majority of variations in expected returns and a volatility of these markets.
Many services measure country risk:
• Bank of America World Information Services
• Control Risks Information Services (CRIS)
• Economist Intelligence Unit (EIU)
• Euromoney
• Institutional Investor (II)
• Standard and Poor’s Rating Group
• Political Risk Services: International Country Risk Guide (ICRG)
• Political Risk Services: Coplin-O’Leary Rating System
• Moody’s Investor Services and others.
Each one of these providers amalgamates a range of qualitative and quantitative information into a single index or rating. Two of perhaps foremost providers of risk ratings are Institutional Investor (II) and International Country Risk Guide (ICRG). Institutional Investor external data, but must derive it from an inductive study of a large group of cases. Life expectancy rates are examples of statistical probabilities.
21 “How investors react to political risk”, Claire A. Hill
credit ratings are based on a survey of leading international bankers who are asked to rate each country on a scale from zero to 100 (100 representing maximum creditworthiness).
Institutional Investor averages these ratings, providing greater weights to respondents with greater worldwide exposure and more sophisticated country analysis systems.
Whenever a survey or expert panel is used to subjectively rate creditworthiness, it is hard to exactly define the parameters taken into account. At any given point in time an expert’s recommendation will be based upon factors the expert feels are relevant. In order to identify the factors that its survey participants have taken into consideration in the past, Institutional Investor asks them to rank the factors that they take into account in preparing country ratings.
The bankers rank factors differently for different groups of countries and rankings have changed over time within country groups.
International Country Risk Guide compiles monthly data on a variety of political, financial and economic risk factors to calculate risk indices in each of these categories as well as a composite risk index. Five financial, thirteen political and six economic factors are used.
Each factor is assigned a numerical rating within a specified range. The specified allowable range for each factor reflects the weight attributed to that factor. A higher score indicates lesser risk. Political risk assessment scores are based on subjective staff analysis of available information. Economic risk assessment scores are based upon objective analysis of quantitative data and financial risk assessment scores are based upon analysis of a mix of quantitative and qualitative information. Calculation of the three individual indices is simply a matter of summing up the point scores for each factor within each risk category. The composite rating is a linear combination of the three individual indices’ point scores. However, the political risk measure (100 points) is given twice the weight of financial and economic risk (50 points each). ICRG, as well as many of the other providers, think of country risk as being composed of two primary components: ability to pay and willingness to pay. Political risk is associated with a willingness to pay while financial and economic risks are associated with an ability to pay.
Among the four indicators, changes in the ICRG political risk ratings display the most pronounced correlations with returns. The ICRG political risk indicator contains the components such as Economic Expectation vs. Reality, Economic Planning Failures, Political Leadership and Law and Order Tradition. Political risk proves to be an important factor for most measures of stock market development, stock returns being strongly related to the changes in the political risk indicator.
CHAPTER THREE
GARCH ANALYSIS
“The only function of economic forecasting is to make astrology look respectable.”
(John Kenneth Galbraith)22
3.1. Introduction
GARCH analysis of Belgrade Stock Exchange index - BELEX15 was conducted using MATLAB’s GARCH Toolbox software.
The GARCH Toolbox models the conditional variance in a time series as a standard GARCH process with Gaussian innovations. It allows a general GARCH (P, Q) form23 with Gaussian innovations for the conditional variance:
¦
¦
P
i
Q
j j t j
i t i
t G A
1 1
2 2
2 N V H
V
(1)
GARCH model requires time series vectors and matrices to be time-tagged series of observations. The daily index values must be converted into a return series using continuous compounding (preferred method for most of continuous-time finance analysis). If the successive price observations at time t and t+1 are denoted as Pt and Pt+1, respectively, continuous compounding transforms a price series {Pt} into a return series {yt} by:
t t
t
t t P P
P
y logP1 log 1log
(2)
The GARCH model also assumes that return series are stationary processes. This may seem limiting but the price-to-return transformation is common and generally guarantees a stable data set for GARCH modeling.
The GARCH default model is the simple conditional mean model with GARCH (1, 1) Gaussian innovations:
t
t C
y H (3)
2 1 1 2
1 / 1 2
t t
t N GV AH
V (4)
22 American economist, Professor of Economics at Harvard 1949-1975. U.S. Ambassador to India 1961-1963.
23 This GARCH model is based on Bollerslev’s original model, and also includes Engle’s original ARCH model as a special case.
In the conditional mean model, Eq. (3), the returns, yt, consist of a simple constant, plus an uncorrelated, white noise disturbance, t. This model is often sufficient to describe the conditional mean in a financial return series.
In the conditional variance model, Eq. (4), the variance forecast, , consists of a constant plus a weighted average of last period's forecast, , and last period's squared disturbance, .
2
Vt
2
1
Vt
2
1
Ht
Although financial return series typically exhibit little correlation, the squared returns often indicate significant correlation and persistence. This implies correlation in the variance process, and is an indication that data can be modeled with GARCH. The default model shown in Eq. (3) and Eq. (4) has several benefits:
- The model requires the estimation of only four parameters (C, , G1, and A1).
Elaborate models often fail to offer real benefits when forecasting.
- The simple GARCH (1, 1) model captures most of the variability in most return series.
Small lags for P and Q are common in empirical applications. Typically, GARCH (1, 1), GARCH (2, 1), or GARCH (1, 2) models are adequate for modeling volatilities even over long sample periods.
3.2 Belgrade Stock Exchange
Belgrade Stock Exchange24 started publishing its BELEXline Index on October 1, 2004, and BELEX15 Index on October 4, 2005. Both indices reached their peak values on May 3, 2007, with BELEXline crossing the 5,000 points mark (5,007.34), while BELEX15 reached 3,304.64 points.
BELEX15 is the leading index of the Belgrade Stock Exchange with the purpose to closely describe movements of the most liquid Serbian shares. It is a free float market capitalization weighted index, consists of shares traded using the continuous trading method, which have satisfied criteria for inclusion into the index basket. The influence of the index components is limited to a maximum of 20% of the total market capitalization of the index on the revision date. BELEX15 is primarily created for improving the investment process, through measuring performances of the most liquid segment of the Serbian capital market and the possibility of comparing potential investment strategies with the index. BELEX15 is calculated and published every BELEX working day in real time.
BELEXline, on the other hand, is positioned as benchmark index of Belgrade Stock Exchange with a purpose to closely describe movements of the broad market and to measure price changes (price index) of shares which are traded on the Belgrade Stock Exchange, which have previously fulfilled criteria for index basket. Index basket consists of common shares traded on the BELEX markets that have fulfilled criteria for the index basket. Index basket can have at least 70 components, and upper limit for the number of components is not determined, for the purpose of better representation of the total market movements.
BELEX15 represents the condition of the strongest and healthiest part of Serbian economy when there are more than 1,800 companies registered on the Belgrade Stock Exchange, but for the big majority of them, it ends there, i.e. there is simply no activity whatsoever with regard to their stocks. 100 stocks that are included in the BELEXline Index make more than 70% of the total trading volume.
Having all that in mind, GARCH analysis was conducted in order to examine the time series data of BELEX15 only. Time series start with the date of its initial publishing (index start value was set at 1000 points) and end on February 28, 2011, thus providing 1360 observations of BELEX15 daily values, making the time series large enough for GARCH analysis.
3.3 GARCH Analysis
As can be noticed on Figure 1 - BELEX15 Daily Values, the graph shows a characteristic unsustainable parabolic rise almost reaching a vertical asymptote.
Figure 1 – BELEX15 Daily Values
MATLAB’s GARCH Toolbox estimates the parameters needed to model the series in several steps:
1. A pre-estimation analysis to determine if the data is heteroskedastic and can be modeled using GARCH;
2. Estimation of the parameters for the default model;
3. A post-estimation analysis to confirm that the chosen model explains the heteroscedasticity present in the data.
In the first step, the daily values of an index are converted into a return series (because GARCH modeling requires a return series), checked for correlation and quantified.
As seen at Figure 2 – BELEX15 Daily Returns, the presence of volatility clustering is clearly present in the raw return series, especially during these periods: mid March to mid May 2007, February to May 2008, mid September to end October 2008, and from mid November 2008 until the end March 2009.
The presence of correlation in the raw return series can be tested by autocorrelation function (ACF) and partial-autocorrelation (PACF) function, which serve as preliminary identification tools and provide some indication of the broad correlation characteristics of the returns.
Figure 2 – BELEX15 Daily Returns
The ACF and PACF graphs of the raw returns of BELEX15 Index show the presence of correlation. More, the ACF of the squared returns indicates significant correlation:
Figure 3 – BELEX15 Sample Autocorrelation Function (ACF) of Squared Returns
Clearly, variance process of BELEX15 Index exhibits some correlation. The ACF of the Squared Returns dies out slowly, indicating the possibility of a variance process close to being non-stationary.
These descriptive checks for correlation were quantified using formal hypothesis tests, such as the Ljung-Box-Pierce Q-test and Engle's ARCH test, both of which confirmed presence of heteroscedasticity and indicated that GARCH modeling is appropriate.
Table 1 – GARCH (1, 1) parameter estimation of BELEX15 returns Mean: ARMAX (0, 0, 0); Variance: GARCH (1, 1)
Conditional Probability Distribution: Gaussian Number of Model Parameters Estimated: 4
Parameter Value Standard Error T Statistic
C 0.00041191 0.00025625 1.6075
K 1.2393 e-005 1.8658 e-006 6.6420
GARCH (1) 0.57881 0.02159 26.8090
ARCH(1) 0.42119 0.029439 14.3074
The toolbox uses the estimation function garchfit to estimate the model parameters, assuming the default GARCH (1, 1) model. After the estimation is completed, the function garchdisp displays the parameter estimates and their standard errors.
Finally, by substituting these values in the definition of the default model, Eq. (3) and (4), the estimation process implies that the constant conditional mean/GARCH (1, 1) conditional variance model that best fits the observed data of BELEX15 Index would be:
t
yt 0.00041191H (5)
2 1 2
1 / -005
2 1.2393e 0.57881 t 0.42119 t
t V H
V (6)
In addition to the parameter estimates and standard errors, the toolbox provides the residuals (innovations), and conditional standard deviations (sigma), which allows inspecting the relationship between the innovations derived from the fitted model, the corresponding conditional standard deviations, and the actual observed returns.
The fitted innovations also exhibit volatility clustering (Figure 4) but the plot of standardized innovations (innovations divided by their conditional standard deviation) shows that clustering is less prominent (Figure 5), i.e. the changes seem to be more evenly spread throughout the observation period:
Figure 4 - Comparison of Innovations, Conditional Standard Deviations and Observed Returns of BELEX15 Index
Figure 5 – Standardized Innovations of BELEX15 Index
Finally, the ACF of the squared standardized innovations plot (Figure 6) shows much less correlation than the ACF of the squared returns prior to fitting the default model:
Figure 6 – Sample Autocorrelation Function of Standardized Innovations
The analysis shows that the chosen default model does explain in a sufficiently satisfactory way the heteroscedasticity in the raw returns of BELEX15 Index, but also suggests that more elaborate model might yield even better results. This suggestion was tested through quantitative checks for correlation (Ljung-Box-Pierce Q-test and Engle's ARCH test) in the fitted innovations, with results implying that the explanatory power of the default model might not be sufficient to precisely describe the BELEX15 financial data series. The presentation of the GARCH statistical analysis indicates that GARCH (1, 1) model efficiently explains the correlation and heteroscedasticity in the time series of BELEX15 Index. Yet, a more elaborate model derived from the original ARCH and GARCH ideas of Engle and Bolerslev might better describe the behavior of financial data series of BELEX15 Index. Similar analysis conducted for another index, BELEXline, brings the same results. However, as the main goal of this thesis was not a GARCH analysis per se, it was conducted to examine whether the financial data series of Belgrade Stock Exchange indices exhibit the characteristic behavior of financial data series present in other markets and derivatives, i.e. heteroscedasticity, volatility clustering and leptokurtosis, which was confirmed.
CHAPTER FOUR
EVENTS ANALYSYS AND OVERWIEV OF SIGNIFICANT EVENTS
“The herd instinct among forecasters makes
sheep look like independent thinkers.”
(Edgar R. Fiedler)25
4.1 Shock Dates Overview
GARCH analysis conducted in previous chapter confirmed that both BELEX15 and BELEXline indices of BSE exhibit characteristic behavior of financial time series. A total of 44 jump (or shock) dates were identified (days when BELEX15 Index moved 4% or more in any direction) during the observation period. Out of these, 20 were the days with movements in the interval between 4-5%, 9 days between 5-6%, 5 days with 6-7% movements, 3 days in the interval 7-8%, 2 days in the interval 8-9%, 1 day with the movement between 9-10%, 2 days with the market movement between 10-11% and, finally, 2 days with the earthquake type of change of more than 12%.
24 days have seen BELEX15 Index move in negative direction, with 20 days in the green. BELEXline followed the direction of BELEX15 Index, with less severe changes.
Accumulated change during only these 44 shock days amounted to -25.508%26. But, during the same period (between October 04, 2005, and December 01, 2009, i. e. the last shock date), all trading days included, BELEX15 lost 28.405%, faring only slightly worse. This confirms what was already stated in the Chapter One - that gains\losses of one’s portfolio will more-less be the same whether trying to be proactive and stay out of the market during shock days or passively watch the market and not act at all.
Table 2– BELEX15 returns change over the observation period
BELEX15 Start
04.10.2005.
On a Last Shock Date
01.12.2009. Change
Index Value 1000 715.95 - 28.405%
25 Edgar Russell Fiedler, was an American economist who served as Vice President of The Conference Board and as Assistant Secretary of the Treasury for Economic Policy (1971 - 1975) during the presidencies of Richard Nixon and Gerald Ford.
26 Taking into account only changes occurring on shock days.
These shocks should have been caused by unexpected political events in Serbia the day(s) before, which, in turn, would support the premise of this paper that Serbian stock market is responsive to and more susceptible to the negative political changes. As described in Chapter One, these events were identified through analysis of news and reports of the leading Serbian daily newspapers.
Table 3– Overview of significant events Date Event(s) BELEX15/BELEXLINE change
Proxy sign/Category 29/03/2007 Government’s decision on prolongation of temporary financing (budget) announced; coalition efforts on government composition continue 4.84/3.00 +1/Int.Pol./Econ. 08/05/2007
T.Nikolic from ultranationalist Serbian Radical Party elected Parliament Speaker with support from DSS; one of the pro-democratic parties walks out of government talks; deadlock on forming a government
-5.40/-3.12 -1/Int.Pol. 09/05/2007
Crisis continues as deadline for government formation nears; signing of the Agreement on Visa Liberalization postponed; T. Nikolic hints on the possibility of imposing the state of emergency
-6.75/-5.38 -1/Int.Pol./Eur. 11/05/2007 DSS, G17 and DS reach an agreement on forming the government; new Parliament Speaker; EU and USA welcome the news on coalition agreement10.83/6.57 -1/Int.Pol/ 20/11/2007
Unsuccessful sale of state-owned insurance company DDOR “Novi Sad”; DSS considers any possible recognition of Kosovo’s independence a hostile act; EU announces that “it will make decisions if no agreement on the future status of Kosovo is reached”
-5.07/-3.1 -1/Econ./For.Pol. 31/01/2008 EU announces the decision to offer signing of the Political Agreement with Serbia on February 07; calls for pro-European votes on presidential elections; USA support EU and Serbia agreement 4.18/2.5 1/Eur./For.Pol. 04/02/2008 B. Tadic wins second round of presidential elections; EU welcomes the victory; high trading volume on BSE 6.14/4.42 1/Int.Pol./Eur. 05/02/2008 Prime Minister V. Kostunica hesitates and stalls to call the Government session to approve signing of the agreement with the EU -4.26/-2.09 -1/Int.Pol./Eur. 06/02/2008 V. Kostunica calls Parliament to reject the agreement with the EU; Brussels says “if you reject the offer it is no longer on the table” -4.34/-2.73 -1/Int.Pol./Eur. 26
06/03/2008 Radical Party pushes for a Parliament Resolution on future relations with EU; DSS supports the motion; Government split - de facto coalition breakdown; Parliament breaks the session -4.97/-2.57 -1/Int.Pol./Eur. 10/03/2008 Kostunica announces the Government is not functioning and that he will “return a mandate to the people”; Government requests President Tadic to call early elections; new coalition talks already started; EU say Serbia chooses between Europe and self-isolation
-4.65/-2.66 -1/Int.Pol. 12/03/2008 DS and G17 want new government without DSS; Kostunica announces the agreement is reached on forming the government with SRS-7.68/-3.56 -1/Int.Pol. 13/03/2008 President Tadic calls for elections for May 11; possible DSS- SRS-NS coalition, DS still without partner -4.91/-3.14 -1/Int.Pol. 17/03/2008 Announcement that DS and G17 have formed a coalition 4.14/2.10 1/Int.Pol. 09/05/2008
Signing of Interim Agreement on Trade and Trade-Related Matters between Serbia and EU on the eve of elections brings support for pro-European coalition; Russia supports SAA; deal with the EU brings new votes for DS as a sign of support
4.35/3.04 1/For.Pol./Econ. 12/05/2008 Easy victory of democratic forces 12.16/9.87 1/Int.Pol. 13/05/2008 Triumph, but how to form the government; SPS to decide on the new Government -4.25/-2.67 -1/Int.Pol. 16/09/2008 SRS’ A. Vucic resigns and confirms that he will join Nikolic; MPs from DSS and NS walk out of the Parliament -6.82/-4.70 -1/Int.Pol. 18/09/2008 SRS motions to impeach Parliament Speaker; Administrative body in the Parliament to decide on Nikolic’s destiny -4.94/-3.59 -1/Int.Pol. 19/09/2008 Ruling coalition majority keeps the mandates of Nikolic’s MPs4.08/2.30 1/Int.Pol. 29/09/2008 FIAT deal will bring new jobs and money to Serbia; SRS has conditions for the continuation of the Parliamentary session; -4.42/-3.28 1/Econ. -1/Int.Pol. 06/10/2008 SRS blocks the adoption oh necessary Schengen laws; rebalance of the budget adopted; EULEX mission in Kosovo problematic for the Parliament -8.44/-5.65 -1/For.Pol./Eur. 1/Econ. 07/10/2008 Serbia in the EU only if it accepts EULEX; new Parliamentary procedures removes the blockade of Parliament’s work -10.86/-6.97 -1/Eur. 1/Int.Pol. 27
28
08/10/2008 Serbia will protect the savings and the currency -9.64/-6.40-1/Econ. 14/10/2008 EULEX acceptance not precondition for EU integration continuation; new loan of 400 mil. USD for the Corridor 10; “FIAT Serbia” company established 12.03/7.31 1/Eur./Econ. 16/10/2008 Escalation of world financial crisis; Serbia asks IMF for help; inflation 10.9% -8.18/-5.44 -1/Econ. 24/10/2008 Dinar falls further to Euro -6.35/-4.51 -1/Econ. 27/10/2008 Negative EU report on approximation to EU; “Metals Bank” in crisis -7.00/-4.51 -1/Eur./Econ. 19/11/2008 Dinar continues to loose value against Euro; Parliament blocked by huge number of amendments -4.13/-2.65 -1/Econ./Int.Pol. 20/11/2008 Opposition parties block Parliament -4.48/-2.51 -1/Int.Pol. 15/12/2008 EU still against unblocking of the Interim Agreement; dinar looses value; 2009 budget proposal -4.50/-2.89 -1/EUr./Econ. 25/12/2008 SRS blocks budget voting, but no government reconstruction; Russia buys Serbian oil company NIS 6.72/3.97 1/Econ./Int.Pol. 29/12/2008 Good arrangement with Russia, stable Government; possible EU support to Serbian budget of 400 mil. USD 5.36/3.45 1/Econ. 23/01/2009 Cheaper credits, no price hikes for gas and electricity; President Tadic says no time for elections; Government announces 1 Bill. Euro stimulant package for economy5.32/2.73 1/Econ./Int.Pol. 19/03/2009 Government to discuss Vojvodina’s prerogatives; extra wealth tax introduced; 4.32/217 ---- 04/05/2009 “TELECOM” Serbia announces 3.9 bln. Dinars in dividends to be paid;Serbia to get on the “white Schengen list” by the end of the year; new IMF loan for Serbia5.05/2.54 1/Econ./Eur. 05/05/2009 Strong sales of Serbian banks’ stocks on BSE 4.44/2.60 ---- 06/05/2009 40 bln. Dinars in new stimulant Government measures for economy 4.68/3.67 ---- 07/05/2009 EU grant of 100 mil. EURO 7.76/4.80 ---- 16/06/2009 EU decides to grant visa-free regime to Serbia; possible unblocking of the Interim Agreement with EU; GDP drop 5% -4.95/-3.52 ---- 26/08/2009 IMF suggests VAT increase, NIS stocks to be distributed to citizens5.05/3.70---- 02/10/2009 Risk for investing in Serbia remains high (DiB Agency) -5.49/-4.46---- 06/10/2009 Support for visa abolition, and European integrations voiced 5.47/4.27 ---- 01/12/2009 Government adopts budget for 2010 5.10/3.50----
4.2 Serbian Stock Market
Shallow Serbian market acted during the first half of 2007 as under a constant adrenaline infusion, due to the introduction of new stocks, increased interest by foreign investors, emergence of domestic institutional investors and newly discovered (but shy) interest for stock market by Serbian citizens. This inflow of capital was driving the demand and overheating the stocks’ prices.
Foreign investors recognized the potential of Serbian securities and market and were initially investing in various sectors, from banking to food industry, agriculture, construction, transportation, which contributed to the explosive growth of total market capitalization.
Serbian investors, on other hand, were inexperienced in investment strategies, as domestic investment and pension funds were just beginning to function in Serbia, having been formed and started working at the beginning of 2007 (after necessary legal regulations were adopted), which in turn proved to be a very bad timing as the market was in strong upward movement in that period. Furthermore, Serbian citizens still preferred bank account savings (which in January 2009, amounted to more than 4 billion Euro) and were not a significant investment factor. 27 The severe market downturn after near-to-vertical growth was the first (and painful) stock markets experience for domestic investors and the “Investments 101” lesson - Buy when market is going down, sell when market is rising, could not be exercised.
Picture 1 – Unsustainable rise of BELEX15 Index
Cooling of the Belgrade Stock Market started already in the 2007, preceding the arrival of the world crisis to Serbia but in 2008, as a result of a crisis, it was clearly visible that foreign capital was fleeing the market. Foreign capital participated with around 50% in market trading volume, but the investors were not very exposed in general terms, considering total value of their investments. When the “decline and fall” of Serbian market became evidently irreversible
27 Although domestic market was in process of intensive development, Serbian citizens still needed to learn about possibilities of investing in securities. Capital market was relatively new opportunity, and some citizens didn't
foreign investors started leaving their positions on the market even if they had to bear losses (switching to more profitable investments, repairing their portfolios compositions, returning to their domicile markets, etc), and their actions influenced the behavior of insecure domestic investors. Under these conditions domestic capital available for investments could not sustain the required level of market liquidity any more, so Belgrade Stock Exchange lost what it needed the most - new market players and investments, necessary for development and growth. Accordingly, the market being still small and underdeveloped and without the Serbian biggest companies’ stocks, and state/municipal and corporate bonds experienced a very painful correction.
BELEX15 Index lost around 30% of its value during the first three months of 2008, falling back to its December 2006 level. Once the avalanche starts, it can not be stopped, and on the Serbian market the initial snowball was composed of accumulated effect of negative expectations, additionally confusing investors already weakened by the global crisis. Local institutional investors, investment and pension funds (still very fragile and in their infancy), had to sell parts of portfolios in order to maintain liquidity. The consequence of ”domino effect” forced local investors, influenced by the general instability (largely emphasized by the media) to sell their stocks.
When the world crisis finally arrived to Serbia infecting first the financial and later spreading to the real sector of the economy, all the factors that characterized Serbian underdeveloped stock market were still present: low offer of high quality financial instruments, bad communication between companies and investors, legal regulations that limited companies’ market exposure and limited emissions of more financially attractive papers and instruments, low domestic demand for stocks, etc. While there were no such financial instruments on the Serbian market as the ones that have caused the world crisis, negative information from the world markets, financial sector crisis and foreign investors’
actions and expectations, all brought insecurity and hesitation to Belgrade Stock Exchange and further aggravated the situation, creating a psychological effect of “market exit”, which additionally inflamed the downturn trend and caused a new wave of investors’ withdrawal.
The crisis had not just one, but perhaps three causes: 1) external factor: 2) internal weakness of Serbian market (low liquidity, low quality and diversity of stocks offered, low level of corporate management, lack of effort by companies to attract the investors, non-harmonized laws not stimulating the development of stock market), and 3) frequent political tensions, which kept the political risk factor at a constantly high level.
Investors would usually wait for the first early statements from companies to provide a signal for buying or selling stocks, or signal the possible change of a stock value. But it seemed that investors in Serbia preferred to make decisions based on the influence of daily news from
