Stock Market Efficiency in Kazakhstan and Russia
Disclaimer: This work has been submitted by a student. This is not an example of the work written by our professional academic writers. You can view samples of our professional work here.
Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of UK Essays.
Published: Wed, 14 Mar 2018
Efficient Market Hypothesis (EMH) continues to be controversial and debatable even in the 21st century. In the financial world many researchers have garnered ever greater empirical evidence of inefficiency, greatly discrediting EMH; however it plays a significant role in modern finance and therefore gives us an incentive for future research. There is a wide consensus that on the early stage stock markets in transition countries are obviously not weak-form efficient, due to the lack of information, irrational behaviour of participants and an overall lack of financial development. However, regulations, market structure and learning processes could create the framework to become relatively efficient markets.
The emerging stock markets in transition economies have attracted increasing attention from researchers. The vast majority of time-varying market efficiency research to date have been focused on European transition economies and no similar study on has been performed for post soviet countries except Russia.
The purpose of the current paper is to analyze financial market of one of the post soviet countries Kazakhstan and conduct comparative analysis of evolving market efficiency in a weak-form sense with Russian. The main objective is to identify which particular stock market becomes more efficient over time.
We used Kazakhstan Stock Exchange (KASE) and Russian Trading System (RTS) stock index prices at daily frequency covering time series from 1st of July 2005 to 1st of July 2010. After the detailed analysis of the structure and regulations of KASE and RTS the analysis is carried out at two stages. First, we estimate an AR (1) model and a GARCH model for the stock indexes. Then we estimate our AR (1) model with time varying coefficients and GARCH type errors. Finally we find varying levels of efficiency and varying speeds of movements toward relevant efficiency.
The main objective of this paper is to test whether the Russian market, the most important amongst the so-called transition economies markets, has evolved towards some degree of efficiency since its foundation.
The EMH has been the central proposition of finance since the early 1970s and is one of the most contoversial and well-studied propositions in all the social sciences.
Regardless of whether or not one believes that markets are efficient, or even whether they are efficient, the efficient market hypothesis is almost certainly the right place to start when thinking about asset price formation. One can then consider relative efficiency.
It is interesting to analyze Russian financial markets for several reasons. First, financial market behavior in Russia can be different than in other emerging markets due to historical, cultural, and institutional factors. Second, Russian markets may offer better diversification benefits (Rockinger and Urga, 2000). Third, since the early 1990s, Russian policy makers have implemented major economic and financial reforms, resulting in the emergence of new financial instruments. A related question in this respect is whether investors in this market react to “news” in a similar fashion as those in advanced market economies. Fourth, because Russia is rich in energy resources, oil price shocks may have destabilizing effects on domestic financial markets. Fifth, a significant drawback with respect to financial market liberalization took place in 1998, and it is interesting to analyze the consequences of this development for the internationalization of Russian financial markets.
PART I: LITERATURE REVIEW
1.1 MARKET EFFICIENCY HYPOTHESIS AND CONCEPT OF EFFICIENY
‘A market in which prices always “fully reflect” available information is called “efficient.”-‘ [Fama (1970)]
The EMH puzzled many researches and economists from 20th century, many research papers have been written to provide empirical evidence either for or against EMH in order to assess its suitability for estimation and determination of stock price.
The EHM was first expressed by Bachelier (1990) in his thesis “The Theory of Speculation” being the first to model stochastic process (Brownian motion). The EMH was developed by Eugene Fama in 1960s. Fama (1965) first defined “efficient market” as a market which adjusts rapidly to new information. In his empirical analysis he concluded that stock market prices follow random walk and series of price changes have no memory. First formal economical argument against EHM was provided by Samuelson (1965). In his article “Proof that properly anticipated prices fluctuate randomly” inspired by Bachelier’s (1990) work, he focused on the concept of stochastic process rather than a random walk.
Roberts (1967) coined the term “efficient markets hypothesis” and made the distinction between weak and strong form tests, these distinctions were developed further by Fama (1970) and became the classic classification of there level of efficiency. In accordance to revolutionary framework of Fama (1970) there are three forms of efficiency:
Weak form efficiency – security prices fully reflect the information contains only a past price movement which means that the future price movements are unpredictable and makes it impossible to trade on profit on the basis of the historic price information.
Semi-strong form efficiency – security prices reflect the information contains all past price movements and all publicly available information which means it is impossible to make superior return based on using publicly available sources because information is already incorporated into security prices.
Strong form efficiency – securities reflect all relevant information regarding its publicly available or not, which means none of the investors will not be able to make superior returns.
Fama (1970) also suggests that the share prices are very efficient and at any given time reflect all available information holding the assumption that there is no transaction cost and all market participants equally interpret the effect of the new information on the current price. However, we know that this assumption is unrealistic in the real financial world. It was argued by Grossman et al (1980) that it is impossible for a market to be perfectly informationally efficient. Information is costly and prices cannot perfectly reflect the information due to the fact that trader who obtained that information simply cannot make money on trading it. Lo et al (1997) shared the same opinion in favor of the notion of relative efficiency by measuring the efficiency of one market against another.
All this controversial results coming from research studies have driven researches to apply various methods and statistical techniques to obtain a more conclusive idea about validity of EMH. Many of the researches reproached their views on EHM AND even Fama (1991) focused on testing informational efficiency and modified his views suggested earlier in Fama (1970). However many studies were conducted to define EHM, for instance, Malkeil (1992), Fama (1998) and Timmermann et al (2004) discuss the EMH from the perspective of a modern forecasting approach. Malkiel (2004) in his book “A random walk down wall street” believes that EMH is almost true. However, Shiller (2005) in his best selling book “Irrational Exuberance” is very doubtful about EMH. It is hard to conclude the dispute about the EMH. However, in accordance to Campbell et al. (1997) it is more reasonable to measure relative market inefficiency rather than efficiency.
1.2 EVOLVING EFFICENCY IN EMERGING MARKETS
“…the Efficient Markets Hypothesis, by itself, is not a well-defined and empirically refutable hypothesis. To make it operational, one must specify additional structure, e.g., investors’ preferences, information structure, business conditions, etc. But then a test of the Efficient Markets Hypothesis becomes a test of several auxiliary hypotheses as well, and a rejection of such a joint hypothesis tells us little about which aspect of the joint hypothesis is inconsistent with the data.” – (Lo et al., 1999)
After all the arguments for and against EMH economists reached the consensus that all the markets even in developed economy considered to be inefficient. However there is an increasing popularly in measuring and analyzing the relative efficiency or informational efficiency between the developed and emerging stock markets. The term changing or evolving efficiency was introduced.
As it is describe in Emerson (1996) there are three types of efficiency: information, operational and allocation efficiency. The concept of informational efficiency can be defined as prediction of how quickly prices will respond to new information. The information is available at a very low cost and it is assessable to all market participants those markets considered to be as informationally efficient. The operational efficiency suggests that transactions cost are at a very low level. In developed financial markets with a high level of completion between financial intermediaries is true. The allocative efficiency suggests that the price with the same of risk will offer the same expected return. It was concluded that it is much easier to test measure informational efficiency of the market rather than operational and allocative efficiency.
Vast majority of research papers are focused on how those emerging markets become relative informationally efficient and which of the particular markets evolve faster and the reasons. The investigations of evolving market efficiency have so far been mainly focused on European transition economies and Russia. There is no any similar research on Kazakhstan stock market.
Cornelius (1994) described in the early days of a new market, it is obvious that market participants are unlikely to act in accord with the efficient market paradigm. Emerging markets attracted researches due to unique opportunity to track changes and analyze how markets learn over time and move towards efficiency. Number of studies was conducted to analyse this phenomenon, as it was noticed above mainly focused on developed markets and emerging markets in Europe and Asia. Previous empirical studies on evolving stock markets efficiency in European emerging markets, for instance, Emerson et al (1996) tested four Bulgarian shares whether and how the markets moved towards informational efficiency covering time period from first week of 1994 to first week of 1996. To find varying levels of efficiency and varying speeds of movement towards efficiency with given sample time varying autoregressive model AR (2) and a generalized autoregressive conditional heteroskedasticity model GARCH-M (1,1) and Kalman filter were used. Zalewska-Mitura et al (1999) carried out analysis by applying empirical data from London and Budapest Stock Exchanges, and showed the changing levels of inefficiency in developing markets as compared to developed markets. To investigate evolving market efficiency the classical test for autocorrelation of returns were extended by combining a multi-factor model with time varying coefficients and the GARCH-M approach. Study confirmed that application of Monte Carlo simulation verified perfect series from the London Stock Exchange. While Budapest Stock Exchange it was confirmed the usefulness for an investigation of the first stages of a market performance. Rockinger et al (2000) considered aggregate stock indexes of the Czech Republic, Hungary, Poland and Russia for the frequency running from 1994 to 1997. It was concluded Russian shows signs of ongoing convergence towards efficiency, while Poland and the Czech Republic were not improving. The Russian stock market was tested for evolving market efficiency in the research paper by Hall et al (2002). By using a time varying parameter model with GARCH in mean effects they tested two indexes of the Russian Stock Market and concluded that RTS index represented by the most liquid stocks were initially inefficient and it took around two and half years to become more efficient than ASPGEN index. However, both of the markets were moving towards efficiency. The most liquid tree shares were selected for further analysis; in this case as well there was mixed evidence over the period in particular with the larges oil holding Lukoil was being more efficient than other selected stocks.
In the majority of the listed above research papers for instance, Hall et al (2002) and Emerson et al (1996) the weak form efficiency for were tested by carrying out simple regression:
rt = β0+∑ βi rt-i + εt (1)
rt rate of return at time t
Weak form of efficiency implies that βi=0, i>0
Given hypothesis usually to be tested by Ordinary Least Squares (OLS) or Generalized Method of Moments (GMM). It was concluded by Emerson et al (1996) and Hall et al (2002) that this is not sensible approach to test emerging markets of Bulgaria and Russia. The main reason is that both of the markets are considered to be inefficient and given regression would test efficiency over the whole period of time. In order to measure inefficiency and measure how quickly emerging markets move towards efficiency the equation was modified as follows,
rt = β0t+∑ βit rt-i + εt (2)
so that allows for the changing parameters have time subscripts to vary over time. This was also concluded in Rockinger and Urga (200).
In light of these studies, current research progresses by initially applying the same regression which is particularly relevant to our sample and carries on with further tests which will be described in the second part.
The recent research was conducted again on US stock markets by Ito et al (2009) realized an empirical study measuring a gradually time varying structure of market inefficiency of S&P 500 stock index form 1955 to 2006. Their method was carried out in two steps, on the first place time varying structure of autocorrelations of the returns were checked on the bases of the Moving Window method, the second step estimated the time varying AR (1) coefficients by using stage space model. The paper confirmed that the US stock markets were the most inefficient during the late 1980s and has become most efficient at around 2000 in the last half a century.
PART II: STRATEGY AND FORMULATION OF A MODEL
The aim of this paper is to investigate financial markets of two post soviet countries Kazakhstan and Russian, and conduct comparative analysis of evolving market efficiency in a weak-form sense. The main objective is to identify which particular stock market becomes more efficient over time.
The strategy of analysis of comparative evolving market efficiency in a weak form since will be based on empirical studies of the Emerson et al (1996), Hall et al (2002) and Zaleshka- Metura et al (1998) which is reasonable due to the fact that the not all emerging financial markets behave similar to each other. Financial Market of Russia were previously assessed in those research papers. It is also reasonable to assume that Kazakhstan and Russian stock markets are pretty much the same. Both of the countries were part of the post soviet union sharing the same history and culture. Both of the financial markets started to develop and coexist separately from each other at the same time; both of the countries’ economies are developing fast due to natural recourses as oil and gas and other mineral recourses.
In order to conduct comparative analysis of two markets, we need to gain understanding how both of the stock market function, what is their structure, regulations and policies. By briefly providing summary of vital information we will proceed with quantitative analysis.
In order to determine which particular market moves towards efficiency first of all, we estimate an AR (1) model and a GARCH model for the stock indexes. GARCH type models takes into account the possibility that the error process for a stochastic series might not prove to have a full set of NIID (normal, identical and independent distribution) properties as it was stated in Emerson et al (1996), Hall et al (2002). In addition, we estimate our AR (1) model with time varying coefficients and GARCH type errors. By selecting the most liquid and reprehensive share from the list of shares which comprises KASE and RTS index, we will carry the same test to compare evolving efficiency in between individual shares from each stock exchange.
The rest of the paper is organised as follows, in the next part, we describe the data and the source of it. By providing brief overview on Russian and Kazakh stocks markets and its origination. We will precede with empirical framework in part four where the whole model is presented and explained, while the empirical results are given and discussed in Section V. Finally, Section VI and VII will summarize our finding and present the limitations of the analysis.
In our test KASE and RTS stock index prices were obtained at daily frequency covering time series from 1st of July 2005 to 1st of July 2010. The time series data used to be non stationarity due to this reason we will conduct test for stationarity and unit root and all the data from will be logarithmic.
PART III: DATA SOURCE AND DESCRIPTION
The data tested below are daily observations on two aggregate indexes of closing prices on the Kazakhstan Stock Exchanges and Russian Trading System stock exchange Index. Both of the indexes are price weighted and comprise series of all listed stocks. All the information and on stock index prices were obtained from publicly available official recourses. In current research we covered time series starting from 1st of July 2005 to 1st of July 2010 with a total of 1242 observations for both of the stock exchange markets. The sample size and data used in this paper will provide new results evolution of Russian stock market since the empirical analysis of Hall et al (2002) and Rockinger et al (2000) and will be the first for Kazakhstan Stock exchange.
Before modeling the data we need to carry out simple descriptive statistics on the empirical data. The specific problems that often encounter in dealing with time series are as follows Defusco (2007): the residuals are correlated instead of being uncorrelated, the mean and/or covariance of the time series changes over time. Due to this fact before using our data we need to ensure that the data is stationary and there is no autocorrelation. By plotting the returns Figure 1 and Figure 2 we can see that the data is non stationary and autocorrelation is present during the sample period for both RTS and KASE markets. As it is well know that time series data are subject to
Figure 1  Figure 2
By generating logarithms and first difference, we transformed the raw data into covariance stationary time series. We applied the Augmented Dickey Fuller (ADF) test to check that data doesn’t contain a unit root. The summary of our test on a unit root is illustrated in table 1.
Since the time series shown in Figure 1 and 2 may contain a unit root, we apply. We assume a model with time trend and a constant and use Schwarz Bayesian Information Criterion (SBIC) as an order selection criterion. Lag 0 is chosen and the test statistics is computed to be −2.70. The corresponding 5% critical value is −3.42. Thus we cannot reject the null that the data contain a unit root. This fact supports our assumption that the AR coefficients in a time varying AR model follow a random walk process.
The simplest method to measure the autocorrelations of stock returns is to apply the AR model to the stock returns data, since the Yule–Walker equation assures sample autocorrelation functions correspond to coefficients of the AR model with each other uniquely.
It can be notices that see that they exhibit quite a few spikes that are far above the average amplitude. These spikes suggest that the two rates of returns possess significant volatility clustering and unequal scatter. But a visual inspection of the time plot of a return series is not sufficient for us to judge on its statistical properties. A way to check whether or not a return series as an error process is an i.i.d (0, s2) sequence would be to look at its summary statistics.
Table 1 sets out these summary statistics over different sample periods. The sub-periods are determined according to whether or not stock prices were subject to government interventions. Before 21 May, 1992, strict restrictions were imposed on daily aggregate price fluctuations on the Shanghai exchange, but not on the Shenzhen exchange (Mookerjee and Yu, 19991, p. 44). The subperiod between 21 May, 1992 and 16 December, 1996 was free of price controls
3.1. THE KAZAKHSTAN STOCK EXCHANGE 
3.1. THE RUSSIAN STOCK EXCHANGE 
PART IV: EMPIRICAL FRAMEWORK
PRILIMINARY DATA ANALYSIS
The framework within the context of this paper deals with Kazakhstan Stock Exchange Index KASE and Russian Stock Exchange Index RTS from 1st of July 2005 to 1st of July 2010. Data has been collected on the daily basis and therefore sums up to 1242 observations. Weak form of efficiency is often tested by carrying out simple regression of form:
rt = β0+∑ βi rt-i + εt (1)
Our test applies the same equation as Emerson et al (1996) and Hall et al (2002) which is optimized for emerging markets like Bulgaria and Russia. This form represents time subscripts and can vary over time
rt = β0t+∑ βit rt-i + εt (2)
The evolution of stock markets in Kazakhstan and Russia in terms of weak-form efficiency can be defined in terms of the time-varying AR (1) model which consists of two equations:
rt = β0t+∑pi=1∑ βit rt-I + εt, εt ~ N(0,ht) (3)
βit = βit-t +uit uit ~ N (0,δ2) and i= 0,1,… p. (4)
rt the rate of daily stock returns at time t computed as the difference between two successive log-levels of the stock price index
εt is assumed to be a white noise process
δ2 for s=0
E(εt) =0; E(εt εt -s) =
0 for s≠0
Although the unconditional variance of εt is a constant δ2, its conditional variance ht could be time-varying, and we will model its behaviour later on.
The test will be conducted in the following 4 steps:
Simple Autoregressive time series model: the AR(P) testing
Unit Root Testing (ADF and KPSS tests)
Kalman Filter or Monte Carlo Simulation
Simple Autoregressive time series model: the AR (P) testing
The first steps in our analysis we need to estimate our AR (1) model for both KASE and RTS index using Ordinary Least Squares (OLS) method. To do so we need to check whether the data series are characterized by ARCH effect. We obtained first difference of logarithmic returns for both stock markets. From the time plot of the series we can see clearly see that there are certain periods with large and smaller volatility by confirming the existence of ARCH effect in both markets. As we can see there is a very high volatility in KASE rather than in RTS returns during whole sample period.
Figure 2  Figure 3
From the figure 2 KASE index, we can observe that the expected value of the magnitude of the disturbance terms is much greater in comparison to RTS index figure 3. We can also observe that high volatility follows by another high changes in returns then slightly lower volatility follows by another low risk, which is then follows by large changes in stock return and again slightly lower volatility. This phenomenon is called volatility clustering Asteriou et al (2007). Therefore it suggests that the volatility is dependent upon past realizations of the asset process and related volatility process. In such case the assumption of homoskedasticity (constant variance) is very limiting and in order to conduct test we need to consider conditional variance.
The most important is the ARCH effect in the residuals of the model. The R –squared is 62.21753 for RTS and 100.6484 for KASE respectively with a probability limit of 0.000. This clearly suggests that we reject homoskedasticity, or that ARCH (1) effects are present. We have also carried out the same test with 6 residuals and the R squared is even higher 190.9889 for RTS and 133.3361 for KASE.
Describe models you estimated variouse tests that I’m planning to carry
FIRST PUT DATA THROUGH SOME PRELIMINARY CHECKING
PART V: ESTIMATION AND RESULTS
Observation in terms of hypothesis and expectations,
PART VI: CONCLUSION
There are four key articles that have contributed significantly to the development of the empirical test of the evolving market efficiency using time varying coefficients and generalized auto-regressive conditional heteroskedastic (GARCH) errors, Monte Carlo Simulation and Kalman Filter.
It is possible to state that there is a gap in the literature that exists for testing the evolving market efficiency in transition economy especially in post soviet countries.
PART VII: LIMITATIONS OF STUDY AND POSSIBLE EXTENTIONS
Recognise your study’s limitations, due to lack of available data or computer programs for specific method you deem apprroperoate and state the posible extentions and further studdyings.
Asteriou D. and Hall S. G. (2007). Applied econometrics a modern approach. Revised edition. Published by Palgrave Macmillan. New York 2007. Ch 14, pp. 249
Bachelier, L. (1900). Theory of Speculation, In Cootner, P.(Ed.), The Random Character of Stock Market Prices, Massachusetts Institute of Technology Press, Cambridge, 1964;
Campbell, J.Y., Lo, A.W., MacKinlay, A.C., 1997. Econometrics of Financial Markets. Princeton University Press, New Jersey.
Cornelius, P.K., 1994. A Note on the informational efficiency of emerging stock markets. Weltwirtschaftliches Archiv 24, 820-828.
Emerson, R., Hall, S.G., Zalewska-Mitura, A., 1996. Evolving market efficiency with an application to some Bulgarian shares. Research memorandum/ Advisory Centre for Education, Project; no.96/18
Fama, E.F. 1965. The Behaviour of Stock Market Prices. Journal of Business 38, no.1: pp. 34–105.
Fama, E., 1970. Efficient capital markets I: review of theory and empirical work. The Journal of Finance, Vol. 25, No. 2, Blackwell Publishing for the American Finance Association, pp. 383-417
Fama, E., 1991. Efficient capital markets II. Journal of Finance 46, 1575-1617.
Fama, E., 1998. Market efficiency, long-term returns, and behavioral finance. Journal of Financial Economics 49, 283-306
Grossman, S. J., and J. E. Stiglitz, 1980. On the impossibility of informationally efficient markets, American Economic Review, 70:3, pp. 393-408
Hall, S. G. and Urga, G and Anna Zalewska-Mitura (1998). Testing for evolving stock market efficiency with application to Russian stock prices. Discussion Paper No. DP 12-98
Lo, A. W. and MacKinlay A. C. (1999). A non-random walk down Wall Street, Princeton University Press, pp. 4-8
Malkiel, B.G., 2004. A Random Walk Down Wall Street: The Time-Tested Strategy for Successful Investing. W. W. Norton & Company, Inc, New York. Ch 11, pp 244-271
Samuelson, P. A (1965). Proof That Properly Anticipated Prices Fluctuate Randomly, Industrial Management Review, 6 (spring, 1965), pp. 41-9
Shiller, R.J., 2005. Irrational Exuberance, Second Edition. Princeton University Press, New Jersey. Ch 10.
Rockinger, M and Urga, G. (2000) The evolution of stock markets in transition economies. Journal of Comparative Economics, 28, 31, pp. 456–72.
Roberts, H. (1967). Statistical versus Clinical Prediction of the Stock Market”, unpublished manuscript, Centre for Research in Security Prices, University of Chicago, May.
Zalewka-Mitura, A and Hall, S.G (1998). AExamining the first stages of market performance: a test for evolving market efficiency. Economics Letters 64 (1999) 1–12
Cite This Work
To export a reference to this article please select a referencing stye below: