HOW DOES POLITICAL INSTABILITY AFFECT MARKET RISK AND THE RISK PREMIUM IN ISRAEL

Rami Saad Autum 2011

Supervisor: Professor Jörgen Hellström Master Thesis, 30 ECTS

HOW DOES POLITICAL INSTABILITY AFFECT

MARKET RISK AND THE

RISK PREMIUM IN ISRAEL

Rami Saad

ACKNOWLEDGMENT

I would like to thank my awesome and brilliant supervisor Professor Jörgen Hellström for his endless patience and guidance, his door was always open to me. Even though this thesis took more than the allocated time Professor Hellström remained supportive and motivational. I will not forget to thank my lovely parents, Emil and Fatineh Saad, for their mental and financial support, during my studies in Sweden.

Abstract

Analysis and assessment of market risk and country risk premium have become a critical component of valuation in recent years in emerging markets. In these markets investors look for a higher rate of return on their investments than in developed countries.

Israel is an attractive country for international investors. This is due to the fact that Israel has one of the highest concentrations of high-tech companies in the world and the tourism industry in the country which is highly attractive for its historical places;

especially in a religious manner.

Israel is a country located in Asia and has experienced many political changes (for many reasons). Thus, it is likely that political decisions taken by the governments affect the risk premium.

In this paper I considered the issue of market risk and the issue of country risk that should be considered explicitly in valuation of the risk premium in emerging countries. Then, the focus, in this paper, is on how the "political instability" in Israel has affected the market risk and the risk premium in the last decade (from the year 5\2001 till 3\2010).

Standard deviation of the returns (estimated by moving average method) of the Israeli stock market were used as a proxy for the market risk. The results show that the political instability affects the market risk. The "Country Default Spread" approach and the "Relative Equity Market Standard Deviation" approach were used to measure the risk premium in Israel. The effect of the political instability on the risk premium was thus found. Further, due to autocorrelation, the robustness of the results was tested by models including lags of the dependent variable. The results from the robustness test show, in most of the analyses, weak and less significant (low confidence level) effect of the political instability on the risk premium.

Contents

1. Introduction……….……….1

2. Background – Israel 2.1 Economic development….………..…...4

2.2 Political situation..….………..…...6

3. Earlier studies…..……….9

4. Method\Theory………...…11

5. Data………...…...18

6. Results 6.1 Results for model 1……….22

6.2 Results for model 2……….24

6.3 Results for model 3……….27

6.4 Robustness test……….………..………….……31

7. Conclusions………....………....………..…...34

Appendix ..………..36

References.…...………40

1. Introduction

Since the world became freer with respect to capital flows during the 1980's¹ when governments deregulated restrictions (regulations) on trade and capital flows, investors started to invest in foreign countries. Before this period, the international investments were under regulations that limited the investors from moving their assets from one country to another. These regulations are almost removed nowadays, especially in developed countries, which make international capital's investors free movers.

When it comes to evaluating their investment opportunities, investors look for the highest return that they can earn consequently with the lowest risk among the opportunities available in the market. In recent years, when the free trade and capital flows policies were applied in most of the developed countries, assessment of market risk and country risk premium has become a critical component of valuation. Every country contains different risks regarding the economic and the political situation. In general, countries located in Asia tend to be more risky than countries located in Europe². With this pattern, investors will require (for) higher return when they invest in risky countries rather than in less risky countries to compensate for holding this risk.

Central parts for the investor’s portfolio are market risk and risk premium. The market risk or the systematic risk (see CAPM model in section 4.1) is the risk that cannot be avoided or diversified from a specific investment. This kind of risk usually derived from the stock market. The standard deviation (square root of variance) of the rate of returns from the stock market is used as the proxy for market risk.

The risk premium in emerging markets is usually measured in comparison with developed countries rather than using accepted and known models. It is an important point that is mention now to understand the methods used in this paper to assess and measure the risk premium for emerging markets. The reason for using this approach to measure the risk premium is the fact that using historical data it is difficult to apply it in the accepted and known models for emerging markets.

Market risk and country risk are affected by two main factors. These two main important factors are the economic and the political stability of every specific country.

If we consider the previous example of a country located in Asia compared to Europe, we can see that generally the first is associated with higher risk than the second. This risk embodied by the fact that Asia is a region where the political situation is more unstable and at the same time less developed than in most of the European countries.

European countries are more developed in the economic aspect compared to Asian countries. For instance, Iraq is a country located in Asia with an unstable regime (until

1 The subject was discussed in the paper Capital Flows to Emerging Markets: The Myths and Realities,

"over time countries realized that the free movement of capital could have widespread benefits" by Bill Block and Kristin Forbes Council of Economic Advisers.

2 this subject was discussed in Asia Risk Monitor: Global risk segmentation and the implications for risk management in the private and public sectors by Daniel M. Hofmann, Group Chief Economist of Zurich Financial Services, in this paper it's mentioned that Asia appears to be vulnerable to list of risks and richer countries tend to be less prone to the same list of risks – Europe included.

the date of this study) and will therefore be classified, in the investors' eyes, as a risky country compared to Germany; hence, investors will require a higher return in order to invest in Iraq.

There are several differences between developed countries and emerging ones. In this paragraph some characteristics mentioned below may help to distinguish between the developed and emerging countries. In the developed countries, in general, a stable economic environment associated with low risk can be found. This is due to clear rules and regulations settled by the developed governments which control the market and the transactions. The developed countries are unionized with the organization so called OECD (Organization for Economic Cooperation and Development). The emerging countries commonly are associated with rapid growth in production and have a potential for higher profit with higher risk than other developed or underdeveloped countries.

Political instability is a situation in a specific county where the political system experiences tensions. This situation can appear in terms of wars, turmoil, elections, or other events that can cause tension in the political regime. These periods are characterized by non-convenience situation and hurt the economic stability in the country.

The purpose of this paper is to study how and to which extent political instability affects market risk and the country risk premium as well. The study is a case study focusing on this issue for Israel.

Israel is located in Asia and classified as an emerging market (up to April 30, 2010). I found that Israel is a suitable country to examine the affect of the political instability on the market risk and the country risk premium for many reasons. The most important reason is the fact that Israel attracts many investors from all over the world, so this thesis could provide important information for those who are looking to invest in Israel. On the other hand, Israel has also experienced many politically instable periods. The purpose of this research is to study the market risk and the risk premium in the Israeli market focusing on political changes under the assumption that Israel is an emerging market.

In Israel, the domestic security situation was difficult in the recent 10 years, as the governments were trying to sign peace agreements with Arab and Muslim(s) countries with which Israel had no prior diplomatic relations. It is worth mentioning that in general, the security situation did not influence the daily running of the industries and other sectors except of special short term cases such as the war in Lebanon³ in the year 2006.

This research is important for all investors who are investing in the Israeli market.

This paper provides investors with a sufficient background about the market and the country risk and with information how these risks change during the political instability periods. It is as well of academic interest as it provides evidence of the link between political instabilities and financial risk.

3 According to ANIMA Investment Network, project funded by European Union.

It was not easy to find historical data in order to calculate risk premiums. The data concerning the Israeli market which was found is strictly limited and only from the past decade. Therefore, I was able only to work with the available restricted data. The problem with the data available for a short time may come with a large standard error.

The politically instable periods that I choose are subjectively selected. The unstable periods cover eight events from 2001 to 2010.

In Section 2, a background of the economic development and political situation in Israel is provided. Section 3 contains a review about the most relevant studies for the current question at hand. Section 4 includes the methods used for measuring market risk and country premium. Section 5 contains the data available for measuring market risk and risk premium. Section 6 shows the results of market and country risk in Israel. Section 7 summarizes the paper and presents the conclusions.

2. Background – Israel

The economic and political background of Israel is introduced to the reader.

2.1 Economic development

On the May 14, 1948 the Jewish Agency claimed independence, one day before the British Mandate expired, and gave the name to the country Israel.

Israel is considered today as one of the most industrial developed and economically advanced countries in Asia. It is worth mentioning that Morgan Stanley Capital International (MSCI) announced that Israel will be classified as a developed country starting in May 2010. However, in this paper Israel is considered as an emerging market. The country has been ranked third in the region on the World Bank's Ease of Doing Business Index (in 2006) and had the second largest number of start up companies after the US.

Despite a shortage of natural resources, developments in the agriculture and industry sectors made Israel in the past 10 years self sufficient in food, especially in grains and beef. Israel imports fuels, raw materials and military equipment, while the country exports fruits, vegetables, pharmaceuticals, software, chemicals, military technology and diamonds. Israel is a leading manufacturer of these products in the world. The tourism sector in Israel, especially for religious people, is an important industry, since beaches, archaeological and historical sites with convenient temperature may be found in all parts of the country. Almost 3 million tourists visit Israel each year.

The gross domestic product (GDP) per capita is ranked around 30^th in the world. It has grown rapidly since Israel claimed independence in 1948. The growth rate of Israel (in fixed prices) from December 1995 to March 2010 is presented in Figure 2.

The quarterly gross domestic product (GDP) per capita (in current prices) from December 1995 to March 2010 is illustrated in Figure 1.

Figure 1: Gross domestic product per capita for Israel 0.0

10.0 20.0 30.0 40.0

Dec-95 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Dec-03 Dec-04 Dec-05 Dec-06 Dec-07 Dec-08 Dec-09

Time

GDP per capita

Figure 2: Growth rate for Israel

According to ANIMA Investment Network, in the past 20 years, Israel has developed an industrialized economy despite being classified as an emerging market and carrying ongoing political tensions in the country region. Nowadays, it is a regional economic power with a GDP of US$ 190 billion (NIS 699 billion) recorded in 2009.

The main reasons for this strong growth were exports, private consumption, and expansion in high technology industries and tourism. The number of foreign investors in the recent 20 years has been also grown rapidly.

The domestic currency in Israel is the New Israeli Shekel (NIS) and it's a free convertible currency in the world. On June 5, 2010, the exchange rate for Shekel to Dollar is 3.853 and to Euro 4.703.

Despite Israel being an emerging market, it presents and offers the economic stability of a developed country and simultaneously it continues to offer growth and profit opportunities of an emerging market.

To encourage both local and foreign investments, the State of Israel offers a wide range of incentives (such as tax benefits and grants) and benefits to investors in industry, tourism and real estate. The government attempts to give special attention to investors in hi-tech companies and also R&D activities. The hi-tech industry in Israel is recognized as one of the world's outstanding technology centers and, as mentioned above, is the second largest in terms of start up companies. A major factor in the success of this sector is the clear government policy of leadership, support and encouragement of industrial R&D of the Office of the Chief Scientist (OCS) at the Israeli Ministry of Industry, Trade and Labor.

The Israel Stock Exchange market called as well Tel Aviv Stock Exchange (TASE) is the only stock exchange in Israel and is located in Tel Aviv city. TASE is supervised by the Israel Security Authority and it is a private company controlled by banks and other corporations. TASE plays important role in the Israeli economy. In this marketplace all types of securities such as stocks, bonds, funds etc. are traded.

-10,0 -5,0 0,0 5,0 10,0 15,0 20,0

dec/95 dec/96 dec/97 dec/98 dec/99 dec/00 dec/01 dec/02 dec/03 dec/04 dec/05 dec/06 dec/07 dec/08 dec/09

Economy Growth Rate

Time

Different kinds of companies are listed in TASE, for instance, companies dealing with commerce and service, industry, real estate and construction, investment and holding, as well as with insurance.

The leading index in TASE is TA-25, which contains the 25 largest stocks in TASE;

TA-100 contains the 100 largest stocks in TASE; and TA-75 contains the stocks of TA-100 not listed in TA-25. In Figure 3 below the weekly TA-100 index from January 2000 to April 2010 is presented.

Figure 3: TA-100 Index for the last decade

Figure 3 shows that there is an increasing pattern in the TA-100 index from 2002 to 2008. In 2008, the effect of the global financial crisis can be clearly visualized. After slower growth in 2008 and 2009, due to the global slumps caused by the financial crisis, the stock market started to recover again according to new records in the stock market in 2009 as can be seen in Figure 3.

2.2 Political situation

Israel has a democratic parliamentary system. It is recognized to have the most democratic government in the Middle East, according to Zionism (Jewish national liberation movement) and Israel. All Israeli citizens over 18 years old have the right to vote for a government. This political system includes, as in every democratic system, legislative, executive and judicial branches. Israel applies the power of separation between legislative, executive and judicial branches. The Knesset is the parliament which includes 120 members assembled by the parties selected in the election. Today there are main three parties in Israel; Likud – right party (Zionist and capitalist), Kadima – central party (right and Zionist) and Labor – central party (left and Zionist).

In order to completely understand the political situation in Israel, it is important to mention the Israel-Palestine conflict. This conflict is an ongoing political tension between two nations and countries. It is unclear when this conflict has appeared but it is mentioned already in the Old Testament of the Bible. The conflict mainly focuses on the controlling of the land, especially Jerusalem. Many attempts from other countries were focused on negotiation in order to find a solution for this conflict. So far, the conflict is still ongoing and the tension between two nations remains.

0 500 1000 1500

06/01/00 25/05/00 12/10/00 22/02/01 12/07/01 29/11/01 11/04/02 29/08/02 09/01/03 29/05/03 16/10/03 04/03/04 22/07/04 09/12/04 28/04/05 15/09/05 02/02/06 22/06/06 09/11/06 22/03/07 09/08/07 27/12/07 06/05/08 25/09/08 05/02/09 25/06/09 12/11/09 01/04/10

Index in points

Time

TA Index - 100

As it was mentioned previously, Israel is highlighted in the media as a place which constantly experiences political changes. Therefore, eight different political tension periods are presented to be used later to measure the effect of political instability periods on the market risk and risk premium:

1- The Second Intifada War took place during the period from 5/2001 to 10/2003.

During this period Israel experienced a period of intensified Palestinian-Israeli violence, which began in late September 2000 and ended in October 2003. I covered the period starting from 5/2001 because of lack of data before this period. The death toll was big and included both military and civilian casualties. Approximately 3,333 Palestinians, 1,000 Israelis and 64 foreign citizens were killed (according to Haaretz Israeli news). According to these numbers it is obvious that this period, the Second Intifada, is an essential event which is worth to focus on and to find out how this event affects the risk premium in Israel.

2- The election for the 17th Knesset took place during the period from 3/2006 to 4/2006. The election results were surprising and unexpected. The voter turnout of 63.2% was the lowest ever.

3- The 2006 Lebanon War took place during the period from 7/2006 to 9/2006. It is also called the Israel-Hezbollah 2006 War and known in Lebanon as the July War and in Israel as the Second Lebanon War. It was a 34-day military conflict in Lebanon and northern Israel. The war started on the July 12, 2006, until United Nations achieved ceasefire on the August 14, 2006. During the war, at least 1,244 people were killed, and at least one million people were left without homes (according to BBC news).

4- The Gaza War took place during the period from 12/2008 to 1/2009. This war took place in Gaza during the winter 2008 for three weeks. The aim of this war was to stop rocket attack from Gaza to Israeli regions. Israeli forces attacked buildings belonging to government, police and military. Over 1,000 Palestinians and 13 Israelis were killed (according to BBC news, Jerusalem).

5- The elections for the 18th Knesset took place during the period from 2/2009 to 3/2009. The elections came after the Prime Minister Ehud Olmert resigned from the Kadima party and the successor Tzipi Livni failed to form a new government. If Olmert remained in his office or if Livni had formed a coalition government, the elections would have been scheduled for 2010 instead.

6- In 9/2009, a UN special mission headed by Judge Richard Goldstone, produced a report accusing both Palestinian militants and Israeli Defense Forces of war crimes and possible crimes against humanity, recommending to bring those responsible to justice. The UN Human Rights Council endorsed the report, criticizing Israel but not Hamas.

7- In 1/2010, Israel experienced serious tension with Turkey on the January 11, 2010 after Israeli Vice president for Foreign Affairs, Minister Danny Ayalon invited the Turkish ambassador Ahmet Oguz Celikkol to the meeting. During the session, Celikkol was seated below Ayalon, with cameras rolling. The aim of this appointment was to reproach Turkish policies against Israel. In response, Turkey wanted to return

the ambassador back from Israel but in the last minute Israel released an apology to Turkey. Currently, the tension is still ongoing.

8- In 2/2010, according to the media, a big controversy was among several countries (especially European countries) concerning the identity of the killer of al-Mabhouh (senior Hamas military commander). Al-Mabhouh was killed after being followed by at least 11 individuals carrying fake passports from various European nations. The countries which were involved because of the fake passports accused Israel for using their passports for their own political interests. In this period a pressure was on the Israeli government and most of the European countries asked to open an investigation.

I focused on these eight periods with intent to analyze the effect of political instability on the market risk and the risk premium in Israel. Even though, there is infinity of political changes throughout the past decade, I found these events to be the most essential concerning political changes and instability.

These eight political instability periods are collected from the available media in the internet. I went through the Israeli news websites such as www.ynet.co.il, www.walla.co.il, www.haaretz.com and www.jpost.com. These entire websites contain daily news covering mainly Israel and the Middle East.

3. Earlier studies

There is a number of papers discussing market risk, risk premium and country risk.

The most relevant study to my thesis is the paper by Aswath (2003). In this paper Aswath focuses on two main questions. The first is whether the companies should take into account country risk when it pertains to investments in emerging countries.

The second is how to measure the exposure to country risk in emerging countries. In this paper two approaches for estimation of country risk premium are presented as well. The first is a historical risk premium approach, while the second is an implied premium approach. In the first approach, the historical potential return of the stock market is compared to the return from risk-free assets in the same periods, difference between the returns gives us the risk premium. The author claims that this approach is not relevant for emerging markets (such as Israel) because of the large standard error that we can have. He proposed an alternative approach, so called “modified historical risk premium”, to measure country risk in order to avoid the noise that can come from the emerging markets’ data. The second approach does not require any historical data but requires the use of the present value equation to extract the required return on equity (see the paper by Aswath (1999). For this thesis part of the first approach (modified historical risk premium) is used, which includes two methods “Country Bond Default Spreads” and “Relative Equity Market Standard Deviations” to measure the risk premium in Israel.

The paper by Soultanaeva (2008) analyzes the impact of political news on the return and volatilities of the Baltic's stock markets. The results indicate that the political news have led to lower uncertainty in the stock market of Riga and Tallinn in the period 2001-2003 regarding foreign and domestic news, except of Russian. Political news from Russia increased volatility in Tallinn stock market in the same period.

Furthermore, it seems that in the period 2004-2007 the effect of political news on the Baltic's stock markets was significantly lower as compared to the period 2001-2003.

Relatively,Vilnius stock market seems to be unaffected by political news in the periods 2001-2003 and 2004-2007. The main conclusion from this paper is that the sensitivity of the Baltic's stock markets seems to decrease over the two sample periods.

The paper written by Hellstrom and co-author Soultanaeva (2011) studies almost the same subject as was mentioned in the previous paper. Accordingly, the authors used different methods to test the causes of stock markets’ jumps followed by political news. The results are in line with those of the previous mentioned paper.

The paper by Hung et al. (2007) tests the jump intensity and volatility in both Taiwan stock and foreign exchange markets during presidential elections. The results indicate that during the presidential elections the jump intensity and the volatility of both markets increase.

The paper by Chan (2001) tests the impact of salient political and economic news on the stock return volatility, the price volatility and the daily volume in the Hong Kong stock market. Subsequently, the author found that the salient political news cause a

negative effect while salient economic news cause a positive effect on the stock market.

The paper by Jean-Claude and Jean (1991) came after famous business and financial magazines “Euromoney” and “International Investor” which published rating of countries "creditworthiness" in 1987. In their paper authors tried to replicate the results from “Euromoney” and “International Investor's” ratings, which was successfully done. Their results indicate that country risk rating responds to the most economic and political variables. For instance, in both magazines they rank less indebted countries higher than more indebted countries. In conclusion, both

“Euromoney” and “International Investor's” results are highly correlated and both are agree on factors or variables that determine the country risk of the assessed countries.

Another paper by Aswath (1999) suggests a new approach to measure risk premium.

All known methods are based on historical data, which is a main component in measuring risk premium, as in the CAPM model and others. In this new method which differs from the known and traditional methods, no historical data are needed.

In addition, the author suggests deriving the equity risk premium from equity prices.

In this method we can take externally (without using historical data) the value of the market, the expected dividends next period, and the expected growth rate. From these values we can extract the required return on equity by using simply the present value equation of the market value, and by subtracting the risk-free rate from the required return on equity we get the risk premium. Aswath mentioned in his paper the disadvantage from using historical data which sometimes tend to be limited and noisy and comes with higher standard errors. Therefore, he suggests this new method to measure risk premiums without the need of historical data.

There is another interesting paper by Ekpenyong and Umoren (2010), in which the authors turn their attention to the political risk issue as an integral part of almost every business. The definition of political risk is wider in this paper. Companies usually use defensive or integrative strategies to cope with political risk. The writers suggest adopting the modified integrative strategy to cope with political risk.

The paper by Clare and Gang (2010) studies the effect of exchange rate and political risks in foreign direct investments. According to the paper, companies investing in developed countries observe the past and the present variation in exchange rates (they assume that in developed countries the political issues are stable over time). When the investments are made in less developed countries (emerging markets) they observe the present and the future variation in exchange rate (they rely more on the expectations since political situation is less stable than in developed countries).

Decreasing political risks will increase foreign direct investments. The main results are negative effect of exchange rate risks to foreign direct investments and positive effect of political stability to foreign direct investments.

For summarizing the main findings from previous papers, I have found that political risks either in terms of election, wars or unstable regime affect the market risk and the risk premium of the country. These papers, mentioned above, found explicit link between the political risk, market risk and risk premiums and some of them encourage companies to take these risks into account when it comes to investments in emerging or other undeveloped markets.

4. Theory\Method

When it comes to evaluating their investment opportunities, investors look for the highest return that they can earn corresponding with the lowest risk among the opportunities available in the market. Hence, the valuation of market risk and risk premium is an important factor for the investors when they decide whether to make a move to a specific market or not, especially in Israel, because of political changes.

There are several ways to calculate the risk premium by using accepted models. In this section the reader is presented to the most two appropriate methods (in section 4.2.2) from the total prior selection (in section 4.1). The way of calculating the market risk is presented (in section 4.2.1) and two ways of calculating the variance of the returns are presented by a moving average method and by using GARCH models.

Clarification of “market risk” and “risk premium” is therefore required.

Market risk reflects the risk of declines or losses in the value of any portfolio due to uncertain factors related to the market. The market value of a well diversified portfolio is affected by so called uncertain macro-economic factors, such as interest rate, stock prices and foreign exchange rate. Market risk is also called systematic risk (see CAPM model in section 4.1). This kind of risk cannot be diversified away from the entire risk in any portfolio. In other words, holding a not well diversified portfolio may include market risk and other risks.

Risk premium reflects the difference in the risks associated with two or more different investment choices faced by the investor. For instance, the investor will choose between risk free bonds and other risky asset. Then, the difference between the return of the specific risky asset to the bond is the risk premium, or one can say that the term

"risk premium" is the reward for holding a risky investment rather than a risk free one.

4.1 Measuring Risk Premium

All risk and return models break the return from any investment into two components.

The first is the "built in" risk for a specific investment (unsystematic risk), the second is the market risk that cannot be diversified or eliminated (systematic risk).

As stated above, there are several different models for measuring the risk premium.

For instance, there is the Capital Asset Pricing Model (CAPM), the Arbitrage Pricing Theory model (APT), and the Multi Factor model⁴. In this paper the CAPM is presented shortly as a background for the reader to understand how the risk premium, in general, is measured since the CAPM model is the most widely used model concerning this subject.

4More details about measuring risk premium models mentioned above are available in the book

“Investments”, 8^th edition, by Bodie Z., Kane A., and J. Marcus A. Chapter 9 p. 279 and chapter 10 p.319.

The CAPM model is used to determine the required rate of return on investing in a risky asset, when investors believe that they should be compensated over the expected value of the theoretical risk free rate. It takes into account the sensitivity of the assets to the market risk, which is usually denoted by Beta and the expected market return (the Beta of the company is the risk of investing in the specific company compared to the risk in investing in the overall market). In this model there are also some assumptions shown like no transaction costs and no taxations.

The CAPM formula is:

Rr= R

f + β * (R M- R

f) Where:

Rr is the required rate of return on the risky asset.

Rf is the risk free rate of return in the market.

RM is the expected return in overall market.

β is the specific Beta of the company or the sector of the assets.

(RM- R

f) is the expected risk premium

When the risk free rate of return is observed, the expected return from the overall market and the appropriate Beta, the required rate of return on the specific assets can be simply measured. Then the CAPM is actually pricing the assets. When the assets are traded below the expected return, then the assets are undervalued and vice versa.

For instance, if β is equal to 1 (β=1), consequently, the risky asset(s) has the same risk or standard deviation as the market portfolio, both move in the same direction. When β is larger than 1 (β>1), consequently, the risky asset(s) is more sensitive than the market portfolio. In this case, if the market portfolio increases in 1%, the risky asset will increase in more than 1%. When β is less than 1 (β<1) the risky asset is less sensitive than the market portfolio. In this case, if the market portfolio increases in 1%, the risky asset will increase in less than 1%.

4.2 Measuring market and country risk

4.2.1 Estimating the market risk

Market risk is usually measured with the standard deviation of returns for an appropriate stock market index. Two methods to estimate the variance of the returns from the stock market are presented here: “moving average” method and the

“GARCH model”. The variance of the returns will be used later to measure the market and the country risk. Here GARCH model is used to test the robustness of the results from measuring the risk premium (only) based on the “moving average”

method.

Moving average:

The “moving average” method is a statistical tool widely used to smooth the noisy data in order to create the trend of the data. It is simply taking the average value over specific time periods or specific subset from the entire data, as follows:

σ

²_t

=(1/M)*∑

This method was used to measure the market risk in Israel (in section 4.2.2) and to measure the country risk in the method "Relative Equity Market Standard Deviation"

(in section 4.2.3) for both Israeli and USA stock markets. It is widely acceptable in measuring the standard deviation of the returns by the “moving average” to assume that the mean of the returns is zero (ȓ≈0). The standard deviation of the rate of return from the stock market is widely used as a proxy of market risk. For this purpose (measuring market risk) I calculated the returns of the Index-100 from the Israeli stock market (TASE). The standard deviation is calculated by the moving average method. The returns are in weekly terms, so I took the first eight weeks and calculated the standard deviation; this is my first observation. The second observation is the standard deviation of the returns from week two to week nine. The third observation is the standard deviation of the returns from week three to week ten and so on. Thus, each observation (standard deviation) is the average of eight returns’ standard deviations. In this way I calculated the standard deviation from the Israeli stock market which will be used later as a proxy of the market risk.

GARCH model:

GARCH (generalized autoregressive conditional heteroskedasticity) is a part of the ARCH family. The main objective of using this method is the fact that the volatility of return/risk premium can vary over time. In other words, the volatility of the return is not constant over time.

To understand the need for models like GARCH, we need first to understand the term

“homoscedasticity”. Homoscedasticity refers in econometrics to the assumption that the variance of the dependent variable is constant and do not vary over time. This assumption is important since when the variance is constant, the estimated coefficients in the model will be consistent and efficient⁵. If we have heteroskedasticity it means that the Ordinary Least Square (OLS) estimator will be inefficient.

When the variance of the dependent variable in the model is not constant, i.e. vary over time, it is called “heteroskedasticity”. This case is the opposite of homoscedasticity. According to the paper by R. Perrelli 2001, in most of the cases in financial data large and small errors occurs in clusters, which means that large errors usually follow large errors and small errors usually follow small errors – clustering.

All these facts allow us to conclude that the variances are not constant over time, it vary over time, so the assumption of homoscedasticity is violated. That is why we need to use the GARCH model to test if the time series have the clustering characteristic.

5 Consult the book Econometric Analysis, 6^th edition, by William H. Greene.

The main issue in this method is to analyze the volatility of the error term over the political instability periods under the assumption of heteroskedasticity. For this purpose the software Stata (statistical package for data analysis) was used to produce the variance of the rate of returns. To proceed with this model I assumed GARCH(1,1), which mean we have lag one for the variance of the error term and lag one for the error term itself.

The estimated standard deviations (square root of variance) from the GARCH model is used to analyze the risk premium from the method "Relative Equity Market Standard Deviation" (presented in section 4.2.3).

The consequent procedure of producing the variance is presented by using the software Stata.

The first step in this procedure is to find the error term, for this purpose I regress the following:

r

t

=α+β*r

_t-1

+ε

_t

where r_t is the rate of return at time t, r_t-1 is the rate of return lagged one period and ε_t is the error term of this regression. To model the conditional variance a GARCH (1,1) model is specified as

ε

t

= σ

t

* u

t

where u_t is standard normally distributed (i.e. u_t

~

N(0,1)) and

σ

²_t

=γ

₀

+γ

₁

* ε

_t-1

+γ

₂

* σ

²_t-1

The error termεt is conditionally heteroscedastic with respect to εt-1,

Var

(

ε

t│

ε

t-1

) = γ

0

+γ

1

* ε

²t-1

4.2.2 Country risk

Israel is classified as an emerging market for the period of this study. In both papers

"Measuring Company Exposure to Country Risk: Theory and Practice" and

"Estimating Equity Risk Premiums" by Aswath (2003), it is not applicable to use short and volatile historical data in order to measure risk premium in emerging markets, as the models mentioned in section 4.1 require. Aswath, in his paper

"Estimating Equity Risk Premiums" (2003) shows the same tendency in Europe when, for example, Germany is a mature country in economy aspect but the market, on the other side, does not have to share the same characteristics of mature market. In the same paper the author estimates risk premiums for several European countries for 26 years from 1970 to 1996 and these results come with 5% standard errors. Thus, we can imagine how much it will be noisy and useless to use historic data to measure risk premium in emerging markets such as Israel. Therefore, I will use the US market as a benchmark since investors often see the US market as one of the safest markets in the world.

For the reasons mentioned above, I adopted the alternative method, by Aswath (2003), to measure the risk premium in an emerging market. This alternative method, called the "Modified Historical Risk Premiums", is presented as follows:

The expected equity premium for emerging markets can be written as:

Expected return = Expected return from mature country + Country risk

The country risk can be represented as the differences between the expected return of the country minus the expected return from mature country. The purpose is to measure the country risk that can reflect the risk premium for every specific country.

The following 2 methods are used widely to measure country risk:

1- Country Bond Default Spreads is the most common and easiest method used in measuring country risk. In this method we look at the yields to maturity from Treasury bonds in the two countries, the mature and the country in question. The difference in the yields to maturity between the two countries should reflect the country risk and the default risk hidden in the specific country.

2- Relative Equity Market Standard Deviation, where we measure the country risk by using the volatility of the stock exchange markets. We believe that stock exchange markets associated with high volatility reflect more risky markets rather than low volatility. If we divide the standard deviation of the stock market in the country by the standard deviation of the stock exchange market of the mature country, we obtain the relative standard deviation.

Relative Standard DeviationSpecific country=

σ

Is

/σ

US

If the relative standard deviation is multiplied by the premium which is used for the mature market, the equity risk premium for the specific country is obtained.

Equity Risk Premium

Israel=Risk Premium

USA*Relative Standard Deviation Is.

For the purpose of my research I adopted the approach of "Country Default Spread"

and the approach of "Relative Equity Market Standard Deviation" for two main reasons. First of all, these approaches are the most widely used measures of country risk. Secondly, collecting data for Israeli market was not that easy and the limitation of time only enabled me to work with the available data that I could obtain. Finally, I found that the "Country Default Spread" and the "Relative Equity Market Standard Deviation" are the most suitable approaches in this case study of Israel.

In the method "Relative Equity Market Standard Deviation" the standard deviation, which comes from the moving average method, and the variance from the GARCH model, are used.

Hence, "Country Bond Default Spreads" and "Relative Equity Market Standard Deviation" are used to study the effect of the political instability on the Israeli market.

In the "Country Bond Default Spreads" method, monthly measured average of the yields to maturity of the Israel Treasury bond for 10 years are compared to USA Treasury bond for 10 years. The differences between the yields are used as the default spread or the risk premium for the Israeli market. In the "Relative Equity Market Standard Deviation" method, the ratio of the volatility on Israeli Equity Market to the volatility on the USA Equity Market is observed. Finally, the changes in the risk premium, through the past decade focusing on the political tension periods, are analyzed.

4.3 Econometric models

To test if the market risk/risk premium is affected by political tension three econometric models are used in order to analyze the data. The first model is specified as:

(1) σt = α + β1 * MRt + β2 * (SD GDP)t + β3 * PIt + εt.

(2) RPt = α + β1 * MRt + β2 * (SD GDP)t + β3 * PIt + εt.

Here, in equation (1) σ is the market risk in Israel, measured by stock market standard deviation. In equation (2) RP is the risk premium of Israel, measured either as bond spread or relative stock market standard deviation. MR (in equation (1) and (2)) is the Moody's rating of Israeli Treasury bonds. SD GDP (in equation (1) and (2)) is the ratio between the volatility in GDP growth in Israel to the volatility in GDP growth in USA. PI (in equation (1) and (2)) is a dummy variable controlling for periods of political instability. For periods of political instability the dummy variable PI take the value 1, otherwise zero. The "β" are associated coefficients for the variables and the

"ε" is a random i.i.d. term.

The need for other variables, except of political tension periods, is to control for different factors that can affect the market risk/risk premium. The market risk/risk premium could be affected from different factors except of political changes, for instance, Moody’s rating and the growth in GDP as well. For these potentially changes, the variables MR and SD GDP are added to the model to control the effects of other factors, so I can extract the effect of political instability.

The second model is specified as:

(1) σt = α + β1 * MRt + β2 * (SD GDP)t + β3 * Elt + β4 * PI2t + εt.

(2) RPt = α + β1 * MRt + β2 * (SD GDP)t + β3 * Elt + β4 * PI2t + εt.

Here, in equation (1) and (2), the variables σ, RP, MR and SD GDP are described in the first model. El (in equation (1) and (2)) is a dummy variable controlling for periods of elections. PI (in equation (1) and (2)) is a dummy variable controlling for periods of political instability (without elections). For periods of political instability the dummy variable PI take the value 1, otherwise zero and for periods of elections the dummy variable El take the value 1, otherwise zero. The "β" are associated coefficients for the variables and the "ε" is a random i.i.d. term.

The purpose of the second model is to check and to control the effect of the political instability periods on risk premium in Israel. It can be argued that elections are not political instability in the definition so I decide also to separate the elections periods.

The third model is specified as:

(1) σt = α + ∑

+ β9 * MRt + β10 * (SD GDP)t + εt.

(2) RPt = α +∑

+ β9 * MRt + β10 * (SD GDP)t + εt.

Here, in equation (1) and (2), the variables σ, RP, MR and SD GDP are described in the first model. PIit (in equation (1) and (2)) is a dummy variable controlling for 8 periods (exactly in the same order as in section 2.2) of political instability. For periods of political instability the dummy variable PIit take the value 1, otherwise zero. The

"β" are associated coefficients for the variables and the "ε" is a random i.i.d. term.

The purpose of the third model is to check and to control the effect of each political instability periods separately on market risk and risk premium in Israel. There are 8 political periods and this model tests to which extend each period of political instability affects the market risk and risk premium in Israel.

It is important to note that the four regressions have been run for every model estimated by ordinary least squares (OLS). The first regression ran equation (1) when the market risk (dependent variable) is the stock market standard deviation estimated by moving average method. The second regression ran equation (2) when the risk premium (dependent variable) is the difference between the yields to maturity on Treasury bonds. The third regression ran equation (2) when the risk premium (dependent variable) is the relative standard deviation on the equity markets estimated by moving average method. The fourth regression ran equation (2) when the risk premium (dependent variable) is the relative standard deviation on the equity markets estimated by GARCH model.

To summarize, equation (1) is concerning market risk, equation (2) is concerning the method "Country Default Spread" and the method "Relative Equity Market Standard Deviation".

5. Data

As it is described in the previous section, the return standard deviations (estimated by the moving average method) of the Israeli stock market will be used as a proxy for market risk. "Country Bond Default Spreads" method and the "Relative Equity Market Standard Deviation" method are used to test the effect of the political instability on the risk premium in Israel. The monthly yields to maturity on bonds issued by Israel Treasury and US Treasury bond for 10 years are collected, starting from the period 5/2001 to 3/2010. These 10 years yields to maturity are to be used for the "Country Bond Default Spreads" method. For the "Relative Equity Market Standard Deviation" method the weekly indexes from Tel Aviv 100 in Israel and from S&P500 in USA in weekly returns are collected from 1/2001 to 3/2010. The monthly yields to maturity of the Israeli Treasury bond are collected from the website of the Bank of Israel (central bank) and the monthly yields to maturity of the US Treasury bond are collected from the website of the USA Central Bank (Federal Reserve System). The weekly rate of returns of the Israeli market are collected from the website of the Tel Aviv Stock Exchange (TASE) and the weekly rate of returns from the S&P500 in USA are collected from the website of Standard & Poor’s (S&P).

The "Country Bond Default Spreads" method focuses on the difference between the yields to maturity that the Israeli's bond gives over the US Treasury bond. The difference expresses the risk premium hidden in the Israeli market. In this method I focused mainly on the difference in the risk premium between the two countries and I gave explanations of the volatility through time while the attention is directed towards the changes in the political instability. For the "Relative Equity Market Standard Deviation" method the analysis of the difference in the standard deviations (estimated either by moving average or GARCH model) between Israel and USA equity markets is given. This shows the volatilities in the market that can reflect the risk premium in Israel.

To this data, information about Moody's rating of the Israeli Treasury bond is added through the time in aim to analyze and explain the changes in the Israeli risk premium regarding any kind of disturbances.

The changes in monthly yields to maturity on Israel and USA Treasury bonds between May 2001 and March 2010 are presented in Figure 4 below.

Figure 4: Changes in monthly yields to maturity on Israel and USA Treasury bonds.

Figure 4 shows an almost stable difference in yields to maturity between USA and Israel. The big difference can be seen between the years 2001 to 2003 mostly followed by the Second Intifada.

Figure 5 below presents the weekly standard deviation (estimated by moving average method) in Israel and USA equity market between January 2001 and March 2010.

Figure 5: Weekly standard deviation in Israel and USA equity markets.

Figure 5 shows that during the last 10 years the US and Israeli equity market standard deviation are in the interval from 0.02 to 0.08, except for the period in the Second part of 2008. It seems to be the results of the financial crisis in the global market 2008- 2009.

0 2 4 6 10 8 12 14

m aj-01 nov -01 m aj-02 nov -02 m aj-03 nov -03 m aj-04 nov -04 m aj-05 nov -05 m aj-06 nov -06 m aj-07 nov -07 m aj-08 nov -08 m aj-09 nov -09

y ields to matu ri ty

Monthly periods

USA Israel

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

3/19/2001 3/19/2002 3/19/2003 3/19/2004 3/19/2005 3/19/2006 3/19/2007 3/19/2008 3/19/2009 3/19/2010

Time

Standard Deviation

USA Israel

In order to estimate risk premium (RP) with the GARCH model, Table 1 shows the outcome from the Stata software for the Israeli stock market and Table 2 shows the outcome from the Stata software for the USA stock market:

Table 1: GARCH model, Israeli stock market

Coef. Std. Err. z

Return:

rt-1 .026 .048 0.54

constant .000 .001 2.55

Arch:

arch L1 .1 .022 4.59

garch L1 .831 .044 19.04

constant .000 .000 2.16

Table 2: GARCH model, USA stock market

Coef. Std. Err. z

Return:

rt-1 -.046 .056 -0.83

constant .002 .001 2.03

Arch:

arch L1 .227 .028 8.06

garch L1 .731 .036 20.35

constant .000 .000 2.94

Table 1 and 2 together show that we have ARCH term (

ε

_t-1) and GARCH term (

σ

²_t-1) for both markets (at 5% level), which means that

γ

1 and

γ

2 are significantly different from zero. This information indicates that we have conditional “heteroskedasticity”

and the data is clustered.

The last step is to produce the fitted or the estimated variances of the error term

σ

²_t, for both Israel and USA,

σ

²_t^Is and

σ

²_t^US, respectively. This step is also done in Stata.

Table 3 gives descriptive statistics for the dependent variables that are analyses in the models used in this study and for the independent variable measuring the ratio in the changes between Israel to USA in "standard deviation in gross domestic product" as well:

Table 3: Descriptive Statistics.

Default Spread

Relative Equity Market SD

(by

“moving average”)

Relative Equity Market SD (by GARCH

model)

Ratio in GDP quarter changes

(SD GDP)

Political Instability

(PI)

Moody’s Rating

(MR)

Mean 2.578 1.289 1.250 2.244 0.393 0.224

Standard

deviation 1.637 0.560 0.358 2.783 0.491 0.419

Variance 2.681 0.314 0.128 7.747 0.241 0.176

Max 7.788 3.545 2.372 12.872 1 1

Min 0.130 0.494 0.464 0.311 0 0

Moody's rating and the political instability independent variables presented in Table 3 are dummy variables that can take values 0 or 1. For political instability the dummy variable PI take the value 1 when there is political instability situation in the specific period, otherwise 0. When Moody's rating upgrade the Israeli Treasury bond, the dummy variable MR take the value 1, otherwise 0. It is worth mentioning that I did not include control variables for the period of financial crisis in the global market 2008-2009. Thus, I assume that both countries, Israel and USA, are affected equally by the financial crisis in this period.

The political tension periods, mentioned in section 2.2, are used to analyze the market risk/risk premium over time. Afterwards, the yields to maturity, standard deviation from Israel and USA equity market and the information in section 2.2 are the basic data used to analyze the effect of the political instability on the market risk/risk premium in Israel.

6. Results

In this section, the results from the empirical analysis are presented. The results are presented separately for the three models. In Model 1, political instability periods are aggregated into one dummy variable. In Model 2, political instability periods are separated in two dummy variables between election periods and other political instability periods. In Model 3, all political instability periods are separated into eight dummy variables to test the effect of each individual period of political tension.

6.1 Results for Model 1:

Table 4 shows the results from running a regression when the dependent variable is the market risk in terms of returns’ standard deviation, estimated by moving average method. The independent variables are Moody's rating to Israeli Treasury bond, the ratio in GDP standard deviation between Israel and USA and the political instability periods as given by:

σ t/ RPt = α + β1 * MRt + β2 * (SD GDP)t + β3 * PIt + εt.

For this regression the residuals are significantly auto correlated and the adjusted R² value is 0.32 with 473 observations. It is important to note, that the standard errors in this table are not robust for serial correlation.

Table 4: Results model 1, market risk

Coefficients Standard Error t Stat

Intercept 0.020 0.001 26.470

MR ^{0.010 0.001 9.920}

SD GDP ^{0.000 0.000 0.527}

PI ^{0.010 0.001 11.468}

Table 4 shows that standard deviation of GDP between Israel to USA has an insignificant effect on Israeli market risk. The effect of the Moody's rating is significantly (at 5% level) positive. Therefore, when Moody upgrades Israeli Treasury bond, the market risk increases. This result is in contrast with my expectation. I would suggest, that when Moody upgrade the Israeli Treasury bond, more investors join the stock market, and that can increase the trading activity (higher volume). Higher volume comes with higher volatility, which is exactly what this regression shows. The effect of the political instability periods has a significant (at 5% level) on the market risk. Tension periods lead to higher volatility in the stock market.

Table 5 shows results from running a regression when the dependent variable is the risk premium/default spread between yields to maturity of Israeli Treasury bond and USA Treasury bond for 10 years. The independent variables are Moody's rating to Israeli Treasury bond, the ratio in GDP standard deviation between Israel and USA and the political instability periods as it was mentioned in section 2.2.

For this regression the residuals are significantly auto correlated and the adjusted R² value is 0.24 with 106 observations. It is important to note, that the standard errors in this table are not robust for serial correlation.

Table 5: Results model 1, Default Spread method

Coefficients Standard Error t Stat

Intercept 1.823 0.275 6.635

MR -0.488 0.350 -1.396

SD GDP 0.104 0.055 1.900

PI 1.664 0.308 5.409

Table 5 shows that Moody's rating has an insignificant effect on Israel risk premium.

The effect of the ratio of the standard deviation of GDP between Israel to USA is significantly (at 10% level) positive. Therefore, the volatility of the gross domestic product in Israel has a significantly positive effect on the risk premium, for instance, if the volatility of the GDP is high, it is indeed indicating a non-stable economical environment in the country increasing the risk premium. The effect of the political instability on the risk premium is 1.664%. This means that investors on average require 1.664% more over the risk free return from investing in Israeli Treasury bond in political instable periods comparing to political stable periods. This later effect is statistically significant (at 5% level).

Table 6 shows results from running a regression when the dependent variable is the relative Israeli equity market standard deviation to USA equity market standard deviation (standard deviation of both countries are estimated by the moving average method). The independent variables are Moody's rating to Israeli Treasury bond, the ratio in GDP standard deviation between Israel and USA and the political instability periods as it was mentioned in section 2.2.

For this regression the residuals are significantly auto correlated and the adjusted R² value is 0.11 with 473 observations. It is important to note, that the standard errors in this table are not robust for serial correlation.

Table 6: Results model 1, Relative Equity Market Standard Deviation (moving average) method.

Coefficients Standard Error t Stat

Intercept 1.514 0.048 31.232

MR ^{-0.481 0.061 -7.832}

SD GDP ^{-0.037 0.010 -3.869}

PI ^{-0.083 0.054 -1.536}

Table 6 shows that Moody's rating have a significantly (at 5% level) negative effect on the difference between the standard deviation of Israel equity market and USA market. This result makes sense in a way when Moody upgrade Israeli Treasury bond investors will believe more in the bond that leads to decreasing risk premium. The effect of the ratio in standard deviation in GDP between Israel to USA is significantly (at 5% level) negative. Therefore, the volatility of the gross domestic product in Israel has a significantly negative effect on the risk premium. For instance, if the volatility of the GDP is high, followed by non stable economical environment in the country, the risk premium will decrease. This result is in contrast with our expectations. We expect to see positive effect instead of negative. We have this unexplained result only in the "Relative Equity Market Standard Deviation" method. The effect of the political instability on the difference between the standard deviation of Israel and USA equity markets is insignificantly negative -0.083%. Someone could claim now