Monetary policy announcement and stock market movement
MSc. in Finance, Asset Management Track
Master thesis
Qianci Li
Student Number: 11391219
Supervisor: Natalya Martynova
July, 2017
Faculty of Economics and Business
Amsterdam Business School
Statement of Originality
This document is written by Student Qianci Li who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
This study mainly focuses on the impact of monetary policy announcement on Chinese stock market, while I also expand the sample to eight markets over the period 2006 to 2016. Adopting an event-study methodology, the results reveal a negative relationship between unanticipated monetary policy and stock return. Stock market is sensitive to unexpected monetary policy announcement. On average, a 100 basis points rise in unanticipated monetary policy action is linked with an
approximately 0.73% cut in stock price in a sample of eight countries. Furthermore, business cycle and the level of financial depth play a key role in determining the magnitude of effect from monetary policy shocks on stock market.
Table of Contents
1. Introduction ... 1 2. Literature Review ... 5 Theoretical Papers ... 5 Tobin’s q Theory ... 6 Wealth Effect ... 6 Balance-sheet Effect ... 7Household Liquidity Effect ... 7
Empirical Papers ... 7
Methodology Approach ... 11
3. Methodology ... 14
4. Data and Descriptive Statistics ... 19
Data ... 19 Descriptive Statistics ... 21 5. Results ... 24 6. Robustness Test ... 27 7. Conclusion: ... 28 References ... 29 Appendix ... 33
1
1. Introduction
Chinese stock market was established in 1990s, since then it has experienced a rapid development. Until April, 2017 the total market value of stocks in Shanghai Stock Exchange and Shenzhen Stock Exchange are 30030.272 billion Yuan and 22717.811 billion Yuan, respectively. The percentage of Chinese stock market value in global stock market value has already exceeded Japan in 2011, the second biggest stock market all over the world.
With the bankruptcy of Lehman Brothers, the 2008 Subprime Mortgage Crisis immediately influenced the main financial markets such as Japan and European. Soon the Subprime Mortgage Crisis became a worldwide financial crisis. As one of the most important economic entity in the world, China’s economy is heavily dependent on global trade. The decrease in global demand led to a decrease in economic, therefore China was faced with severe challenge at that time. Due to the lagged effect of transmission mechanism, the influence of the financial crisis on China started to appear in the late 2008. In the third quarter of 2008. Chinese economic growth experienced an accelerated decline. According to Asian Development Outlook, Chinese economic growth decreased from 11.9%, 2007 to 10%, 2008 and will continue decreasing to 9.5% in 2009. The first quarter of 2009 witnessed a 6.1% GDP growth rate, which was the lowest growth rate over the past 10 years. The stock market took a hit as well, with an over 70% decrease from 6124 to 1660 (Shanghai Composite Index). To prop up the economy, Chinese government responded quickly and started to implement monetary policies frequently, including decreasing the reserve-deposit ratio and interest rate (since September, 2008, the central bank of China has cut the reserve ratio five times and the interest rate four times). These measures received a remarkable result. In 2009, China’s GDP increased 9.2% with an index increase over 50%.
Therefore, this thesis studies whether the announcement of monetary policy has a significant impact on Chinese stock market by using ten-year data. I expect that interest rate change is significantly negative related with stock market movement in China. A rise in interest rate results in a decline of expected stock return in Chinese stock market. The relationship between stock market and
monetary policy has drawn great attention of policy makers as well as scholars and been discussed since late 1960. Now it is still one of the hottest topic in economic and finance field. It is broadly accepted that monetary policy announcement has a significant impact on stock market. The
objective of monetary policy is to control money supply and the inflation level. According to Allsopp and Vines (2000), the principle instrument of monetary policy that central bank uses in developed countries is the short-term nominal interest rate.
2
Researchers believe that the change in stock prices can be predicted from the previous change of money supply. The decrease of reserve requirement ratio is associated with an increase of money supply through money multiplier. Numerous studies prove the notion that stock market movement is firmly related to money supply. Sprinkel (1964) compares the breaking points of US stock market index with that of the movement of money supply. His finding suggests that both changes in money supply and stock prices are in accordance with the turning points of business cycle. He draws a conclusion that the speed of stock market reacted to money supply movement is different in bear market and bull market, 15 months and 2 months after the monetary growth, respectively. In theory, the change of supply of money can directly affect the amount of money can be used in stock market. First, the supply of money can affect investor’s expectation about monetary market, which could change the money flow into stock market and then affect the stock return. Second, the increase of money supply breaks the portfolio balance and affects the marginal utility of stock. Investors pursing the maximization of utility will adjust the weight of stock. Thus partial of the money goes to stock market to seek for more returns. Lastly, when the money supply increases, it will decrease and change investors’ expectation of the discount of future cash flow. In the
meantime, enterprises will increase their investment, which will lead to a rise of future cash flows and then improve stocks’ internal value.
According to the standard view, stock price reflects investor’s expectation about a firm’s future performance. Base on discounted cash flow model, stock price is equal to the present value of future cash flows. Keran (1971) points out that the appropriate discount rate in this model is interest rate. Therefore, increasing interest rate means a higher discount rate, which could lead to a decline in stock price. Under expansive monetary policy, central bank cuts interest rate, which leads to a rise of stock price. In the meanwhile, a drop in interest rate implies the financing cost of a firm declines which could increase expected cash flow. In general, an expansive monetary policy can lead investors to view stocks as less risker assets and thereby require a lower return. Therefore, stock price reacts negatively to monetary policy.
According to efficient market hypothesis, all related information have been incorporated into asset price; market is relatively efficient. Roberts divided market into three different forms: “weak,” “semi-strong,” and “strong”. China, as the second largest stock markets in the world, has developed rapidly these years. However, there are still some problems exposed such as the lack of construction of the soft environment, corresponding laws and regulations. Therefore, China’s stock market is considered as a weak form market, which holds that stock price cannot reflect all publicly available information instantly. Thus I expect there will exist time jag between the monetary policy
3
announcements and the stock market movement. This paper will use the event study methodology to examine the one-day response of the stock market measured by main national stock index to the changes in monetary policy announcements proxied by policy rate from 2006 to 2016. Following Kuttner (2001), I decompose monetary policy announcement into unanticipated and anticipated element. The unanticipated monetary policy decision can derivate from the change of market interest rate, in this context, the three-month interbank offered rate is employed.
In general, market conditions are different between developing and developed countries. However, most of the current studies focus on how monetary policy affect stock return in advanced economies such as US and European Countries. There is a lack of literature in emerging markets. This paper attempts to examine the response of Chinese stock market to unanticipated monetary policy decisions. The study of China, the largest emerging market in the world, may contribute to fill the literature gap. I add other seven countries to compare as well, covering emerging markets as well as the developed economies. The difference in development of financial system among different countries is taken into consideration. Besides, the release of monetary policy in China is very different from that of other countries. In China, the monetary policy is suddenly announced without any official schedules, while in other countries such as the U.S., the announcement of monetary policy follows a pre-determined schedule of Federal Open Market Committee (FOMC) meetings. With this research, I hope the conclusions may help monetary policymakers to have an idea up to what extent their decisions could influence stock market and formulate effective policy decisions. Besides, from the perspective of market participants, the study may help them with more precise estimation of the stock market reaction to monetary policy and then more effective asset allocation strategy can be made.
My thesis contributes to the current literature from several aspects. First, limited studies focus on emerging market, while most literatures focus on developed markets, especially in the U.S. My thesis makes a contribution to fill the literature gap. Besides, current papers which conduct a
multi-countries analysis mainly investigate the heterogeneity stock market reaction to federal rate change. Although this is important with the economic globalization trend, it is also necessary to study how stock market reacts to domestic monetary policy announcements. Papers related to the connection between Chinese stock market and Chinese monetary policy fail to isolate the unexpected part from real monetary policy announcement. As mentioned by Kuttner (2001). The expected monetary policy decision has already been priced, stock market only reacts to monetary policy shocks. Therefore, decomposing monetary policy announcements is essential when study the relationship.
4
Furthermore, I also examine whether the degree of financial market development (the depth of financial market) and business cycle influence the monetary policy transmission efficiency. This thesis provides empirical evidence with respect to the relationship between stock market movement and monetary policy announcement. I adopt an event-study methodology in eight countries including China, Thailand, Indonesia, Malaysia, Mauritius, South Korea, Norway and UK during the period 2006-2016. The empirical results are in line with previous literatures. On average, a 1% rise in unanticipated policy rate change can cause a 0.73% cut in stock price in sample
countries. The magnitude of the effect from monetary policy actions on stock market varies from business cycle. The effect is strengthened during financial crisis period. Financial depth also plays an important role in determining the magnitude. The deeper of a country’s financial system, the weaker of the monetary policy influence on stock market.
This paper is structured as follows. Section 2 discusses the current literature of related field, including both theoretical and empirical paper and the most adopted methodology in this topic. Section 3 describes the methodology and hypothesis I make. Section 4 is about the data selection criteria as well as some descriptive statistics. Section 5 presents the result and related. Section 6 focuses on the robustness check of the result. Section 7 makes a brief summary of this paper.
5
2. Literature Review
Theoretical Papers
First I would like to give a brief overview of the theoretical basis about how monetary policy affects stock market. The quantity theory of money holds the assumption that money supply is exogenous and consumer’s preference remains constant in short term. Variables which could influence money supply are independent with variables that affect money demand. Therefore, central bank can influence money supply by implementing monetary policy such as adjusting the reserve requirement ratio. When the money supply increases, there will be surplus in the quantity of money and will increase the opportunity cost of holding currency. It will encourage investor to buy more shares. With more capital flows into security market, the stock price decreases. Change in money supply can affect investor’s sentiment of holding stock asset.
Reserve requirement ratio money supply surplus in money opportunity cost of holding currency investor’s desire toward stocks stock price
As for the Keynes’s theory of money demand, he brings up with the liquidity preference theory, a preference that people would rather sacrifice the interest income to hold the currency. He sums up three motives of demand for liquidity: the transaction motive, the precautionary motive and speculative motive. For the speculative motive, on the one hand, assume the money supply holds constant, the decrease of interest rate could decrease the discount rate, lead to an increase of stock price. On the other hand, the financing cost would decline and therefore corporate could invest more and expand their productivity to improve their profitability, which would increase the net value of a firm then the stock price goes up.
Policy rate discount rate stock price
Policy rate financing cost firm investment profit firm value stock price
In regard to the monetary policy transmission mechanism, it exhibits different mechanism during different period. Conventional monetary policies take commercial bank as intermediate to influence substantial economy through credit channels, interest rate channels and exchange rate channels. Recently, with the development of capital market, it holds a central position in economy. The
6
conventional monetary policy transmission mechanism is weakening while in the meantime, stock market increasingly shows its power in monetary policy transmission process. In the very beginning,
Rozeff and Michael (1974) come up with three different channels which are the liquidity effect, the
earnings effect and the risk premium. Then Mishkin (1996, 2001) presents an overview of the transmission mechanisms of monetary policy, both conventional channels and unconventional channels. He summarizes four different channels in transmission mechanism, which involve Tobin’s q Theory, wealth effect, balance-sheet effect and household liquidity effect. In 1999, Ralph Chami makes a supplement to transmission mechanism from the aspect of inflation. Mishkin (2010) further improves the transmission mechanism by dividing it into two basic categories: neoclassical channels and non-neoclassical channels.
Among all the channels, Tobin’s q Theory and balance-sheet effect explain the mechanism through how money supply affect investment, while wealth effects and household liquidity effect explain this through consumption.
Tobin’s q Theory
Tobin’s q is the ratio that firm’s market value divided by its replacement cost and it is commonly regarded as an important measurement of firm’s performance or future growth rate. When q > 1, market value is greater than replacement cost, which indicates the cost of physical investment is relatively cheaper. Firms could release stocks with higher price and are encouraged to invest more in capital, vice versa. When central bank implements loose monetary policy, the increase of money supply leads to an increase of stock price, followed by a rise of Tobin’s q. Investment expense and aggregate social output will go up as well, which results in the increase of national income. As shown in below:
Money supply stock price q investment expenditure aggregate output
Wealth Effect
Consumption expenditure depends on the wealth an investor owns, including human capital, financial asset and fixed asset. Stock is the most important component of financial asset. When central bank implements expansionary monetary policy, increased interest rates affect the discount rate in the calculation of stock price. The stock price goes up. Investor would expect a higher stock return and income, thus increases consumption.
7
Balance-sheet Effect
There exists information asymmetry in the imperfect competitive market. Adverse selection occurs when a company presents a poor balance sheet. Company owns more information than the bank (lender). The lower of the firm’s net value, the more likely the moral risk arises. At this point, firms tend to invest in high-risk project. Once the project fails, it would be really difficult for the bank to recover loans. When implementing restrictive monetary policy, firms need to pay more interest because of the increase in interest rate. When stock price goes down, the net value of those firms decline with it. The adverse selection arises and banks reduce the loan. In the end the investment and output are affected.
Money supply stock price net value loan investment expenditure aggregate output
Household Liquidity Effect
Liquidity theory holds the view that except wealth, the liquidity of asset one owns can influence the total consumption expenditure as well. The higher of the liquidity, the less likely a family gets into financial distress. When consumers expect a higher future income, they prefer to hold more fixed asset and durable consumer goods. On the contrary, when they predict they may get into financial distress in the future, they would hold more liquidity asset such as stocks. Therefore, when stock price goes up, consumers hold comparatively more liquidity asset and thus increase investment and consumption.
Money supply stock price liquidity asset probability of financial distress consumption expenditure aggregate output
Empirical Papers
There is an extensive empirical literature discussing the response of stock market to monetary policy shocks. Relationship between stock market movement and monetary policy announcement has been examined by many researchers (Sprinkel, 1964; Homa and Jaffee, 1971; Bernanke and Kuttner, 2005). Related literatures grow up during the past twenty years; most papers (Jenson and Johnson, 1996; Reinhart and Simin, 1997; Ehrmann and Fratzcher, 2004; Kontonikas and Kostakis, 2013) focus on U.S. stock market to evaluate how the monetary policy announcements influence stock returns, while there are only a very few papers focus on other countries. (Guray et al. , 2013; Bredin and
8
Hyde, 2007). Previous empirical evidence broadly supports the notion that monetary policy action is associated with an inverse movement of stock market.
A lot of researchers focus on explaining stock market movement through economic and monetary elements. A study by Rozeff (1974) reveals that stock market movement is linked with simultaneous monetary policy change. In 1988, Fama and French test the predictable power of economic factors in stock return. Their results suggest that dividend yield, default spread and the term spread can help to forecast stock market movement. They further prove that business condition can also help to explain the stock returns in the following year. Their work is extended by Jensen, Mercer and Johnson (1996) by dividing the market into two regimes, expansionary regimes and restrictive regimes. They explore the long term stock market reaction to business conditions under different monetary environment from February 1954 to December 1992 in the U.S. Their research reveal that business conditions can influence stock market differently. Stock returns show a higher correlation with business conditions under restrictive regime than expansionary regime. Kurov (2010) also divides the market into different regimes, bull and bear regime and gets the similar result as Jensen, Mercer and Johnson (1996). He finds that under bear market, corresponding to the restrictive regime in Jensen et al. (1996)’s paper, the stock market shows a stronger response to the monetary policy actions as a result of investor sentiment shift. JR Booth and Lena (1997) believe that firm size matters in determining the monetary policy effect on stock return, that smaller companies are firmly dependent on bank financing thus are more easily influenced by monetary policy shocks. Their study presents a similar result as Jensen and Janson (1996). They also introduce two measurement of monetary policy actions in the same paper. One is the directional change of the discount rate proposed by Jensen, Mercer and Johnson (1996), another one is federal funds rate. Their findings reveal both measurements exhibit a strong explanatory power for stock return. Therefore, when testing the stock market responsiveness to monetary policy actions, scholars mainly choose policy rate change (e.g. federal funds rate in US) as a proxy for monetary policy.
Recent empirical papers confirm these results from two aspects, interest rate and money supply. For example, Ehrmann and Fratzscher (2004), Farka (2009) and Bjørnland and Leitemo (2009) all study how interest rate affects US stock market. But the final result is slightly quantitative different. Ehrmann and Fratzscher(2004), using credit and interest rate as measures, show that an unexpected 50 basis points cut in U.S. policy rate can rise stock price by 3% on announcement day. Bernanke and Kuttner (2005) find a 25 basis points rise in unanticipated policy rate target is estimated to decrease broad stock indexes by around 1% immediately. A research by Bjørnland and Leitemo (2009) shows monetary policy rate interdepend with stock market. With a 10 basis points tightening in monetary
9
policy, stock price falls by around 2%. While most researchers use daily stock price, Ioannidis and Kontonikas (2008) focus on monthly data. They conduct a multi-nation analysis on 13 OECD countries from 1972 to 2002 to assess the relationship between monetary policy and stock market movement. Their finding shows that stock return is firmly associated with monetary policy. Using a different method, accounting for endogeneity and omitted variable bias, analyzing intra-day data, Farka (2009) finds that an unanticipated 1 % increase of the FED’s target rate results in a decline of 5,6 % in the stock market. All those research results indicate that monetary policy decisions have a significant negative influence on stock price.
To better understand the monetary policy transmission mechanism via stock market, in 2002, Poole, Rasche and Thornton examine market anticipation of monetary policy. By calculating the changes of bank bill rate with different maturities under monetary policy shifts with a sample period over 1988-1999, they find that market participants can predict the policy decision better in recent years. One can be explained by the fact that Federal Reserve Bank of the United States started to apply forward guidance to implement monetary policy in 1994. In 2010, Kurov comes up with an interesting idea to emphasize why monetary policy can affect stock market. He finds evidence that monetary policy announcements can significantly influence investor sentiment, which will influence stock price in the end.
According to EMH, all publicly available information is already incorporated in the share price of a stock. Assuming semi-strong efficiency, the market should adjust the stock price instantaneously when new information emerges and disregard news that is highly anticipated. An important
distinction in the literature is therefore that of anticipated and unanticipated policy changes. Kuttner (2001) first introduces the approach to decompose the monetary policy into surprise component and nonsurprise component. He finds a strong and reliable connection between monetary policy surprise and market interest rate, therefore he constructs the surprise component by taking the time-weighted difference between the futures fund rate on the announcement day and the day prior to the change. Calculation equation is shown as below:
∆𝐼𝐼𝑡𝑡𝑢𝑢=𝐷𝐷 − 𝑑𝑑 (𝑓𝑓𝐷𝐷 𝑑𝑑0− 𝑓𝑓𝑑𝑑−10 )
Where 𝐷𝐷 represents the number of days in the month 𝑑𝑑 represents the announcement day in the month
10
Kuttner (2001) applies this method in bond yield and finds bond yield reaction to unexpected component in policy rate is significant. After that, Kuttner and Bernanke (2005) apply the same methodology in the stock market. They study both the whole equity market and industry portfolio’s immediate response to monetary policy and find a significant stock market reaction to the
unanticipated monetary policy actions. With this new measurement their finding shows an unexpected 25 basis point decline of federal funds rate is linked with an approximately 1% rise in stock prices. Unexpected monetary policy announcement shows the strongest explanatory power in stock market reaction. In a similar vein, Ehrmann and Fratzscher (2004)’s study reveals that an increase of 50 basis points in federal funds rate leads to a 3% decline in stock return and the reaction is stronger when there is no anticipated change during the period 1994-2003. The impact of
monetary policy surprise on stock market shows asymmetry and heterogeneity. Bredin, Nitzsche and Hyde (2007) adopt the measurement proposed by Bernanke and Kuttner (2005) and apply in UK market. They use change in 3-month sterling LIBOR futures as a substitution for surprise monetary policy action and find a qualitatively similar result as Bernanke and Kuttner (2005). A problem with this approach is that there are difficulties in determining to what degree a policy change is
anticipated or not. Some researches therefore make the assumption that all policy changes are unanticipated. This is the assumption made in a study conducted by Stevenson (2002) on the German stock market, analyzing bank stocks.
With respect to the measurement of unanticipated monetary policy announcement, for countries outside US, scholars then propose several different substitutions of federal funds futures such as 90-day treasury bills, interbank offered rate, bank acceptance, etc. Roley and Sellon (1995) examine the relationship between long-term interest rate and monetary policy actions and find long-term rate can predict future policy action. Therefore long-term rate is a good proxy for anticipated policy rate change. Correspondingly, the short-term interest rate can be seen as a measurement of
unanticipated policy rate change. Gürkaynak and Sack (2007) compare 6 different instruments in measuring expected monetary policy. Their finding proves that federal funds futures shows the best performance in forecasting policy rate with both a long-term period and short-term period.
However, for policy announcement interval longer than six weeks, other market interests rates present similar predict power as federal funds futures.
In regard to multi-nation analysis, Ehrmann and Fratzscher (2009) expand their sample into 50 countries worldwide based on their previous study in 2004. They analyze the heterogeneity in responding to federal rate change among Australia, Canada and other fifty countries, covering both developing countries and developed countries. Their result shows, on average, an 1% increase in US
11
federal funds rates lead to an approximately 3.8% decline in global stock market. There exists the cross-country as well as the cross-sector heterogeneity. Stock market in China, for example, hardly response to US monetary policy action, while in other countries such as Korea and Sweden, the stock markets change significantly. The degree of financial openness in a country plays an important role in the transmission channels efficiency. As for the cross-sector heterogeneity, information
technology sector shows the strongest response among all sectors. Similar study conducted inside in European area by Giulia (2012) investigates the stock market reaction to ECB monetary policy announcements, from conventional and unconventional aspects. Ioannidis and Kontonikas (2008) study the stock market response to monetary policy shocks in 13 OECD countries from 1972 to 2002. Their finding proves the notion that stock market is negatively related with monetary policy. One important contribution of their work is the co-movement among international stock markets is taken into the regression model. However, one drawback of this paper is they use the monthly data while employing the event-study method. Low data frequency may result in endogeneity and omitted variable bias. Ricci (2015) focuses on the reaction of bank industry to ECB monetary policy actions. He divides the sample period into conventional period and unconventional period from June 2007 to June 2013. The result shows bank sectors are more sensitive to unconventional monetary policy shocks. Large banks in European area not only answer to ECB monetary policy announcements, but also to federal funds rate change.
Methodology Approach
A great number of researches have been conducted to evaluate the connection between monetary policy shocks and stock returns and this relationship has been identified. The empirical results show that stock market reacts negatively under restrictive monetary policy and positively under loose monetary policy. Many empirical methods have been used to analyze this issues, including event study and VAR method.
When assessing the relationship between monetary policy announcements and stock market reaction, early literatures mainly use regression model taking money supply shifts as the proxy for monetary policy actions. However, pointed by Nessen, Sellin and Sdberg (2001), the results can be affected by the endogenous problem. According to Rigobon and Sack (2003), the movement of stock price can also drive policy makers to modify their policy decisions. In this case, the event study approach is performed which looks at the change in stock return on announcement day. In 1989, Cook and Hahn first utilize an event study framework to examine the immediate response of bond to monetary policy. By using the bond return on announcement day, they find monetary policy shift is
12
significantly connected with bond market. The research of Thorbecke (1997) proves that
expansionary money policy increases stock prices by employing event study approach. Kutter (2001) makes further improvement on Thorbecke (1997) by isolating the unanticipated monetary policy from raw monetary policy. Alexander Kurov (2008) follows Kuttner (2001) to compute the unanticipated element of federal monetary policy. Their finding is in line with previous research. Furthermore, their finding suggests that the effect of federal funds rate shift on stock market is stronger in bear market. Bredin, Hyde and Nitzsche (2007) and Gregoriou and Kontonikas. (2009) apply the same measurement constructed by Bernanke and Kuttner (2005) in UK stock market and get similar results. With the same methodology, the strong negative impact from interest rate change on stock price in Pakistan is proved by Rahman and Mohsin (2011). Bomfim and Reinhart (2002), Ehrmann and Fratzscher (2004), Corallo (2006) all adopt the event study methodology to test the monetary policy influence on stock market.
Rigbon and Sack (2004) offer a new estimator to avoid the endogeneity and omitted-variable issue raising up in the related field. This estimator is based on the heteroscedasticity existed in high-frequency data and is an alternative to the event study approach which requires relatively weaker assumptions. Their study shows that the outcome based on event-study contains bias that make the impact smaller. Their study indicates that a 25 basis point rise in federal funds rate leads to a 1.7% decline in the S&P 500 based on US data from 3rd January, 1994 to 26th November, 2001, very similar
as Ehrmann and Fratzscher (2004). Ehrmann (2011) and Bohl and Siklos (2008) utilize the
heteroscedasticity-based approach and find similar results with Rigbon and Sack (2004) in European markets.
Another most commonly adopted estimation method is the VAR, Vector Autoregression method. Since Sims (1980) proposed this model, it has been widely used for empirical research in
macroeconomics. In VAR models, there is no need to specify endogenous and exogenous variables and flexible dynamics are allowed. Geske and Roll (1983) first examine the casual relationship between monetary policy and stock returns using VAR. Thorbecke (1997) and Mihov (1998) follow the same methodology to study monetary policy transmission. Thorbecke (1997) finds that stock price goes up under expansionary monetary policy. Kontonikas and Kostakis (2013) employ the VAR framework to examine if the characteristics of stock portfolios can influence the outcome of
unanticipated monetary policy, using US data through 1967 for 2007. Their finding is consistent with Booth and Booth (1997), that small capitalization stocks are more sensitive to monetary policy surprises.
13
Rudebusch (1998) questions the reliability of VAR approach. He argues that this approach is deficient especially when taking into account structural benchmarks. Using VAR to evaluate the relation between stock price and monetary policy announcements cannot capture information provided by central bank in time. VAR heavily relies on data frequency and thus the results are unreliable and fragile. He believes the VAR equation is not a good model to reflect the real federal monetary policy and the residuals from this regression show little correlation with monetary policy shocks.
14
3. Methodology
Previous literatures, both theoretical and empirical, have shown that monetary policy has a significantly influence on stock price movement. But measuring the stock market response to
monetary policy announcement is difficult with the endogeneity problem and omitted variables bias. The movement of stock price can also drive policy makers to modify their policy decision (Rigobon and Sack, 2003). To overcome these problems, Cook and Haan (1989), Bernanke and Kuttner (2005), and Hausman and Wongswan (2006) all employ the event study method, which is the most widely adopted method in this field. Therefore, to dissolve the endogeneity problem, I adopt the same event study approach to conduct my research, testing the stock market response to monetary policy shocks. Using a narrow event window might eliminate the omitted variable bias. In this paper, I choose interest rate as the indicator of monetary policy.
One benefit of using event study is that the estimation results are more accurate by using high frequency data compared with other approaches. The employ of this method only focuses on a short period after the announcement day, therefore it can remove the endogeneity bias at the most extent. For other methods applied in this field, they choose weekly or monthly data, which is more difficult to obtain a precise estimate because of data frequency.
I adopt several regression models in this paper. The baseline regression model, first proposed by Cook and Hahn (1989), examines the one-day stock market reaction to the changes in policy rate.
𝑅𝑅𝑡𝑡 = 𝛼𝛼 + 𝛽𝛽∆𝐼𝐼𝑡𝑡+ 𝜀𝜀
Where 𝑅𝑅𝑡𝑡 represents the stock return after the announcement
𝜀𝜀 represents other factors that influence stock return after announcement
It should be mentioned that in China, The PBC releases monetary policy decisions after the stock market closed. Therefore, the regression model above is modified as follows:
𝑅𝑅𝑡𝑡+1 = 𝛼𝛼 + 𝛽𝛽∆𝐼𝐼𝑡𝑡+ 𝜀𝜀
Where 𝑅𝑅𝑡𝑡+1 represents the stock return on the next transaction day after the announcement
The first hypothesis test is therefore: Hypothesis 1
15
Null hypothesis: The monetary policy announcements show significant negative impact on Chinese stock market.
Alternative hypothesis: The monetary policy announcements show insignificant influence on Chinese stock market.
In the spirit of Kuttner (2001), the second regression model used here is to measure the stock market reaction to unanticipated monetary policy decisions. According to efficient market
hypothesis, which the event study relies firmly on, all related information has been incorporated into the asset price. Investor can obtain information from prior press conference and unofficial
conversation and therefore forming ideas about the interest rate changes in future. It is unlikely for stock market to react to expected monetary policy. Therefore, the stock market should only
response to the unanticipated monetary policy shocks. It is essential to divide monetary policy into surprise and nonsurprise component.
When decomposing monetary policy, most literatures follow Kuttner (2001) by taking the difference of federal funds futures rate, scaled up by time. Due to the fact that China first released government bond futures in September, 2013, I follow Gürkaynak and Sack (2007) when measure the unexpected element by calculating the difference between the market interest rate on the day and prior day of announcement. This behavior can avoid endogeneity and simultaneity issues to the greatest extent. The change of interest rate can be written as follows:
∆𝐼𝐼𝑡𝑡 = 𝐼𝐼𝑡𝑡− 𝐼𝐼𝑡𝑡−1 = (𝐸𝐸𝑡𝑡−1(𝐼𝐼𝑡𝑡) − 𝐼𝐼𝑡𝑡−1) + (𝐼𝐼𝑡𝑡− 𝐸𝐸𝑡𝑡−1(𝐼𝐼𝑡𝑡)) = ∆𝐼𝐼𝑡𝑡𝑒𝑒+ ∆𝐼𝐼𝑡𝑡+1𝑢𝑢
Where ∆𝐼𝐼𝑡𝑡 denotes the change of interest rate
∆𝐼𝐼𝑡𝑡𝑒𝑒 denotes the expected element of the monetary policy
∆𝐼𝐼𝑡𝑡𝑢𝑢 denotes the unexpected element of monetary policy
𝐸𝐸𝑡𝑡−1(𝐼𝐼𝑡𝑡) denotes the expected interest rate change
Considering the special situation that Chinese monetary policy is released after the stock market closed, the unanticipated element of monetary policy should be calculated based on the following day of announcement as well. Therefore, the second regression model used here would be:
𝑅𝑅𝑡𝑡+1 = 𝛼𝛼 + 𝛽𝛽𝑒𝑒∆𝐼𝐼𝑡𝑡𝑒𝑒+ 𝛽𝛽𝑢𝑢∆𝐼𝐼𝑡𝑡+1𝑢𝑢 + 𝜀𝜀
16 The corresponding hypothesis is as follows,
Hypothesis 2
Null hypothesis: Chinese stock market reacts significantly to both anticipated and unanticipated interest rate change.
Alternative hypothesis: Chinese stock market only reacts significantly to unanticipated interest rate change.
While other papers only focus on the one-day movement of stock market, I decide to investigate the one-day, two-day, three-day, four-day and five-day response respectively. Unlike the developed markets, China, as the biggest emerging stock market in the world, has been considered as a weak-form efficient market, proved by Laurence, Cai and Qian (1997).
According to EMH, weak-form effective market fails to react immediately to publicly available information. Thus, Chinese stock market will not reflect the monetary policy shocks promptly. It takes times to fully digest the public news, in this case, the monetary policy decisions. The corresponding regression model would be:
𝑅𝑅1𝑑𝑑,2𝑑𝑑,3𝑑𝑑,4𝑑𝑑,5𝑑𝑑= 𝛼𝛼 + 𝛽𝛽𝑒𝑒∆𝐼𝐼𝑡𝑡𝑒𝑒+ 𝛽𝛽𝑢𝑢∆𝐼𝐼𝑡𝑡+1𝑢𝑢 + 𝜀𝜀
where 𝑅𝑅1𝑑𝑑,2𝑑𝑑,3𝑑𝑑,4𝑑𝑑,5𝑑𝑑 stands for the one-day, two-day, three-day, four-day and five-day reaction
respectively.
Thus, my third hypothesis would be: Hypothesis 3
Null hypothesis: Chinese stock market responses promptly to the monetary policy actions on the announcement day.
Alternative hypothesis: Wider event window performs better in explaining the stock market reaction to monetary policy announcements.
My main study object is China. However, the existing heterogeneity makes me examine the heterogeneous response across countries. Therefore, I expand the sample country by adding Thailand, Malaysia, Indonesia, Korea, Norway, Mauritius and UK, including both developed
economies and emerging markets. Similar regression models are used as before, except whether to use the lag return and lag market interest rate depends on announcing time, varies cross countries.
17
Scholars believe stock market movement is also influenced by the economy condition. A growing body of research supports the view that stock market shows a stronger response to monetary policy announcement during restrictive regimes. (Jensen and Johnson, 1996; Kurov, 2010). Thus I consider the effect of business cycle on stock market reaction to monetary policy decisions. To test if the influence is different during financial crisis, I add financial crisis as a dummy variable into the regression model and interact crisis dummy with surprise policy. The 2008 financial crisis period is defined as from December, 2007 to June, 2009, given by official website of NBER. If the
announcement was released during this period, the dummy variable equals to one, otherwise zero. 𝑅𝑅𝑖𝑖.𝑡𝑡+1= 𝛼𝛼𝑖𝑖+ 𝛽𝛽𝑖𝑖𝑒𝑒𝛽𝛽𝑒𝑒∆𝐼𝐼𝑖𝑖,𝑡𝑡𝑒𝑒 + 𝛽𝛽𝑖𝑖𝑢𝑢∆𝐼𝐼𝑖𝑖.𝑡𝑡𝑢𝑢 + 𝛽𝛽𝑖𝑖.𝑑𝑑𝑢𝑢 𝐷𝐷𝑐𝑐𝑐𝑐𝑖𝑖𝑐𝑐𝑖𝑖𝑐𝑐∗ ∆𝐼𝐼𝑖𝑖,𝑡𝑡𝑢𝑢 + 𝛽𝛽𝑑𝑑𝐷𝐷𝑐𝑐𝑐𝑐𝑖𝑖𝑐𝑐𝑖𝑖𝑐𝑐+ 𝜀𝜀𝑖𝑖
where 𝐷𝐷𝑐𝑐𝑐𝑐𝑖𝑖𝑐𝑐𝑖𝑖𝑐𝑐 is the dummy variable of financial crisis
𝛽𝛽𝑖𝑖.𝑑𝑑𝑢𝑢 is the parameter of interest
Hypothesis about the impact of financial crisis is as follows: Hypothesis 4
Null hypothesis: Financial Crisis does not affect the effectiveness of monetary policy transmission to stock market.
Alternative hypothesis: Financial Crisis leads to a stronger movement of stock market influenced by monetary policy shocks.
Another important issue is that whether the stock price movement depends on a country’s financial depth. Financial depth refers to the development of a country’s financial market. The regression model employed to test the influence of financial market depth would be:
𝑅𝑅𝑖𝑖,𝑡𝑡 = 𝛼𝛼𝑖𝑖+ 𝛽𝛽𝑖𝑖𝑒𝑒∆𝐼𝐼𝑖𝑖,𝑡𝑡𝑒𝑒 + 𝛽𝛽𝑖𝑖𝑢𝑢∆𝐼𝐼𝑖𝑖,𝑡𝑡𝑢𝑢 + 𝛾𝛾𝑖𝑖𝐹𝐹𝐷𝐷𝑖𝑖,𝑡𝑡∗ ∆𝐼𝐼𝑖𝑖,𝑡𝑡𝑢𝑢 + 𝛿𝛿𝑖𝑖𝐹𝐹𝐷𝐷𝑖𝑖,𝑡𝑡+ 𝜀𝜀𝑖𝑖
Where 𝑅𝑅𝑖𝑖,𝑡𝑡 denotes the percentage change of stock price index in country i either in announcement
day or the day after announcement day depends on the specific time. ∆𝐼𝐼𝑖𝑖,𝑡𝑡𝑒𝑒 and ∆𝐼𝐼
𝑖𝑖,𝑡𝑡𝑢𝑢 denote the expected and unexpected monetary policy element in country i,
respectively.
18 𝛾𝛾𝑖𝑖 denotes the parameter of interest
Official website of World Bank provides an indicator of financial depth, which equals to private credit relative to gross domestic product (GDP).
𝐹𝐹𝐷𝐷𝑖𝑖,𝑡𝑡 =𝐺𝐺𝐷𝐷𝑃𝑃𝑃𝑃𝑃𝑃𝑖𝑖,𝑡𝑡 𝑖𝑖,𝑡𝑡
Where 𝑃𝑃𝑃𝑃𝑖𝑖,𝑡𝑡 represents private credit of Country i at time t and 𝐺𝐺𝐷𝐷𝑃𝑃𝑖𝑖,𝑡𝑡 represents GDP of country i at
time t.
Hence, the last hypothesis would be: Hypothesis 5
Null hypothesis: The response of emerging stock market to unanticipated monetary policy actions has no relationship with financial depth.
Alternative hypothesis: Financial depth has a significant influence on stock market reaction to anticipated monetary policy.
19
4. Data and Descriptive Statistics
Data
The sample period runs for January, 2006 through December, 2016, including 2008 global financial crisis. Main research object is China. For further comparison, I expand the country of interest including Thailand, Malaysia, Indonesia, Mauritius, Korea, Norway, UK, covering emerging markets as well as the developed economies. Due to time constrains, I cannot include all countries into this research. Therefore, I randomly pick sample countries based on the list offered by the official website of World Bank.
All data is available from Datastream, official website of each country’s central bank and official website of World Bank. The data sources used in this paper are summarized in table 2. To make sure the data is reliable and accurate, I cross compare to check data from different resources and make further processing to conform data.
Reaction to monetary policy is examined with respect to stock return. Currently, the main stock indexes in China include Shanghai Stock Exchange (SSE) Composite Index, Shenzhen Stock Exchange (SZSE) Composite Index and China Securities Index (CSI) 300. Among them, SSE Composite Index is the most comprehensive and familiar to people. Besides, Shanghai Stock Exchange was set up earlier than Shenzhen Stock Exchange and has a comparable high market value. It is more sensitive to different changes. Most studies related to Chinese stock market employ SSE Composite Index as their research object. Therefore, I use the return of SSE Composite Index as the proxy for Chinese stock market reaction. Among previous literatures, Cook and Hahn (1989) and Rigobon and Sack (2004) apply the Federal Funds rate as the measurement of Federal Open Market Committee (FOMC) monetary policy. In this context, Chinese monetary policy is proxied by interest rate change. To be more specific, the benchmark interest rates include lending rate and deposit rate. I use China Lending Rate 6M to 1Y (hereafter referred to as interest rate) as the policy rate. The sample period covers 24 times of interest rate change in China. The first announcement to change the interest rate was on 17th March, 2007. PBC announced a 0.27% increase in interest rate. The announcements are not always released in workdays and are released after the stock market closed. Therefore, when calculating the stock market reaction, I use record of next transaction day. Most current literatures follow Kuttner (2001), Bernanke and Kuttner (2005) when calculating the expected monetary policy. This approach calculates the time weighted difference between federal funds futures rate on the day of announcement and prior day of announcement. However, Chinese government first released the government funds futures in 6th September, 2013. To obtain as much samples as I can, making the
20
outcome more reliable, I adopt another measurement of monetary policy surprise by taking the difference between the market interest rate on the day and prior day of announcement. As for the market interest rate, 3-month SHIBOR (Shanghai Interbank Offered Rate) is applied. Then subtract it from real interest change to get the nonsurprise element.
In Thailand, the main stock index is Stock Exchange of Thailand Index (SET Index). It is important to note front that Bank of Thailand (BOT) changed the monetary policy instrument from 14-day repo rate to 1-day repo rate in January 2007. Bangkok Interbank Offered Rate (BIBOR) is used as measurement of unanticipated policy rate change.
In Malaysia, the main stock index is FTSE Bursa Malaysia Kuala Lumpur Composite Index (FBMKLCI) and policy interest rate is the overnight interest rate set by Bank Negara Malaysia (BNM). Kuala Lumpur Interbank Offered Rate (KLIBOR) is the proxy for market interest rate.
In Indonesia, the Bank Indonesia rate is applied as the policy interest rate and main stock index is Jakarta Composite Index (JCI). As other countries, the Jakarta Interbank Offered Rate (JIBOR) is used. Monetary policy decision is announced 1pm local time after the submit of JIBOR. Therefore, the JIBOR change needs to be lagged by one day.
The main stock index in Mauritius is SEMDEX. According to the official website of Bank of Mauritius (BOM), the monetary policy decision is released in 3pm local time while the stock market closes at 1:30pm. Therefore, the percentage of stock return is based on next transaction day. The policy interest rate in Mauritius is key repo rate collected from BOM. As for the market interest rate, Port Louis Interbank Offered Rate (PLIBOR) is used, computed by Thomson Reuters at 12am local time. Therefore, when calculating the surprise policy rate change, one-day lag need to be taken. Another thing need to be mentioned here that no data available for Mauritius PLIBOR before 5th April, 2011.
Therefore, monetary policy decisions before that will be taken as fully surprise.
Besides for the emerging market, I also include South Korea, United Kingdom and Norway as my research object to see if the outcome is different in developed economies. The main stock index in South Korea is Korea Composite Stock Price Index (KOSPI). The indicator of monetary policy is call rate before March 2008, and after that, it becomes Bank of Korea base rate. Still, the interbank offered rate in Korea, SIBOR, is used as the market interest rate. Monetary policy decisions are released around 10:30 am local time. Policy rate and main stock index in UK is bank rate and FTSE 100 Index, respectively. The market interest rate used here is three month LIBOR. As for Norway, the policy rate is key policy rate released at 10am. Main stock index in Norway is Oslo Stock Exchange
21
(OBX) index and the market interest rate used here is NIBOR (Norway Interbank Offered Rate). The fix time for NIBOR is 12 noon CET each trading day therefore no lag in NIBOR is needed.
With respect to the influence of financial crisis, the 2008 global financial crisis period is defined as from December 2007 to June 2009 by the official website of the National Bureau of Economic Research (NBER). If the announcement was released during this period, the dummy variable equals to one, otherwise zero. Quarterly data of financial market depth is available from Datastream. However, for Mauritius only annual data is provided. Considering the special situation, I assume Mauritius financial depth increases at a constant speed each quarter.
Descriptive Statistics
Figure 1 and figure 2 describes the characteristics of SSE Composite Index. The distribution of SSE CI is left skewed and leptokurtic. Figure 2 illustrates the high volatility clustering of stock return in China especially between the end of 2007 to 2009, falling into the category of crisis period. Same feature of other countries can be seen from figure 3. Figure 4 shows the change in 3-month interbank offered rate in all eight countries and exhibits the same volatility during crisis period. It saw a rapid increase in the beginning of financial crisis, reached the peak around July 2008, then declined until June 2009 and remained relatively stable until now.
From the monetary policy surprises distribution of China in figure 5 and figure 6, it is clearly that there is one outlier among all surprises, which is the monetary policy announcement in 26th
November, 2008 when People’s Bank of China decided to decrease the interest rate by 0.27%. The appearance of this outlier may be affected by the 2008 Subprime Crisis in America. Besides, from figure 2, the stock return in China exhibits high volatility during the financial crisis. Thus, to get a better estimation, I further add financial crisis as dummy variable to investigate whether business cycle will influence the strength of policy actions on stock market as a robustness check.
Table 3 and Table 4 summarize the descriptive statistics of stock return and market interest rate proxied by 3-month interbank offered rate for each country studied. Among sample countries, the average daily stock return for Indonesia is the highest, followed by China and Thailand while it is the lowest for UK, South Korea and Malaysia. Chinese stock market exhibits the highest volatility with a standard deviation 0.017 among all eight countries, followed by Norway. Stock market in Malaysia and Mauritius are less volatile. As for the interbank offered rate (presented in Table4), JIBOR is the highest with an average 7.78% while LIBOR is the lowest, approximately 2.01%. Overall, average stock return and interbank offered rate in developing countries is higher than that in developed countries.
22
Table 5 presents the descriptive statistics of policy rate change in sample countries. The adjustment range in China is the smallest with the lowest value -0.27% and highest value 0.27%. While public bank in Norway declines its policy rate to the most extent among all countries, with a drop of 1.75%. Mauritius increases its policy rate to the highest level, approximately 0.75%.
Descriptive statistics in Table 6 shows the average absolute value of policy rate change and surprise policy rate change in the eight countries. The unanticipated policy rate change is very close to real policy change in Malaysia, which implies investors mainly have no anticipation about policy rate change. Therefore, the monetary policy decisions released in Malaysia can be taken as fully surprise. To the contrary, the average absolute value in 3-month LIBOR and 3-month NIBOR change on announcement day is far small than real policy rate change, only 9 basis points and 7 basis points, respectively. This indicates in UK and Norway, investors forecast nearly all monetary policy decisions. Overall, monetary policy decisions are easier to be predicted by public in developed economies than in emerging markets.
Table 7 summarizes main characteristics of pooled samples. During the sample period, the minimum value of policy rate change is -1.75% and the maximum value of it is 0.75%. The corresponding surprise is -1% and 0.75%, respectively.
Table 8 presents the descriptive statistics of policy rate change, the surprise and nonsurprise policy shocks and daily return after the announcement as well as the financial depth in China. The sample period includes 24 monetary policy announcements. The average policy rate change s -0.01%, with the minimum value of -0.27% and the maximum value 0.27%. the corresponding one-day stock return varies from -4.55% to 3.81%. The average surprise of policy decisions, measured by change in 3-month SHIBOR, is -1.92 basis points. Stock return has an average mean of 0.392%, 0.679%, 1.08% and 1.55% for event window two-day, three-day, four-day and five-day after monetary policy released, respectively.
Table 9-15 summarize the descriptive statistics of all variables estimated in the regression model in Thailand, Indonesia, Malaysia, Mauritius, Korea, Norway and UK on announcement day or the day after, depending on the policy released time, interbank offered rate fix time and stock market closing time. During the sample period, Indonesia adjusts its policy rate up to 44 times while
Malaysia only adjusts for 9 times. China actually owns the highest financial depth value. However, in tables presented, the average financial depth in China is very similar as in Korea, still the highest in eight countries. This is because People’s Bank of China has remained the policy rate unchanged since 23th October, 2015. In the meantime, financial market depth in Indonesia is the lowest with an
23
average value 26.79, which indicates that the financial services provided in Indonesia is not adequate to meet the requirement of economic growth.
Figure 7-14 gives the plots about policy rate change and unanticipated policy rate change in each country. It is clearly can be seen from the figures that every country experienced a decline in policy rate as well as the market interest since 2007 and an increase since 2009, exactly during the financial crisis period. Monetary policy committees in these countries all adopted corresponding measures to get through the financial crisis and help recover economy. Except for Indonesia, the plot in figure 9 suggests that Bank Indonesia kept increasing BI’s Rate until 4th December, 2008. After that it started
to decline the policy rate continuously for nine times till 5th August, 2009. Cut in policy rate during
this period was unpredictable for all countries. Close attention should be paid to this special period because those extreme values may affect the estimation result. Therefore, I add financial crisis as dummy variable to control and eliminate this bias.
24
5. Results
Main results regarding to the first three hypotheses are presented in Table 16, the influence from monetary policy on Chinese stock market. The initial test is running a simple regression on stock return with raw monetary policy rate change. Column (1) presents the result based on the first regression model about the relation between raw monetary policy change and Chinese stock market reaction (measured by SSE CI). The coefficient of interest is 2.001, positive and insignificant. In theory, there should be a negative correlation between policy rate change and stock movement. Therefore, we can conclude that there is no significant influence of policy rate change on stock return in China. When distinguishing the unanticipated policy rate from policy rate change, the estimated result becomes negative. Column (2) displays the magnitude of impact from unanticipated monetary policy decision to stock price and proves that the surprise monetary policy measured by 3-month SHIBOR is negative correlated with stock return, which is in line with Kuttner (2001),
Bernanke and Kuttner (2005). Unanticipated monetary policy decisions account for the movement in stock market. The result implies a 100 basis-points increase in unexpected monetary policy rate is associated with an approximately 3.8% decline in Chinese stock price. R square is 0.118, which indicates that unanticipated monetary policy announcement can explain 11.8% of stock market movement on announcement day. However, the result is not significant as expected. P value of surprise in column (2) is 0.248, which means that there is 75.2% possibility the result is significant and to reject null hypothesis. Limited observations can be an explanation toward this. In general, monetary policy shock is helpful to explain stock market movement. Column (3) only investigates how stock market moves toward unanticipated monetary policy announcements. As shown in the table, the relationship is negative, consistent with conjecture of previous literatures.
Values in column (2) (4) (5) (6) and (7) report the stock market reaction to policy rate change during different event window. Although coefficient of interest is negative in column (2) (4) and (5), the corresponding p value is pretty high in column (4) and (5), 0.795 and 0.957 respectively.
Furthermore, the absolute value of coefficient of interest in column (2) (-3.819) is larger than that in column (4) (-1.194) and column (5) (-0.289). The R square values also indicate that one-day return presents the best performance when capturing the related information and fitting well with the regression model. My previous hypothesis is that Chinese stock market needs time to fully digest all related information. However, results presented in this table suggest the null hypothesis cannot be rejected. Chinese stock market is relatively efficient. One can be explained by the fact that Chinese monetary policy is released after the stock closes. Therefore, when calculating the stock reaction, data from next transaction day is used. Investor has at least one day, if the announcement is released in Friday then more than two days, to decide whether to hold or sell their equity. Stock
25
market has more time than preconceived to react to the shock. Therefore, the one-day reaction to policy decisions is the most significant compared to 2-day, 3-day 4-day and 5-day reaction. Stock market is not lagging when reflecting policy decisions. If the announcement is released during trading hour, or in extreme case just before the closing time, the one-day response may not be an ideal independent variable.
For the cross-country analysis, results are given by table 17. First column presents the magnitude of influence from raw policy rate change to stock return in pooled countries and second column reports the regression result of decomposed monetary policy decision on stock price. The third column takes the unanticipated policy rate change as the only independent variable in the regression model. The sample covers 197 observations in 8 countries from 2006 to 2016. Although it has a negative value, the overall estimation in column (1) is insignificant and small, approximately -0.00675. When dividing the monetary policy rate change into surprise and nonsurprise components, the observed value of coefficient for surprise in column (2) has a significant negative value, suggesting that there is a significant reverse relation between stock price and unanticipated policy rate shock. This finding is in line with previous research that stock markets only react to surprise rate change. Countries studied presented an inverse link between unexpected monetary policy and stock price movement. The estimated stock market response to surprise rate change for all eight countries is -0.73, meaning that an average 100 basis point cut in unanticipated policy rate change would cause a 0.73% rise in stock price in the pooled countries. Observed results of coefficient for nonsurprise policy rate decision provide a brief measure of market efficiency. Efficient Market Hypothesis assumes public information is immediately incorporated into stock price. According by Kuttner (2001), Bernanke and Kuttner (2005), the expected policy rate change has already been priced into the stock market. Therefore, the coefficient of nonsurprise in column (2) should be zero in an ideal world. If the market is efficient, then the coefficient of nonsurprise should be zero. The empirical results show that sample markets are not perfect efficient.
Column (4) takes the country dummy into consideration. If the country investigated is developing country, the dummy equals to zero and vice versa. Unanticipated monetary policy announcement exhibits a strong predict power in forecasting stock market movement with a significant level 95%. The coefficient of interaction term between country dummy and surprise is positive, proving that the effect of monetary policy action is weaken in developed countries. Developing countries own a relatively perfect financial system. Investors and firms have access to multiple financing channels and therefore are less affected by the policy rate change.
26
To explore the different reaction of stock market to monetary policy actions under different business condition, I add an interaction term between crisis dummy and surprise monetary policy decision. the crisis dummy equals to one if the announcement is released during crisis period from December 2007 to June 2009. The results are reported in column (5). A study by Kurov (2010) indicates that there exist asymmetries in the reaction of stock market to monetary policy announcement. The coefficient of this interaction term is negative, indicating when the monetary policy is announced during financial crisis period, the impact from monetary policy announcement on stock market is strengthened. Financial crisis can increase the magnitude of stock market reaction to surprise monetary policy announcement. The stock markets are more sensitive to monetary policy shocks during crisis period. For all countries, if monetary policy decision was released during 2008 financial crisis, 100 basis points policy rate cut would lead to 1.018% more increase in stock market than non-crisis period. The regression results prove that the effect of monetary policy varies from business cycle, in line with Kurov (2010).
When financial depth is taken into consideration, the overall impact from unexpected policy rate change may be affected. My hypothesis is supported by column (6), presenting the result about whether the level of financial depth would affect stock market reaction to monetary policy decisions. Overall, the coefficient of interaction term between financial depth and unanticipated policy rate change is 0.0137. This means financial depth plays an essential role in the monetary policy transmission process and can mitigate the negative influence on stock price brought by monetary policy decision. A deep financial market can impair the overall effect of monetary policy decisions on stock market. Financial depth captures the relationship between financial sector and the economy, proxied by private credit to GDP. Close connected with income level, financial depth measures to what degree can financial market meet the economy growth’ requirement. If financial depth is too low, potential economy growth may get restrained. The deeper the financial market, the stronger poverty reduction and further economy growth a country would experience. The dampening effect of financial depth is easy to understand since investors and firms have multiple funding channels in a country with deep financial market. Therefore, the change of policy interest rate may not have the same impact as policy makers’ expectation, investors and companies can raise money from other sources. Strengthening in financial depth can weaken monetary policy’s impact on stock market. Overall, we can conclude that the real effect of monetary policy varies with the level of financial depth.
27
6. Robustness Test
In this section, to test if the results are robust, I run the same regression models as before but with a shorter sample period to see if the outcome is in line, by tightening the sample period to July, 2009- December, 2016 (the post-crisis period). Regression results are displayed in table 18.
Table 18 presents the regression results in China with narrow sample period. The regression result is somehow different with the full sample period. First, coefficient in column (1) is negative, indicating a 100% basis points cut in raw policy rate can lead to around a 1.377 rise in stock price, while the value in full sample period is positive. Nevertheless, the estimated value is insignificant. Compared to the figure in column (2) and (3), the outcome in the first regression model is still consistent with previous literature, stock market is negatively related with unexpected monetary policy decisions. The magnitude of impact from unanticipated monetary policy in China is much stronger in a narrow sample period. An overall 25 basis points rise in unanticipated policy rate change causes a roughly 3.84% drop in stock price, almost four times larger than before. Even though the coefficient of interest is negative in column (2) and (4), the p value of column (2) is much smaller than that in column (4). Therefore, we can conclude that the one-day stock return can capture related monetary policy information.
The overall impact of unanticipated monetary policy decisions is weaker in the narrow sample period, provided in table 19. For the pooled samples, 100 basis points rise in unanticipated policy rate change can result in an approximately 0.42% fall in stock price as exhibited in column (2). This effect is weaker in developed economies, consistent with the previous results. Central bank in developed economies tend to hire forward guidance as a tool to implement monetary policy.
Forward guidance provides forecast of future interest rate provided by officers in central bank, given from official or unofficial conversation. Central bank initiative release information about future monetary policy. Among our sample countries, Norges Bank adopted this approach in 2005, followed by UK which officially started to employ this tool at 7th August, 2013. Central bank wants to impact
public’s expectation about future policy rate. Therefore, the monetary policy will not cause a volatile stock market movement compared to that in emerging markets.
Column (5) shows the robustness test of the influence by financial depth. The overall result is similar with previous that financial depth can weaken the impact of monetary policy decision on stock market. The coefficient of the interaction term is 0.003. In another aspect, not ideally, proves that adding this interaction term can cause a reverse movement in stock market. Financial depth plays an important role in the transmission mechanism of monetary policy.
28
7. Conclusion:
The aim of this thesis is to examine the relationship between stock market movement and monetary policy decisions. I utilize an event-study approach, eliminating the effect of endogeneity to the largest extent. A vital part in this study is decomposing monetary policy into surprise and
nonsurprise element to measure the stock market reaction to unanticipated monetary policy in the spirit of Kuttner (2001). The sample period is from 2006 to 2016, main study object is China. The empirical result is in line with previous literature that unanticipated monetary policy actions show a powerful explanation in stock price. Statistical results indicate that in China a hypothetical surprising 100 basis points increase in interest rate lead to a 3.81% cut in stock price.
Then I extend the data to eight countries, covering both emerging markets and developed markets. I find that following a 100 basis points increase in policy rate, stock return decreases by 0.73%. I also examine whether business cycle can affect from monetary policy announcement on stock market by interacting crisis dummy with surprise policy shock. Results prove that the effect of monetary policy varies from business cycle. During crisis period, stock market reacts more significantly to monetary policy announcement. Furthermore, I evaluate the importance of financial depth by studying its relation to the stock market. I add an interaction term between financial depth and surprise policy rate change into the regression model. Regression results indicate that monetary policy effect on stock market is weaken by financial depth. A deep financial market can mitigate the reverse relationship between monetary policy decision and stock market. After that, I conduct the
robustness test by shortening the sample period into post-crisis period, from July 2009 to December 2016. The robustness test shows that previous results are reliable and robust to different sample period.
Overall, unanticipated monetary policy announcement is negatively related with stock market movements. This effect is strengthened during crisis period and dampened by financial depth. The effect of monetary policy announcement on stock market is asymmetric.
