影響墨西哥S&P/BMV IPC的外部因素 - 政大學術集成

全文

(1)國立政治大學商學院國際經營管理英語碩士學位學程 International MBA Program College of Commerce National Chengchi University. 立. 政治大碩士論文. ‧. ‧ 國. 學. Master’s Thesis. io. sit. y. Nat. 影響墨西哥S&P/BMV IPC的外部因素 n. al. er. External Factors Affecting the S&P/BMV IPC. Ch. engchi. i Un. v. Student: Jorge Federico Gonzalez Antuna Advisor: Jason C.H. Tsai. 中華民國一◯ 九年五月 May 2020. DOI:10.6814/NCCU202000683.

(2) 影響墨西哥S&P/BMV IPC的外部因素 External Factors Affecting the S&P/BMV IPC. 研究生：高岱辰. Student: Jorge Federico Gonzalez Antuna. 指導教授：蔡政憲. Advisor: Jason C.H. Tsai. 立. 政治大. ‧ 國. 學. 國立政治大學. 商學院國際經營管理英語碩士學位學程. ‧. 碩士論文. er. io. sit. y. Nat. A Thesis. Submitted to International MBA Program. n. al. v. National ni C Chengchi University. hengchi U. in partial fulfillment of the Requirements for the degree of Master in Business Administration. 中華民國一◯ 九年五月 May 2020. DOI:10.6814/NCCU202000683.

(3) Acknowledgements I wish to express my deepest gratitude to my thesis advisor, Professor Jason C.H. Tsai who helped and guide through the process of writing this thesis. Emily Huang, Li-chi and Jasmine Lu from IMBA office whose assistance has been boundless through the time I have studied my master program at National Chengchi University. TaiwanICDF and my family, whose support made this possible, immeasurable thanks to all of you.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. i DOI:10.6814/NCCU202000683.

(4) Abstract External Factors Affecting the S&P/BMV IPC By Jorge Federico Gonzalez Antuna This paper’s objective is to investigate the influence of external factors to the Mexican economy on the Mexico’s “Índice de Precios y Cotizaciones” (S&P/BMV IPC), the principal index of the Mexican Stock Exchange. The sampling period of this study is from January 2015 to. 政治大 existence of relationship between 立 S&P/BMV IPC and the exogenous variables.. December 2019. We use multiple linear regression and cointegration approaches to test the. ‧ 國. 學. Findings from the multiple linear regression model suggest that exchange rate has a negative influence in S&P/BMV IPC value, while cointegration approach suggests that only the index. ‧. value of the previous period has influence in the index value of the subsequent period.. n. er. io. sit. y. Nat. al. Ch. i Un. v. Keywords: Mexican Stock Exchange, S&P/BMV IPC, international crude oil price, Federal Reserve, Donald Trump. engchi. ii DOI:10.6814/NCCU202000683.

(5) TABLE OF CONTENTS 1. Introduction .......................................................................................................................... 1 1.1. Background ...................................................................................................................... 1 1.2. Research Questions ......................................................................................................... 5 1.3. Research Objective .......................................................................................................... 5 1.4. Limitation of Research .................................................................................................... 5. 政治大. 2. Literature Review ................................................................................................................. 6. 立. 2.1. Mexican Stock Exchange ................................................................................................ 6. ‧ 國. 學. 2.1.1. History .................................................................................................................... 6. ‧. 2.2. Índice de Precios y Cotizaciones S&P/BMV IPC ........................................................... 7. Nat. sit. y. 2.3. Factors that Influence S&P/BMV IPC ............................................................................ 8. n. al. er. io. 2.3.1. International Crude Oil Price .................................................................................. 8. i Un. v. 2.3.2. S&P 500 .................................................................................................................. 9. Ch. engchi. 2.3.3. Exchange Rate USD/MXN..................................................................................... 9 2.3.4. Donald Trump Tweets........................................................................................... 10 2.4. Previous Research on Factors affecting the S&P/IPC BMV ......................................... 11 3. Data and Methodology ....................................................................................................... 12 3.1. Research Method ........................................................................................................... 12 3.2. Object of Research......................................................................................................... 12 3.3. Types and Sources of Data ............................................................................................ 12. iii DOI:10.6814/NCCU202000683.

(6) 3.4. Variable Definition ........................................................................................................ 12 3.4.1. S&P/BMV IPC ..................................................................................................... 12 3.4.2. International Crude Oil Price ................................................................................ 13 3.4.3. S&P 500 Index ...................................................................................................... 13 3.4.4. Exchange Rate USD/MXN................................................................................... 13 3.4.5. Donald Trump Tweets........................................................................................... 13 3.5. Methodologies ............................................................................................................... 14. 政治大. 3.5.1. Multiple Linear Regression Model ....................................................................... 14. 立. 3.5.2. Linear Regression Assumptions ........................................................................... 14. ‧ 國. 學. 3.5.3. Hypothesis Testing MLR ...................................................................................... 15 3.5.4. ARDL Bound Co Integration Test ........................................................................ 15. ‧. 3.5.5. Diagnostic Tests ARDL ........................................................................................ 17. y. Nat. al. er. io. sit. 3.5.6. Hypothesis Testing ARDL .................................................................................... 18. n. 4. Results and Discussions ...................................................................................................... 19. Ch. engchi. i Un. v. 4.1. Multiple Linear Regression ........................................................................................... 19 4.1.1. Descriptive statistics ............................................................................................. 19 4.1.2. Correlation Matrix ................................................................................................ 19 4.1.3. Multiple Linear Regression Equation ................................................................... 20 4.1.4. Hypotheses Test Result MLR ............................................................................... 20 4.1.5. Overall Significance Test MLR ............................................................................ 21 4.1.6. Assumptions Test MLR ........................................................................................ 23 4.2. ARDL Bound Co Integration Test ................................................................................. 24. iv DOI:10.6814/NCCU202000683.

(7) 4.2.1. Descriptive statistics ............................................................................................. 24 4.2.2. Correlation Matrix ARDL .................................................................................... 26 4.2.3. Unit Root Test ARDL ........................................................................................... 26 4.2.4. ARDL Co Bound Integration Test ........................................................................ 27 4.2.5. Long Run Equation ARDL ................................................................................... 28 4.2.6. Error Correction Model ARDL............................................................................. 28 4.2.7. Hypotheses Test Result ARDL ............................................................................. 29. 政治大 ECM Diagnostics tests ......................................................................................... 31 立. 4.2.8. Overall Significance Test ARDL .......................................................................... 30. 學. ‧ 國. 4.2.9.. 5. Conclusions ......................................................................................................................... 33. ‧. 6. References............................................................................................................................ 34. y. Nat. n. al. er. io. sit. 7. Appendix A .......................................................................................................................... 37. Ch. engchi. i Un. v. v DOI:10.6814/NCCU202000683.

(8) List of Figures and Tables Figure 1: Mexico’s Crude Oil Production 2011-2019 ................................................................ 2 Figure: 2 U.S. vs. Mexico Stock Market, March 2009 – March 2013 ....................................... 3 Figure 3: USD/MXN Exchange Rate, 2015-2020 ...................................................................... 4 Figure 4: MXN/USD’s Record Low after Trump’s Announcement of Major Tax Border ......... 5 Figure 5: S&P/BMV IPC vs. Mexico GDP ................................................................................ 8. 政治大. Figure 6: Energy Consumption by Type, 2018 ........................................................................... 8. 立. Figure 7: Normal Probability Plot MLR .................................................................................. 23. ‧ 國. 學. Figure 8: Monthly Movement of Variables ARDL ................................................................... 25 Figure 9: CUMSUM Stability Test. .......................................................................................... 32. ‧. Figure 10: Normality Plot ECM ............................................................................................... 32. io. sit. y. Nat. n. al. er. Table 1 Result of Descriptive Statistics MLR .......................................................................... 19. Ch. i Un. v. Table 2 Correlation Matrix MLR.............................................................................................. 19. engchi. Table 3: Multiple Linear Regression Coefficients .................................................................... 20 Table 4: ANOVA MLR ............................................................................................................. 22 Table 5: Multicollinearity Test MLR ........................................................................................ 23 Table 6: Heteroskedasticity Test Result MLR .......................................................................... 24 Table 7: Result of Descriptive Statistics ARDL ....................................................................... 24 Table 8: Correlation Matrix ARDL .......................................................................................... 26 Table 9: Results of Unit Root Test ARDL ................................................................................ 27 Table 10: ARDL Co Bound Integration Test Result ................................................................. 27. vi DOI:10.6814/NCCU202000683.

(9) Table 11: Long Run Equation IPC ARDL ................................................................................ 28 Table 12: Error Correction Model ARDL................................................................................. 28 Table 13: Overall Significance Test ARDL .............................................................................. 31 Table 14: Breusch-Godfrey Serial Correlation LM Test .......................................................... 31 Table 15: Heteroskedasticity White Test .................................................................................. 31. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. vii DOI:10.6814/NCCU202000683.

(10) 1. Introduction 1.1. Background A stock exchange first than anything is a market. Therefore, the price of what is traded in it, and money, in the form of securities such as stocks, bonds and others financial instruments, are subject to the supply and demand economic model. Which in summary determines the price as follows: if the demand of a good increases, then its price will increase; if the supply of a good increases then its price will decrease. Companies listed in the Mexican Stock Exchange aim to obtain financial resources from. 政治大 buy these securities may share the wealth of profitable businesses through dividends and stock 立 price increments. On other hand, investors could incur capital losses in case the company’s. investors by placing representative securities of their capital (shares of stocks). Investors who. ‧ 國. 學. shares they bought, is unprofitable or troublesome.. ‧. These stocks as any good in any market have a price or market value, which varies everyday. y. Nat. according supply and demand. If numerous investors want to acquire a specific stock, that stock. sit. market value will increase as long as this trend continues and few investors would like to sell. n. al. er. io. that stock. In the opposite way, if the supply of a company’s stock is greater than the demand,. v. then numerous investors would like to sell that stock with a very few willing to acquire it, This. Ch makes its market value go down.. engchi. i Un. This dynamic takes place every day at the Mexican Stock Exchange. However, what motivates investors to buy or sell a specific stock? Many factors potentially influence these decisions, such as macroeconomics perspectives, political and social events, particular economic structure of Mexico and company specific matters among many others. Analyzing the whole stocks market value listed in Mexican Stock Exchange would require remarkable resources; hence, this paper aims to analyze the influence of external factors into the S&P/BMV IPC index. A stock market index is a measure of a stock market, or a smaller subset of the market, that helps investors compare current price levels with past prices to evaluate market performance [Caplinger, 2020]. The S&P/BMV IPC aims to measure the performance of the biggest stocks 1 DOI:10.6814/NCCU202000683.

(11) with greatest liquidity that are listed in the Mexican Stock Exchange. The index provides a representative, invertible and replicable sample of the Mexican stock market (Sánchez, 2018). This research aims to analyze whether four external factors, the variables out of control of the Mexican Central Bank, affect the S&P/BMV IPC. The first variable is international price of crude oil. Mexico is the 12th largest oil producer in the world (U.S. Energy Information Administration, 2019); it ranks the same in terms of export. Even though oil production in the country has dropped during the last years, as shown in figure 1, oil crude revenues account for 4.8% exports earnings in 2019 (INEGI, 2020) and represented 17.2% of the Mexican government financial resources (Sandoval, 2020). High oil prices are good for the public. 政治大. finances to support the social and healthcare government programs, as well as to drive job. 立. creation and investment since high oil prices make it profitable for oil companies to exploit oil. ‧ 國. 學. deposits. In contrast, high oil prices hit businesses and consumers with higher transportation and manufacturing costs, reducing the profitability of companies, and thus lowering their stock. ‧. price.. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. Source: tradingeconomics.com. Figure 1: Mexico’s Crude Oil Production 2011-2019 The second variable is the S&P 500 index. The relationship between the U.S. and Mexico varied over time, from symbiotic economic partnership to conflictual relationship regarding issues such as immigration policies, illegal gun trade and crime. Mexico is currently the third largest U.S. goods trading partner with $611.5 billion in total (two-way) goods trade during 2018 (U.S. Department of Commerce, 2018).. 2 DOI:10.6814/NCCU202000683.

(12) U.S. policy has often strongly influenced the Mexican economy and currency indirectly through large-scale shifts in the U.S. economic policy stance [FXCM, 2017]. When the Fed increases the discount rate, the companies in the U.S. may experience a decrease in their earnings since they will tend to spend less, decrease their earnings, this would lead to the stock price reduction. The S&P500 measures the stock performance of the 500 large companies listed in the stock exchange in the U.S., many experts consider it one of the best representations of the U.S. stock market (Kenton, 2019). Due to the close economic relationship with the U.S., the Mexican stock market will follow the movement of the U.S. stock market. Figure 2 below, shows a tight relationship between U.S. and Mexico stock markets.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. Source: seekingalpha.com. Figure: 2 U.S. vs. Mexico Stock Market, March 2009 – March 2013 The third variable is the exchange rate between the USD and the MXN Mexican peso. An exchange rate is the rate at which one currency will be exchanged for another. It is also regarded as the value of one country's currency in relation to another currency [O'Sullivan, 2003]. The USD/MXN exchange rate is free floating; therefore, it is under the supply and demand dynamic. We have selected the USD since it is the most dominant currency in the international reserves, its transactions are almost the half of the world transactions and the economic and commercial relations between the U.S. and Mexico are rather close. The currency exchange rate is one of the most important determinants of a country's relative 3 DOI:10.6814/NCCU202000683.

(13) level of economic health. Since the end of the fix rate, the behavior of the MXN variability towards the USD has become an indicator of Mexican economy competitiveness and performance [Gutierrez Ortiz, 2011]. A low value of the MXN is an indicator that the Mexican economy is not healthy since people avoid the risk by acquiring USD securities.. 立. 政治大. ‧ 國. 學 Source: macrotrends.com. ‧. Figure 3: USD/MXN Exchange Rate, 2015-2020. y. Nat. sit. The fourth variable is the Donald John Trump tweets. Donald Trump is the current and 45th. er. io. president of the United States. On June 16, 2015, he announced his candidacy for President of. n. al. iv n C U was amongh other e n gissues. c h iMexico. the U.S., and discussed about illegal immigration, offshoring of American jobs, U.S. commercial balance deficits,. mentioned 13 times with. statements such as: “When do we beat Mexico at the border? They’re laughing at us, at our stupidity. And now they are beating us economically. They are not our friend, believe me. But they’re killing us economically” and; “I’ll bring back our jobs from China, from Mexico, from Japan, from so many places.” (Time, 2015). He communicated heavily on Twitter during the 2016 presidential election campaign and has continued to use this channel during his presidency. Political analysts inside and outside the U.S. did not take seriously Trump’s campaign at beginning but when his rival for the Republican Party’s primary election Ted Cruz dropped out on May 2016, the exchange rate between the Mexican currency, MXN, and the U.S. dollar began to respond to Trump’s comments. Because of its sensitivity to policy remarks by Trump,. 4 DOI:10.6814/NCCU202000683.

(14) the U.S. dollar-Mexican peso exchange rate became associated with the term the "Trump trade" (FXCM, 2017). The major interest of this research is to determine whether Trump’s announcements through Twitter, a mayor communication channel of U.S. president, have impacts on S&P/BMV IPC.. 立. 政治大. ‧ 國. 學. Source: finance.yahoo.com. Figure 4: MXN/USD’s Record Low after Trump’s Announcement of Major Tax Border. ‧ y. Nat. 1.2. Research Questions. n. al. er. io. 2. Does S&P 500 influence the S&P/BMV IPC?. sit. 1. Does the international crude oil price influence the S&P/BMV IPC?. i Un. v. 3. Does the exchange rate between the USD and the MXN influence the S&P/BMV IPC?. Ch. engchi. 4. Do Donald Trump’s tweets influence the S&P/BMV IPC?. 1.3. Research Objective Find the impact of international crude oil price, S&P 500, USD/MXN exchange rate and Donald Trump’s tweets on S&P/BMV IPC used as a proxy of the Mexican Stock Exchange (BMV).. 1.4. Limitation of Research The limitation of research is defined to focuses on S&P/BMV IPC that is influenced by international crude oil price, S&P 500, USD/MXN exchange rate and Donald Trump’s tweets period from January 2015 to December 2019.. 5 DOI:10.6814/NCCU202000683.

(15) 2. Literature Review 2.1. Mexican Stock Exchange The Mexican Stock Exchange (BMV) is a public institution that operates under a concession granted by the Ministry of Finance (SHCP) and observes the Mexican Securities Law. It is one of two stock exchanges in Mexico with the other being BIVA – “Bolsa Institucional de Valores”. BMV is the second largest stock exchange in Latin America with a total market capitalization of over US$ 530 billion. It is the physical place where the trades made by the brokerage firms are executed and registered. Investors buy and sell stocks and debt securities through intermediaries, called brokerage firms. Its principal index is “Índice de Precios y Cotizaciones” S&P/BMV IPC.. 立. 政治大. The Mexican Stock Exchange (BMV) is a forum where the organized securities market. ‧ 國. 學. transactions are held. Its main objective is to facilitate the securities transaction, market development and competitiveness.. ‧. The companies that require monetary resources (money) to finance its operations or expansion. y. Nat. projects, might find it at the securities market, by issuing securities (stocks, bonds, commercial. sit. paper, etc.) to the investors (issued) that are traded (bought and sold) at the Mexican Exchange,. er. io. in a transparent and free competitive market with equal opportunities for all participants.. al. n. iv n C U worldwide are h eestablished n g c h iinstitutions. The stock exchanges. by the society for its own. benefit. Investors go to them as an option to protect and increase their monetary savings while supply monetary resources that in turn allow corporations and governments to finance productive and development projects that in turn generate jobs and wealth. In this sense, the Mexican Exchange encourages the development of the country since; it also helps to allocate savings into productive investments that is the source of growth and employment in Mexico.. 2.1.1. History From 1880 to 1900, Mexico City’s downtown streets, held meetings where brokers and entrepreneurs bought and sold all sorts of assets. Later, exclusive groups of stockholders and issuers involved themselves, trading behind closed doors in different parts of the city. On June 14, 1895, a group of brokers formed a partnership under the name Bolsa de Valores Mexico 6 DOI:10.6814/NCCU202000683.

(16) (Mexico’s Stock Exchange). On August 28, 1933, the deed and statutes of the Bolsa de Valores de Mexico, S.A. were approved in accordance with the concession authorized, which for the first time included the stock exchanges. The exchange is supervised by the National Council on Securities (nowadays the National Banking and Securities Commission). By 1975 the Securities' Market Law, prompted the Bolsa de Valores de México to change its name to Bolsa Mexicana de Valores (Mexican Stock Exchange). On 19 April 1980, the S&P BMV/IPC was first published.. 立. 政治大. In 2006, the Mexican securities market was opened to foreigners through the MexDer system. 2.2. Índice de Precios y Cotizaciones S&P/BMV IPC. ‧. ‧ 國. 學. for them to operate from anywhere in the world (BMV Group, 2015).. sit. y. Nat. A reference index aims to represent the performance of a stock market, assets or market segment. As mentioned before in this research, the S&P/BMV IPC aims to measure the performance of. io. n. al. er. the 35 biggest and greatest liquidity stocks listed in the Mexican Stock Exchange. Similar to. i Un. v. other indexes in the world, the S&P/BMV IPC can be regarded as an economic indicator since. Ch. engchi. the stock market has a strong relationship with the economic variables. According to Sanchez 2018, the correlation between the annual variation of Mexican GDP and that of the S&P/BMV IPC is 57% during the period from 1979 to 2017.. 7 DOI:10.6814/NCCU202000683.

(17) 立. 政治大 Source: Sánchez 2018. ‧ 國. 學. Figure 5: S&P/BMV IPC vs. Mexico GDP. ‧. 2.3. Factors that Influence S&P/BMV IPC. Nat. sit. y. 2.3.1. International Crude Oil Price. io. er. In 2018, worldwide energy consumption was 13,978 millions of tons of oil equivalent [Enerdata, 2019], being oil the major source of energy accounting for 32% as show in figure 6 below.. al. n. iv n C Energy Consumption U 2018 h e n g c hbyi Type 10% Oil 32% Gas 26%. Electricity Coal 9%. 23%. Biomass. Figure 6: Energy Consumption by Type, 2018 , 8 DOI:10.6814/NCCU202000683.

(18) The price of oil refers to the spot price of a barrel of benchmark of oil, a reference for buyers and sellers, benchmarks such as Brent Crude, West Texas Intermediate (WTI) and Isthmus. The price of barrel varies according its grade, heavier crude oils are more expensive than lighter crude oils. WTI is the benchmark for the Mexican crude oil. Even though the oil price is subject to the supply and demand dynamic, the Organization of Petroleum Exporting Countries, OPEC, an intergovernmental organization of 14 nations that account for 44 percent of global oil production and 81.5 percent of world’s proven oil reserves have a major influence on global oil prices, acting as a cartel to reduce market competition. In 2019, Mexico was the 12 th largest crude oil producer [U.S. Energy Information Administration, 2020], 2.71 percent of global production, however, the country is not part of the OPEC.. 政治大. 立. 2.3.2. S&P 500 in. ‧ 國. exchange. the. U.S.. It. is. a. market. 學. The S&P 500 measures the stock performance of 500 large companies listed on the stock value-weighted. index;. its. components. ‧. are weighted according to the total market value of their outstanding shares. Due to the large number of companies across all sectors included in the S&P 500, investors perceive it as the. Nat. sit. y. most representative stock index in the U.S.. er. io. The U.S. is over one-half of the global stock market, according to S&P Global BMI measure. As the U.S. economy grows, it propels the stock markets of all other countries—and the U.S.. n. al. Ch. i Un. v. stock market is largely driven by consumer spending. Therefore, the more a country exports to. engchi. the U.S., the more sensitive it becomes to U.S. growth (Gunzberg & Edwards, 2018).. 2.3.3. Exchange Rate USD/MXN Exchange rate is regarded as the value of one country's currency in relation to another currency [O'Sullivan, 2003]. Aside from interest rate and inflation, the currency exchange rate is one of the most important factors of a country’s economic health level. Several factors influence the exchange rate changes: -. The supply and demand dynamic between the currency of two countries, appreciating or depreciating the price of nation currency relatively to other country’s currency.. -. Officially devaluation of national currency conducted by the government of a country.. -. Officially revaluation of national currency conducted by the government of a country. 9 DOI:10.6814/NCCU202000683.

(19) -. Foreign currency supply and demand, supply coming from goods export, capital import. Demand coming from goods imports, capital export.. -. Differentials in inflation, a country with a consistently lower inflation rate exhibits a rising currency value, as its purchasing power increases relative to other currencies.. -. Differentials in interest rate, by manipulating interest rates, central banks affect inflation and currency values. Higher interest rates offer lenders in an economy a higher return relative to other countries. Therefore, higher interest rates attract foreign capital and cause the exchange rate to rise. The opposite phenomena occurs when the central banks decreases the interest rate.. -. Public deficit, a country government can get monetary resources to pay for public sector. 政治大. projects and governmental funding by borrowing money. This stimulate the domestic. 立. economy but large debts encourage inflation, decreasing the exchange rate of the. ‧ 國. -. 學. country.. Position of balance of payments, the balance of payments is a statement of the prices of. ‧. imports and exports of goods, services and capital between a country and the rest of the. Nat. er. io. sit. decrease in relation to its trading partners.. y. world. If the price of exports is smaller than that of its imports, the currency's value will. 2.3.4. Donald Trump Tweets a. n. iv l C n hengchi U used to draw attention to his book, “Think Like a Champion”. From 2013 onwards, Trump's. Donald Trump opened his twitter account @RealDonaldTrump on May 4, 2009. He initially Twitter activity pattern significantly increased in the volume of his tweets and the rate of politically charged rhetoric [Fandos, 2017]. For the 2016 presidential campaign, Donald Trump used its twitter account as an important political communication channel. Political analysts consider that the use of twitter contributed for his victory in the 2016 U.S. presidential elections [Ingram, 2017]. Even though there is an official tweeter account for the U.S. President, @POTUS, President Donald Trump prefers to use his personal twitter account to issue major announcements himself over social media, bypassing the White House Press Secretary. President Donald Trump’s tweets are a great way to know what the president is thinking, doing or thinking of doing, they are also a great source of conflict and controversy. His tweets have effect in the stock market, 10 DOI:10.6814/NCCU202000683.

(20) for example, on August 17, 2017 post on Amazon: ‘Amazon is doing great damage to tax paying retailers. Towns, cities and states throughout the U.S. are being hurt - many jobs being lost! [Trump, 2017] Afterwards, the market capitalization of Amazon declined by $6 billion [Chapman, 2017].. 2.4. Previous Research on Factors affecting the S&P/IPC BMV Lopez-Juaréz, Ladrón de Guevara-Cortés and Madrid-Paredones (2019) in their research about a set of macroeconomic variables that could explain the behavior of the stock market in Mexico, they select the S&P/IPC BMV as the market proxy. The results showed that the international. 政治大. oil price and the exchange rate are factors that influence the S&P/IPC BMV behavior during the period from January 02, 2008 to December 31, 2015. Their research also included the Banco. 立. de Mexico interest rate but according to their results, there was not sufficient statistical evidence. ‧ 國. 學. to support the interest rate’s influence on S&P/IPC BMV.. ‧. Singhal, Choudhay, Biswal (2019) investigate the dynamic relationship among international oil prices, international gold prices, exchange rate and stock market index in Mexico. Findings of. Nat. sit. y. their study suggest that oil price negatively in the long-run influence the stock market index,. al. er. io. exchange rate and gold prices are significantly correlated with stock prices, tending to move in. n. the same direction. They used daily data from January 2006 to April 2018 for their study.. Ch. engchi. i Un. v. Sardy, Cova (2019) findings show that Trumps tweets influence financial markets, using S&P 500 index as the market proxy, at daily returns level. They first collect all economy related Trump’s tweets per day, subsequently, they assign one of three sentiment labels, positive, neutral or negative, determined by whether President Trump had worded in a positive, neutral or negative tone. The daily tweet sentiment was then calculated by assigning each sentiment label with a score; positive: +1, neutral: 0, negative -1, and summing the sentiment of each economy-related tweet of the day. Regarding Trumps tweets effect in foreign markets Guo, Jiao, Xu (2019) research paper analyze their influence on the Chinese stock market. Their findings suggest that after his presidential inauguration, Trumps tweets containing positive sentiment towards China increase the stock prices of Chinese manufacturing companies listed on the Shanghai Stock Exchange and the Shenzhen Stock Exchange. 11 DOI:10.6814/NCCU202000683.

(21) 3. Data and Methodology 3.1. Research Method This research focuses on the influence of international crude oil price, S&P 500, exchange rate MXN/USD and Donald Trump’s tweets on S&P/BMV IPC. This research uses two models, multiple linear regression approach (MLR) auto regressive distributed lag model approach (ARDL). The analysis uses more than one independent macroeconomic variable to explain the factors influencing S&P/BMV IPC.. 3.2. Object of Research. 政治大 international crude oil price, S&P 500, exchange rate USD/MXN and Donald Trump’s tweets 立. The object of this research is the price of S&P/BMV IPC as dependent variable. Meanwhile, as independent variables. The observation is conducted by time series data amounted 60 months. ‧ 國. 學. from June 2015 to December 2019.. ‧. 3.3. Types and Sources of Data. sit. y. Nat. The source of data are secondary data from Thomson Reuters platform EIKON and Trump twitter archive website. The research uses monthly S&P/BMV IPC value, international crude. io. n. al. er. oil price, S&P 500 value, exchange rate USD/MXN and Trump’s tweets sentiment from January 2015 to December 2019.. Ch. engchi. i Un. v. 3.4. Variable Definition 3.4.1. S&P/BMV IPC S&P/BMV IPC measure the performance of the 35 biggest and greatest liquidity stocks listed in the Mexican Stock Exchange; therefore, we use it as a benchmark of the market movement in Mexico. We use price data; price is given by the formula below:. It = Index in period t Pit = Stock price of company i in period t. 12 DOI:10.6814/NCCU202000683.

(22) Qit = Stocks of company i in period t Fi = Adjust factor i = 1, 2, 3… 35 It is represented by IPC in our research. For multiple linear regression analysis, in order to ensure the stationarity we use the return of S&P/BMV IPC: Return IPCt = (IPCt – IPCt-1) / IPCt-1. 3.4.2. International Crude Oil Price This research uses West Texas Intermediate (WTI) price as a proxy for international crude oil price movements. It is measured in USD/barrel and represented by ReOIL. For multiple linear. 政治大. regression analysis, in order to ensure the stationarity we use the WTI crude oil return:. 立. ReturnOILt = (OILt – OILt-1) / OILt-1. ‧ 國. 學. 3.4.3. S&P 500 Index. ‧. The S&P 500 index is widely considered as the best representation of the U.S. stock market. It is represented as SP500 in our research. For multiple linear regression analysis, in order to. y. Nat. sit. ensure the stationarity we use the return of S&P 500:. n. al. er. io. ReSP500t = (SP500t – SP500t-1) / SP500t-1. Ch. 3.4.4. Exchange Rate USD/MXN. engchi. i Un. v. Majority of the export and import from Mexico is done in US Dollars, therefore for exchange rate movements the Mexican peso (MXN) and US dollar (US $) currency pairs is used. They are measured as USD/MXN and represented by ReEX. For multiple linear regression analysis, in order to ensure the stationarity we use the exchange rate difference between periods: ReturnEXt = (EXt – EXt-1) / EXt-1. 3.4.5. Donald Trump Tweets Consists of the Tweets posted by Donald Trump from his personal twitter account @RealDonaldTrump, and the sentiment of each Tweet. The data was collected from Trump. 13 DOI:10.6814/NCCU202000683.

(23) Twitter Archive website (http://www.trumptwitterarchive.com/archive), contained the date and content. In the 60-month period of study, Trump wrote 266 tweets related to Mexico. We determined a sentiment for each tweet. Sentiment was evaluated according to Trump’s sentiment towards Mexico; we assigned each tweet with one of three sentiment labels, positive, neutral, or negative, determined by whether he had worded it in a positive, neutral, or negative tone. The monthly tweet sentiment was then calculated by assigning each sentiment label with a score; positive: +1, neutral: 0, negative -1, and summing the sentiment of each Mexico-related tweet. It is represented by TRUMP in our model.. 3.5. Methodologies. 立. 政治大. 3.5.1. Multiple Linear Regression Model. ‧ 國. 學. The first approach this research uses is multiple linear regression analysis; its aim is to measure the impact of different factors on S&P/BMV IPC. The multiple regression equation is:. ‧. ReIPC =  +1 ReOIL + 2 RetSP500 + 3 ReEX + 4 TRUMP + . sit. y. Nat. Where: :. er. io. ReIPC: Return on S&P/BMV IPC.. al. iv n C Regression coefficient variable. U h e n gofceach h i independent n. Constant coefficient..  1 -  4: ReOIL:. Return in international price of oil.. ReSP500:. Return of S&P500 index.. ReEX:. Return Change in exchange rate USD/MXN.. TRUMP:. Sentiment of Donald Trump’s tweets towards Mexico.. :. Standard error.. 3.5.2. Linear Regression Assumptions Multiple linear regression makes several key assumptions: 1. Linear: Predictor variables in the regression have a straight-line relationship with the outcome variable. Scatterplots can show whether there is a linear or curvilinear. 14 DOI:10.6814/NCCU202000683.

(24) relationship, another method is by complying with the normality and homoscedasticity assumptions, if the residuals are normally distributed 2. Normality: Multiple regression assumes that the residuals are normally distributed. 3. No Multicollinearity: Multiple regression assumes that the independent variables are not highly correlated with each other. This assumption is tested using variance inflation factor (VIF) values. If the VIF values are under 5 it means the variables are not highly correlated. 4. Homoscedasticity: This assumption states that the variance of error terms are similar across the values of the independent variables. A plot of standardized residuals versus. 政治大. predicted values can show whether points are equally distributed across all values of the independent variables. When the scatter plot shows no pattern, it means. 立. homoscedasticity assumption is met.. ‧ 國. 學 ‧. 3.5.3. Hypothesis Testing MLR. Ho1: 1 = 0; There is no influence of international price of oil on S&P/BMV IPC.. Nat. sit. y. Ha1: 1 ≠ 0; There is influence of international price of oil on S&P/BMV IPC.. er. io. Ho2: 2 = 0; There is no influence of S&P 500 index on S&P/BMV IPC.. al. Ha2: 2 ≠ 0; There is influence of S&P 500 index on S&P/BMV IPC.. n. iv n C Ho3: 3 = 0; There is no influence of h exchange rate USD/MXN e n g c h i U on S&P/BMV IPC. Ha3: 3 ≠ 0; There is influence of exchange rate USD/MXN on S&P/BMV IPC.. Ho4: 4 = 0; There is no influence of Trump’s tweets about Mexico on S&P/BMV IPC. Ha4: 4 ≠ 0; There is influence of Trump’s tweets about Mexico S&P/BMV IPC.. 3.5.4. ARDL Bound Co Integration Test In order to test the relationship and measure the impact of different factors on S&P/BMV IPC this research uses ARDL analysis to compare its findings with MLR analysis. The co integration concept surges in order to know if variables are indeed, related. The ARDL approach has several advantages:. 15 DOI:10.6814/NCCU202000683.

(25) 1. ARDL model contains the lagged value(s) of the dependent variable, the current and lagged values of regressors as explanatory variables. 2. It uses a combination of endogenous and exogenous variables. Explanatory variables are also used as dependent variables. 3. Unlike other methods of testing co integration, ARDL bound can be used irrespective of whether the variables of the study are I(0) or I(1) or a combination of both. However, the series must not be I(2), otherwise we cannot use ARDL bound test. 4. Once the lag order of the model is identified, co integration can be tested using bound test procedure of ordinary least square (OLS). From the bound test result, if variables are cointegrated, it is necessary to specify both short-run (ARDL) and long run (VECM). 政治大. models. If they are not cointegrated it is necessary to specify the short-run (ARDL). 立. model only.. ‧ 國. 學. 5. This model performs better and avoids issues of weak power in modelling the cointegrating relationship with small samples (Romilly et al., 2001; Pesaran, 1997).. ‧. Initially unit root test (ADF and Phillip Person) are applied on both the series at the levels and. Nat. sit. y. at the first difference to check the stationary of the variables and to ensure that none of the. io. er. variables is integrated of order (2). In the second step, we develop the ARDL model according to specifications based on the Akaike information criterion (AIC). This is followed by bounds. n. al. Ch. i Un. v. testing to check for a co-integrating relationship between the dependent and the explanatory variables.. engchi. Regressions are estimated as following: IPCt =  +1IPCt-1 + 2OILt-1 + 3SP500t-1 + 4EXt-1 + 5TRUMPt-1 + ∑𝑛𝑖=1 ∝6 IPCt-1 + ∑𝑛𝑖=1 ∝7 OILt-1 + ∑𝑛𝑖=1 ∝8 SP500t-1 + ∑𝑛𝑖=1 ∝9 EXt-1 + ∑𝑛𝑖=1 ∝10 TRUMPt-1 + 1t  OILt =  +1IPCt-1 + 2OILt-1 + 3SP500t-1 + 4EXt-1 + 5TRUMPt-1 + ∑𝑛𝑖=1 β6 IPCt-1 + ∑𝑛𝑖=1 β7 OILt-1 + ∑𝑛𝑖=1 β8SP500t-1 + ∑𝑛𝑖=1 β9 EXt-1 + ∑𝑛𝑖=1 β10 TRUMPt-1 + 2t SP500t =  +1IPCt-1 + 2OILt-1 + 3SP500t-1 + 4EXt-1 + 5TRUMPt-1 + ∑ni=1 γ6 IPCt-1 +. 16 DOI:10.6814/NCCU202000683.

(26) ∑ni=1 γ7 OILt-1 + ∑ni=1 γ8 SP500t-1 + ∑ni=1 γ9 EXt-1 + ∑ni=1 γ10 TRUMPt-1 + 4t EXt =  + 1IPCt-1 + 2OILt-1 + 3SP500t-1 + 4EXt-1 + 5TRUMPt-1 + ∑ni=1 μ6 IPCt-1 + ∑ni=1 μ7 OILt-1 + ∑ni=1 μ8 SP500t-1 + ∑ni=1 μ9 EXt-1 + ∑ni=1 μ10TRUMPt-1 + 3t Where: IPC:. S&P/BMV IPC value. IPC:. Difference of S&P/BMV IPC value. OIL:. International price of oil. OIL:. 治政Difference in international price of oil 大. 立. ReSP500:. ‧ 國. y. sit. Difference of exchange rate USD/MXN Sentiment of Donald Trump’s tweets towards Mexico. Difference of sentiment of Donald Trump’s tweets towards Mexico. er. al. n. TRUMP:. io. TRUMP:. Exchange rate USD/MXN. Nat. EX:. Difference of S&P 500 value. ‧. EX:. 學. SP500:. S&P 500 value. Ch. :. engchi. i Un. v. Standard error. Sentiment of Trump’s tweets is treated as an exogenous variable, it does not respond to the return of IPC, international oil price, return of S&P 500 nor the exchange rate USD/MXN.. 3.5.5. Diagnostic Tests ARDL The diagnostic tests serial correlation, normality and heteroscedasticity test are conducted to check the goodness of fit of the model. The Cumulative sum (CUSUM), and the CUSUMSQ (cumulative sum of squares) stability tests were applied and the statistical are inside the critical bound at 5%, indicating that the regression equation is stable.. 17 DOI:10.6814/NCCU202000683.

(27) 3.5.6. Hypothesis Testing ARDL IPC Equation Ho1: 1 = 2 = 3 = 4 = 5 = 0; There is no long run relationship among the variables in the equation. Ha1: 1 ≠ 2 ≠ 3 ≠ 4 ≠ 5 ≠ 0; There is long run relationship among the variables in the equation. OIL Equation Ho1: 1 = 2 = 3 = 4 = 5 = 0; There is no long run relationship among the variables in the equation.. 政治大. Ha1: 1 ≠ 2 ≠ 3 ≠ 4 ≠ 5 ≠ 0; There is long run relationship among the variables in the equation. SP500 Equation. 立. Ho1: 1 = 2 = 3 = 4 = 5 = 0; There is no long run relationship among the variables in the. ‧ 國. 學. equation.. Ha1: 1 ≠ 2 ≠ 3 ≠ 4 ≠ 5 ≠ 0; There is long run relationship among the variables in the equation.. ‧. EX Equation. y. sit. equation.. Nat. Ho1: 1 = 2 = 3 = 4 = 5 = 0; There is no long run relationship among the variables in the. er. io. Ha1: 1 ≠ 2 ≠ 3 ≠ 4 ≠ 5 ≠ 0; There is long run relationship among the variables in the. al. n. iv n C U the joint significance of lagged In order to test the hypothesis, WaldhF-test e nisgused. c h iIt detects equation.. values of variables in the equation and provide us with the F-statistic and upper and lower critical values. The evidence of co integration is found when F-statistics is above the upper critical value and vice versa. In case F-statistics is between upper and lower bound values, result is inconclusive. After the co integration is established among variables, cointegrating equation for IPC is estimated using long-term error, also known as error correction term in the Error correction model.. 18 DOI:10.6814/NCCU202000683.

(28) 4. Results and Discussions 4.1. Multiple Linear Regression 4.1.1. Descriptive statistics. 政治大. 立. ‧ 國. 學. Mean Std. Dev. Skewness Kurtosis Range Maximum Minimum Jarque-Bera Probability. Table 1 Result of Descriptive Statistics MLR REIPC REOIL RESP500 REEX 0.00079 0.00543 0.00813 0.00477 0.03567 0.09031 0.03455 0.03458 -0.502 -0.2274 -0.4918 0.3155 3.73454 3.04591 3.76545 3.00268 0.19146 0.42259 0.17476 0.15922 0.07911 0.20241 0.08299 0.09057 -0.1123 -0.2202 -0.0918 -0.0687 3.86909 0.52247 3.88341 0.99544 0.14449 0.7701 0.14346 0.60792. TRUMP -1.3167 3.56295 -0.903 3.27169 15 5 -10 8.33854 0.01546. From table 1, among all variables sentiment of Trump’s tweets presents the higher volatility. ‧. return of oil price, return of exchange rate USD/MXN, return of S&P 500 index and return of. al. er. io. sit. y. Nat. S&P/BMV IPC have very similar variability.. n. 4.1.2. Correlation Matrix. ReIPC ReOIL ReSP500 ReEX TRUMP. Ch. engchi. i Un. v. Table 2 Correlation Matrix MLR ReIPC ReOIL ReSP500 ReEX 1 0.354 0.006 0.425 0.001 -0.449 0.000 -0.121 0.356. TRUMP. 1 0.454 0.000 -0.180 0.170 0.060 0.650. 1 -0.327 0.011 -0.051 0.699. 1 0.100 0.446. 1. 19 DOI:10.6814/NCCU202000683.

(29) If p-value ≤ α it means the correlation is statistically significant, with a significance level of 5%, ReSP500 shows a significant correlation with Return of IPC and ReOIL. ReEX shows a negative significant relationship with Return of IPC and Return of SP500. ReOIL shows a positive significant correlation with Return of IPC and Return of S&P 500.. 4.1.3. Multiple Linear Regression Equation Table 3: Multiple Linear Regression Coefficients Coefficient Std. Error T P-value -0.001034 0.00439 -0.236 0.8147 0.078644 0.049 1.60 0.1169 0.228914 0.133895 1.710 0.093 -0.342891 0.121514 -2.822 0.0066 -0.000886 0.00112 -0.791 0.4325 R-square = 32.5% Adjusted R-square = 27.6%. 立. 政治大. VIF 1.277 1.44763 1.15249 1.1618. 學. ‧ 國. Model Constant ReOIL ReSP500 ReEX TRUMP. ‧. The estimated equation is. Return IPC = − 0.001034(−0.236) + 0.078644(1.60) ReOIL + 0.228914∗(1.71) ReSP500 -. y. Nat. sit. 0.342891∗(−2.99)ReEX - 0.000886(−0.791) TRUMP + . n. al. er. io. The R-square means that 32.7% of the variation in return of S&P/BMV IPC is explained by the. v. return of international price of oil, return of S&P 500, return in the exchange rate USD/MXN and Donald Trump’s tweets.. Ch. engchi. i Un. 4.1.4. Hypotheses Test Result MLR T- test is used to prove the influence of each independent variable on de dependent variable. In this research que use a level of significance α=.05, having a confidence level of 95%. The decision rule is as follows: 1) If the significance level (P- value) ≤ 0.05, Ho is rejected and Ha is accepted. 2) If the significance level (P-value) > 0.05, Ho is accepted.. 20 DOI:10.6814/NCCU202000683.

(30) Influence of International Crude Oil Price From table 5, return of international price of oil has a t-value of 1.60 and a p-value of 0.1169. Since 0.1165 > 0.05, there is no statistical evidence to reject Ho, which implies that change in the international price of oil is not statistically significant related to change of the return of S&P/BMV IPC. Influence of S&P 500 Index From table 5, return of S&P 500 index has a t-value of 1.71 and a p-value of 0.260. Since 0.093 > 0.05, there is no statistical evidence to reject Ho, which implies that return of S&P 500 index. 政治大. is not statistically significant related to change of the return of S&P/BMV IPC. It is significant. 立. 學. ‧ 國. at a confidence level of 90%.. Influence of Exchange rate USD/MXN. ‧. From table 5, return of exchange rate USD/MXN has a t-value of -2.82 and a p-value of 0.006. Since 0.006 < 0.05, there is sufficient statistical evidence to reject Ho, which implies that return. y. Nat. sit. of exchange rate USD/MXN is statistically significant related to change of the return of. al. er. io. S&P/BMV IPC. When the monthly exchange rate USD/MXN increases (MXN devaluates) by. n. one unit, then the monthly return of S&P/BMV IPC decrease by 3.42%.. Ch. Influence of Donald Trump Tweets. engchi. i Un. v. From table 5, Donald Trump tweets have t-value of -0.791 and a p-value of 0.4325. Since 0.4325 > 0.05, there is no statistical evidence to reject Ho, which implies that Donald Trump Tweets are not statistically significant related to change of the return of S&P/BMV IPC.. 4.1.5. Overall Significance Test MLR The objective is to test if there is a significant relationship between the dependent variable return of S&P/BMV IPC and the independent variables international oil price, S&P 500, exchange rate USD/MXN and Donald Trump tweets all together.. 21 DOI:10.6814/NCCU202000683.

(31) Ho: β1 = β2 = β3 = β4 = 0 (No linear relationship). Ha: At least one βi ≠ 0 (Linear relationship; at least one independent variable affects the dependent variable). The decision rule is as follows: 1) If the significance level (P- value) ≤ 0.05, Ho is rejected and Ha is accepted. 2) If the significance level (P-value) > 0.05, Ho is accepted. Table 4: ANOVA MLR Model. DF. Regression. ‧ 國. Mean Square. F. 政治大 4 0.024391 0.0061 55 0.050677 0.0009 59 0.075068. 6.62. P 0.000. 學. Residual Total. 立. Sum of Squares. ‧. From table 5, overall model has a F-value of 6.62 and a p-value of 0.000. Since 0.000 < 0.05, there is sufficient statistical evidence to reject Ho. At least one independent. Nat. sit. y. variable has linear relationship with S&P/BMV IPC. From previous sections, that variable is. er. io. the exchange rate USD/MXN. If we do the test with a confidence level of 90%, then the variables that have a linear relationship with S&P/BMV IPC are the return of S&P 500. n. al. Ch. i Un. (positive) and the exchange rate USD/MXN (negative).. engchi. v. 22 DOI:10.6814/NCCU202000683.

(32) 4.1.6. Assumptions Test MLR Normality Null hypothesis: The residuals are normally distributed. 9 Series: Res iduals Sample 2015M01 2019M12 Obs ervations 60. 8 7 6. Mean Median Maximum Minimum Std. Dev. Skewnes s Kurtos is. 5 4 3. 立. 2. -0.04. -0.02. 0.00. 0.02. 學. 0 -0.06. ‧ 國. 1. 政治大 0.04. 0.06. 2.31e-19 0.000952 0.072252 -0.057559 0.029308 0.311101 3.256151. Jarque-Bera 1.131874 Probability 0.567828. ‧. Figure 7: Normal Probability Plot MLR. y. Nat. io. null hypothesis that residuals are normally distributed.. n. al. Multicollinearity Test. Ch. engchi. er. sit. The normal probability plot shows a probability 0.567828, since 0.567828>0.05, we accept the. i Un. v. Table 5: Multicollinearity Test MLR Predictor Coefficient VIF Constant -0.001034 ReOIL 0.078644 1.277 ReSP500. 0.228914. 1.447627. ReEX TRUMP. -0.342891 -0.000886. 1.152487 1.161803. The predictor variables show a VIF < 5, which indicates that multicollinearity is not affecting the regression analysis results.. 23 DOI:10.6814/NCCU202000683.

(33) Homoscedasticity Test Heteroskedasticity Test: White Null hypothesis: Homoskedasticity Table 6: Heteroskedasticity Test Result MLR F-statistic 1.151325 Prob. F(4,50) 0.3436 Obs*R-squared 15.82356 Prob. Chi-Square(4) 0.3243 Scaled explained SS 14.9991 Prob. Chi-Square(4) 0.3782 The prob. F is 0.3436 since 0.3436>0.05, we accept the null hypothesis, the model presents homoskedasticity.. 立. 政治大. 4.2. ARDL Bound Co Integration Test. ‧ 國. 學. 4.2.1. Descriptive statistics. io. y. 365.52 1.48847. 3.56295. 2.03138 2.75319 1.75682 2.93813 0.30856 0.1654 0.20225 -0.7868 10347.4 40.53 1310.76 5.8487 51210.5 74.15 3230.78 20.832 40863.1 33.62 1920.03 14.9833 3.29763 0.42586 4.2728 6.20061 0.19228 0.80821 0.11808 0.04504. 3.27169 -0.903 15 5 -10 8.33854 0.01546. n. al. er. 2747.02 9.18582. sit. Nat. Mean Standard Deviation Kurtosis Skewness Range Maximum Minimum Jarque-Bera Probability. ‧. Table 7: Result of Descriptive Statistics ARDL IPC OIL SP500 EX TRUMP 45693.6 53.4575 2459.86 18.4299 -1.3167. Ch. engchi. i Un. v. From table 7, among all variables IPC presents the higher volatility followed by SP500, Oil, Trump and exchange rate.. 24 DOI:10.6814/NCCU202000683.

(34) IPC. OIL. 52,000. 80. 50,000. 70. 48,000. 60. 46,000 50. 44,000. 40. 42,000 40,000. 30. I II III IV I II III IV I II III IV I II III IV I II III IV 2015 2016 2017 2018 2019. I II III IV I II III IV I II III IV I II III IV I II III IV 2015 2016 2017 2018 2019. SP500. EX. 3,600. 21 20. 3,200. 19. 2,800. 18. 2,400. 立. 17 16 15 14. I II III IV I II III IV I II III IV I II III IV I II III IV 2015 2016 2017 2018 2019. 學. ‧ 國. 2,000 1,600. 政治大 I II III IV I II III IV I II III IV I II III IV I II III IV 2015 2016 2017 2018 2019. TRUMP. -12. y. sit. al. n. -8. io. -4. Nat. 0. er. 4. ‧. 8. C II III IVhI. I II III IV I II III IV I II III IV I 2015 2016 2017 2018. II III IV e2019 ngchi. i Un. v. Figure 8: Monthly Movement of Variables ARDL Figure 8 shows monthly movements of the variables under study. It is noticed that there is a steep fall in oil prices starting the second quarter of 2015 and in the last quarter of 2018. Exchange rate presents an upward trend. The high appreciation of the USD against MXN during the last quarter of 2016 coincides with the victory of Donald Trump in the U.S. presidential election on November 8 2016. IPC shows and upward trend until the middle of 2017, it shows a downward trend after. ReSP500 maintains and upward trend.. 25 DOI:10.6814/NCCU202000683.

(35) 4.2.2. Correlation Matrix ARDL. Table 8: Correlation Matrix ARDL IPC OIL SP500 EX 1. IPC OIL. 0.221 0.090 0.024 0.854 0.113 0.389 0.145 0.268. SP500 EX TRUMP. 立. TRUMP. 1 0.679 0.000 0.287 0.026 0.041 0.753. 1 0.588 0.000 0.163 0.213. 政治大. 1 0.265 0.040. 1. ‧ 國. 學. If p-value ≤ α it means the correlation is statistically significant, with a significance level of 5%, EX shows a significant positive correlation with of SP500 and OIL, while SP500 shows a. ‧. positive significant correlation with OIL. IPC shows a significant correlation with OIL at a 90%. er. io. sit. y. Nat. confidence level. TRUMP and EX show a positive significant correlation.. 4.2.3. Unit Root Test ARDL a. n. iv l C n hengchi U integration of the variables. In case of ARDL test, the variables must not be I(2), otherwise the Before proceeding with the ARDL-Bounds test, unit root test is used to check the order of. results will be spurious. We use two unit root tests namely ADF and PP, the tests are applied on the levels and at first difference by including constant and time trend. Table 3 shows the results. In case of ADF and PP, null hypothesis is that the series is non-stationary.. 26 DOI:10.6814/NCCU202000683.

(36) Table 9: Results of Unit Root Test ARDL Unit Root (Levels) ADF PP IPC -2.664 -2.6543 0.2549 0.2589 OIL -2.497 -2.6851 0.3286 0.2464 SP500 -3.0957 -3.6394 0.1169 0.0357 EX -3.1886 -2.9867 0.0976 0.1446 TRUMP -5.2316 -5.1665 0.000* 0.000* Unit Root(First Difference) IPC -8.5617 -9.9042 0.000* 0.000* OIL -7.024 -7.0507 0.000* 0.000* SP500 -9.5402 -10.655 0.000* 0.000* EX -8.7 -12.8503 0.000* 0.000* TRUMP -12.3967 -16.1882 0.000* 0.000* *Indicates that null hypothesis is not accepted at 5% significance level.. 立. 政治大. ‧. ‧ 國. 學. er. io. sit. y. Nat. al 4.2.4. ARDL Co Bound Integration Test n. iv n C h e n gincTable Results of ARDL Bound test are presented h i U4. Cointegration is found among the variables when S&P/BMV IPC value is kept as dependent variable. This indicates that whenever there is a shock in the system, oil price, S&P 500 value, exchange rate and Trump’s tweets will move first and their movement is followed by the stock market.. Dependent Variable IPC OIL SP500 EX. Table 10: ARDL Co Bound Integration Test Result Co integration Null FHypothesis Lag Structure Statistics 5.706 1 = 2 = 3 = 4 = 5 = 0 (5, 1, 5, 6, 2) 8.427 1 = 2 = 3 = 4 = 5 = 0 (1, 0, 5, 6 , 3) (5, 3. 4. 1, 1) 1.780 1 = 2 = 3 = 4 = 5 = 0 (1, 1, 0, 0, 2) 2.156 1 = 2 = 3 = 4 = 5 = 0. Outcome Co integration Co integration Inconclusive No Co integration. 27 DOI:10.6814/NCCU202000683.

(37) 4.2.5. Long Run Equation ARDL Long run coefficient estimates of co integration equation for S&P/BMV IPC is presented in Table 11. Table 11: Long Run Equation IPC ARDL IPC Equation. Coefficient. Std. Error. C 8563.1100 4163.3200 IPC(-1) 0.7993 0.0788 OIL(-1) -6.3558 32.4770 SP500(-1) -1.2855 0.9872 EX(-1) 226.6330 178.7507 TRUMP(-1) 23.0428 59.4888 *Indicates statistical significance at 5% level.. 立. t-Statistic. 政治大. p-value. 2.0568 10.1379 -0.1957 -1.3022 1.2679 0.3873. 0.0446 0.000* 0.8456 0.1985 0.2104 0.7001. At a level of significance of 5%, the lagged value of S&P/BMV IPC influence positively the. ‧ 國. 學. current value of S&P/BMV IPC.. 4.2.6. Error Correction Model ARDL. ‧. After the co integrated relationship is established, Error correction model is estimated (see Table 6) to study the short run dynamics. Results indicate that error correction term is. y. Nat. sit. significant at 5% level of significance. This indicates that after a shock in the system, IPC value. er. io. return to their equilibrium.. al. ∆𝐼𝑃𝐶𝑡 = 12.4387(0.0593) + 0.7323∗(2.4117) ∆𝐼𝑃𝐶𝑡−1 − 14.1393(−0.2921) ∆𝑂𝐼𝐿𝑡−1. n. iv n C − 2.5513(−0.9238) U (−0.2984) ∆𝐸𝑋𝑡−1 h e∆𝑆𝑃500 n g c𝑡−1h −i 105.3109. − 23.4608(−0.4482) ∆𝑇𝑅𝑈𝑀𝑃𝑡−1 − 0.9984∗(−3.0616) 𝐸𝐶𝑀𝑡−1 Table 12: Error Correction Model ARDL IPC Equation Coefficient. Std. Error. 12.4387 209.6810 C 0.7323 0.3037 IPC(-1) 14.1393 48.4035 OIL(-1) -2.5513 2.7618 SP500(-1) -105.3109 352.9774 EX(-1) -23.4608 52.3408 TRUMP(-1) -0.9984 0.3261 ECM(-1) *Indicates statistical significance at 5% level.. t-Statistic. p-value. 0.0593 2.4117 0.2921 -0.9238 -0.2984 -0.4482 -3.0616. 0.9529 0.0195* 0.7714 0.36 0.7666 0.6559 0.0035*. 28 DOI:10.6814/NCCU202000683.

(38) 4.2.7. Hypotheses Test Result ARDL T- test is used to prove the influence of each independent variable on de dependent variable. In this research que use a level of significance α=.05, having a confidence level of 95%. The decision rule is as follows: 1) If the significance level (P- value) ≤ 0.05, Ho is rejected and Ha is accepted. 2) If the significance level (P-value) > 0.05, Ho is accepted. Influence of International Crude Oil Price Long Run Equation. 政治大. From table 5, international price of oil has a t-value of -0.1957 and a p-value of 0.8456. Since. 立. 0.8456 > 0.05, there is no statistical evidence to reject Ho, which implies that change in the. ‧ 國. 學. international price of oil is not statistically significant related to S&P/BMV IPC value. Error Correction Model. ‧. From table 6, international price of oil has a t-value of 0.2921 and a p-value of 0.7714. Since. y. Nat. 0.7714 > 0.05, there is no statistical evidence to reject Ho, which implies that change in the. n. al. er. io. sit. international price of oil is not statistically significant related to S&P/BMV IPC value.. Influence of S&P 500 Index Long Run Equation. Ch. engchi. i Un. v. From table 5, S&P 500 value has a t-value of -1.3022 and a p-value of 0.1985. Since 0.1985 > 0.05, there is no statistical evidence to reject Ho, which implies that S&P 500 value is not statistically significant related to S&P/BMV IPC value. Error Correction Model From table 6, S&P 500 value has a t-value of -0.9238 and a p-value of 0.36. Since 0.36 > 0.05, there is no statistical evidence to reject Ho, which implies that S&P 500 value is not statistically significant related to S&P/BMV IPC value.. 29 DOI:10.6814/NCCU202000683.

(39) Influence of Exchange rate USD/MXN Long Run Equation From table 5, international exchange rate USD/MXN of 1.2679 and a p-value of 0.2104. Since 0.2104 > 0.05, there is no statistical evidence to reject Ho, which implies that change in the international price of oil is not statistically significant related to S&P/BMV IPC value. Error Correction Model From table 6, international price of oil has a t-value of -0.2984 and a p-value of 0.7666. Since 0.7666 > 0.05, there is no statistical evidence to reject Ho, which implies that exchange rate. 政治大. USD/MXN is not statistically significant related to S&P/BMV IPC value.. 立. Influence of Donald Trump Tweets. ‧ 國. 學. Long Run Equation. ‧. From table 5, Donald Trump tweets have t-value of 0.3873 and a p-value of 0.7001. Since 0.7001 > 0.05, there is no statistical evidence to reject Ho, which implies that Donald. Nat. sit. y. Trump Tweets are not statistically significant related S&P/BMV IPC value.. er. io. Error Correction Model. al. n. iv n C h e nevidence Since 0.6559 > 0.05, there is no statistical g c h i toUreject Ho, which implies that Donald. From table 6, Donald Trump tweets have t-value of -0.44482 and a p-value of 0.6559. Trump Tweets are not statistically significant related S&P/BMV IPC value.. 4.2.8. Overall Significance Test ARDL The objective is to test if there is a significant relationship between the endogenous variable S&P/BMV IPC value and the exogenous variables all together. Ho: β1 = β2 = β3 = β4 = 0 (No linear relationship). Ha: At least one βi ≠ 0 (Linear relationship; at least one independent variable affects the dependent variable). The decision rule is as follows: 1) If the significance level (P- value) ≤ 0.05, Ho is rejected and Ha is accepted.. 30 DOI:10.6814/NCCU202000683.

(40) 2) If the significance level (P-value) > 0.05, Ho is accepted. Table 13: Overall Significance Test ARDL Model F P Long Run 24.55391 0.000 ECM 1.9041 .0980 From table 8, Long Run model has a F-value of 24.55391 and a p-value of 0.000. Since 0.000 < 0.05, there is sufficient statistical evidence to reject Ho. At least one independent variable has linear relationship with S&P/BMV IPC. From previous sections, the lag of. 政治大. S&P/BMV IPC has influence on the current S&P/BMV IPC.. 立. From table 13, ECM has a F-value of 1.9819 and a p-value of 0.0877. Since 0.0877 > 0.05,. ‧ 國. If we do the test with a confidence level. 學. there not is sufficient statistical evidence to reject Ho.. of 90%, then the variables that have a linear relationship with S&P/BMV IPC are the difference. ‧. lag value of S&P/BMV IPC and the ECM term.. y. Nat. sit. 4.2.9. ECM Diagnostics tests. n. al. er. io. Table 14: Breusch-Godfrey Serial Correlation LM Test Null hypothesis: No serial correlation at up to 1 lag F-statistic Obs*R-squared. Ch. 0.265577 0.306442. i Un. v. Prob. F(1,50) Prob. Chi-Square(1). engchi. 0.6086 0.5799. Prob F > .05, accept null hypothesis, the model presents no serial correlation.. Table 15: Heteroskedasticity White Test Null hypothesis: Homoskedasticity F-statistic Obs*R-squared Scaled explained SS. 1.075032 28.52140 30.04991. Prob. F(27,30) Prob. Chi-Square(27) Prob. Chi-Square(27). 0.4216 0.3845 0.3119. Prob F > .05, accept null hypothesis, the model presents homoskedasticity.. 31 DOI:10.6814/NCCU202000683.

(41) Stability Test: Cumsum 30. 1.4 1.2. 20. 1.0. 10. 0.8 0.6. 0. 0.4. -10. 0.2. -20. 0.0. -30. -0.2. IV 2015. I. II. III IV. 2016. I. II. III IV. I. 2017. II. III IV. I. II. 2018. CUSUM. III IV. IV 2015. 2019. I. II. III. IV. I. II. 2016. III. IV. I. 2017. III. IV. 2018. CUSUM of Squares. 5% Significance. II. I. II. III. IV. 2019. 5% Significance. Figure 9: CUMSUM Stability Test. Normality test. 立. 政治大. Null hypothesis: Residuals are normally distributed. 10. sit. n. al. 4 2 0 -5000. er. io. 6. y. Nat. 8. ‧. ‧ 國. 學. 12. -4000. -3000. -2000. Ch -1000. engchi 0. 1000. i Un. v. 2000. 3000. Series: Residuals Sample 2015M03 2019M12 Observations 58 Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis. 5.88e-14 106.1596 2830.771 -4574.245 1436.399 -0.509666 3.725323. Jarque-Bera 3.782405 Probability 0.150890. Figure 10: Normality Plot ECM Probability > .05, accept null hypothesis, residuals are normally distributed.. 32 DOI:10.6814/NCCU202000683.

(42) 5. Conclusions According to our research results, from multiple linear regression with a 95% confidence level, we found that only exchange rate USD/MXN influences negatively the S&P/BMV IPC. This finding implies that whenever the USD appreciates over the MXN, investors will prefer to withdraw their assets from S&P/BMV IPC seeking for higher rate of returns in assets not offered in MXN. With a 90% confidence level, besides the monthly change in exchange rate USD/MXN, the return of S&P 500 index also influences the S&P/BMV IPC but in a positive way, implying that S&P/BMV IPC follows S&P 500 index.. 治政 the probability of spurious regression of non-stationary 大 series, indicate with 95% confidence 立 level, that in the long run, only the lag value of S&P/BMV IPC influences positively its current These findings contrast with results from auto-regressive distributed lag model, which reduces. ‧ 國. 學. value. From the error correction model, again the lagged value of S&P/BMV IPC is significant, while the error correction term is negative and significant, indicating that after a shock in the. ‧. system, S&P/BMV IPC value returns to its equilibrium with a speed of 99.84%.. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. 33 DOI:10.6814/NCCU202000683.

(43) 6. References BMV Group. (2015). About Us. Mexico City, Mexico. Caplinger, D. (2020, January 18). What Is a Stock Market Index? The Motley Fool. Retrieved from https://www.fool.com/knowledge-center/what-is-a-stock-index.aspx Chapman, B. 2017. Donald Trump wipes $5.7bn off the value of Amazon with single tweet. Independent. Aug 16. Retrieved from https://www.independent.co.uk/news/business/news/donald-trump-amazon-value-5-. 政治大 Global Energy Statistical Yearbook 立. billion-drop-off-shares-tweet-social-media-jeff-bezos-a7896266.html Enerdata.. (2019).. 2019.. Retrieved. from. ‧ 國. 學. https://yearbook.enerdata.net/total-energy/world-consumption-statistics.html Fandos, N. 2017. Trump's View of Syria: How it Evolved, in 19 Tweets. NYTimes. April 7.. ‧. Retrieved from https://www.nytimes.com/2017/04/07/us/politics/donald-trump-syria-. sit. y. Nat. twitter.html. n. al. er. io. Federal Reserve System. 2020, January 30. About the Fed. Washington DC. Retrieved from. i Un. https://www.federalreserve.gov/aboutthefed.htm. Ch. engchi. v. FXCM. 2017. What Is President Trump’s Impact On The Mexican Peso? February 1. Retrieved from https://www.fxcm.com/uk/insights/president-trumps-impact-mexicanpeso/ Goodman, P. S., Bradsher, K., & Gough, N. 2017. The Fed Acts. Workers in Mexico and Merchants in Malaysia Suffer. NY Times. March 6. Retrieved from https://www.nytimes.com/2017/03/16/business/federal-reserve-interest-rates-chinamexico.html Gunzberg, J., & Edwards, T. (2018, June 20). Why is the S&P 500® Relevant Globally?. 34 DOI:10.6814/NCCU202000683.

(44) Retrieved from https://www.spglobal.com/en/research-insights/articles/why-is-the-sp500-relevant-globally Guo, S., Jiao, Y., & Xu, Z. 2019. Trump's Effect on the Chinese Stock Market. Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3541523 Gutierrez, R. d., & Ortiz, E. (2011). El efecto de la volatilidad del peso mexicano en los rendimientos y riesgo de la Bolsa Mexicana de Valores. Contaduría y Administración 58, 89-119. Retrieved from. 政治大. http://www.scielo.org.mx/scielo.php?script=sci_arttext&pid=S018610422013000300005 March).. Balanza. ‧ 國. (2020,. Comercial. de. https://www.inegi.org.mx/temas/balanza/. México.. 學. INEGI.. 立. Retrieved. from. ‧. Ingram, M. 2017. Here's Why Facebook Is Partly to Blame for the Rise of Donald Trump.. y. sit. n. al. er. io. trump/. Nat. Fortune. May 24. Retrieved from https://fortune.com/2016/11/10/facebook-blame-. i Un. v. Kenton, W. (2019, May 18). S&P 500 Index – Standard & Poor's 500 Index. Investopedia.. Ch. engchi. Retrieved from https://www.investopedia.com/terms/s/sp500.asp López-Juárez, G. I., Ladrón de Guevara-Cortés, R., & Madrid-Paredones, R. M. (2019). Factores que Explican el Comportamiento del Mercado Accionario Mexicano. Clío América, 268-278. O'Sullivan, A. 2003. Economics: Principles in action. New Jersey: Pearson Prentice Hall. Sánchez, M. 2018. El S&P/BMV IPC cumple 40 años. Retrieved from https://us.spindices.com/ Sandoval, A. 2020. La caída de los precios del petróleo amenaza las finanzas públicas de México. Alto Nivel. March 03 Retrieved from. 35 DOI:10.6814/NCCU202000683.

(45) https://www.altonivel.com.mx/finanzas/la-caida-de-los-precios-del-petroleo-amenazalas-finanzas-publicas-de-mexico/ Sardy, M., & Cova, S. C. (2019, June). How Presidential Tweets Affect Financial Markets. US. Retrieved from https://www.researchgate.net/publication/333852219 Singhal, S., Choudhary, S., & Biswal, P. C. (2019). Return and volatility linkages among International crude oil price, gold price, exchage rate and stock markets: Evidence from Mexico.. Resources. Policy,. 255-261.. Retrieved. from. 政治大 Time, 2015. Here's Donald Trump's Presidential Announcement Speech. June 16 Retrieved 立 www.elsevier.com/locate/resourpol. ‧ 國. 學. from https://time.com/3923128/donald-trump-announcement-speech/ Trump, D. J. (2017, August 17). Amazon is doing great damage to tax paying retailers. Towns,. ‧. cities and states throughout the U.S. are being hurt - many jobs being lost! Retrieved. sit. y. Nat. from https://twitter.com/realDonaldTrump/status/897763049226084352. n. al. er. io. U.S. Department of Commerce. International trade in goods and services, 2018 Retrieved from. i Un. v. https://www.commerce.gov/tags/international-trade-goods-and-services. Ch. engchi. U.S. Energy Information Administration. Production of Crude Oil including Lease Condensate, 2019. Retrieved from https://www.eia.gov/. 36 DOI:10.6814/NCCU202000683.

(46) 7. Appendix A Research Data IPC OIL SP500 EX 40950.58 48.24 1994.99 14.9903 44190.17 49.76 2104.5 14.9833 43724.78 47.6 2067.89 15.2609 44582.39 59.68 2085.51 15.3427 44703.62 60.29 2107.39 15.3737 45053.7 59.47 2063.112 15.7374 44752.93 47.12 2103.835 16.1077 43721.96 49.2 1972.185 16.7502 42632.54 45.09 1920.027 16.9173 44542.76 46.59 2079.36 16.492 43418.55 41.65 2080.407 16.5715 42977.5 37.04 2043.937 17.17 43630.77 33.62 1940.24 18.1 43714.93 33.75 1932.226 18.125 45881.08 38.34 2059.741 17.2786 45784.77 45.92 2065.296 17.17 45459.45 49.1 2096.964 18.4664 45966.49 48.33 2098.855 18.268 46660.67 41.6 2173.604 18.7465 47541.32 44.7 2170.947 18.7801 47245.8 48.24 2168.272 19.375 48009.28 46.86 2126.153 18.859 45315.96 49.44 2198.81 20.567 45642.9 53.72 2238.827 20.715 47001.06 52.81 2278.867 20.832 46856.79 54.01 2363.638 20.1 48541.56 50.6 2362.718 18.72 49261.33 49.33 2384.196 18.81 48788.44 48.32 2411.797 18.6192 49857.49 46.04 2423.409 18.1228 51011.87 50.17 2470.3 17.7917 51210.48 47.23 2471.65 17.882 50346.06 51.67 2519.36 18.25 48625.53 54.38 2575.264 19.147 47092.44 57.4 2647.58 18.629. y. sit. io. n. er. Nat. al. ‧. ‧ 國. 立. 政治大. 學. Date Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 Jul-15 Aug-15 Sep-15 Oct-15 Nov-15 Dec-15 Jan-16 Feb-16 Mar-16 Apr-16 May-16 Jun-16 Jul-16 Aug-16 Sep-16 Oct-16 Nov-16 Dec-16 Jan-17 Feb-17 Mar-17 Apr-17 May-17 Jun-17 Jul-17 Aug-17 Sep-17 Oct-17 Nov-17. Ch. engchi. i Un. v. 37 DOI:10.6814/NCCU202000683.

(47) 49354.42 50456.17 47437.93 46124.85 48358.16 44662.55 47663.2 49698.01 49547.68 49504.16 43942.55 41732.78 41640.27 43987.94 42823.81 43281.28 44597.32 42749.16 43161.17 40863.09 42622.5 43011.27 43337.28 42820.18 43541.02. 政治大. n. engchi. 19.648 18.5905 18.8305 18.15 18.7125 19.908 19.909 18.643 19.08 18.708 20.3295 20.396 19.6403 19.1058 19.2775 19.4215 18.9402 19.612 19.2162 19.1265 20.0575 19.7275 19.228 19.562 18.925. y. sit. io. Ch. er. Nat. al. 2673.611 2823.81 2713.831 2640.866 2648.049 2705.275 2718.37 2816.289 2901.518 2913.978 2711.744 2760.174 2506.847 2704.101 2784.491 2834.4 2945.831 2752.063 2941.761 2980.379 2926.458 2976.737 3037.564 3140.981 3230.782. ‧. ‧ 國. 立. 60.42 64.73 61.64 64.94 68.57 67.04 74.15 68.76 69.8 73.25 65.31 50.93 45.41 53.79 57.22 60.14 63.91 53.5 58.47 58.58 55.1 54.07 54.18 55.17 61.06. 學. Dec-17 Jan-18 Feb-18 Mar-18 Apr-18 May-18 Jun-18 Jul-18 Aug-18 Sep-18 Oct-18 Nov-18 Dec-18 Jan-19 Feb-19 Mar-19 Apr-19 May-19 Jun-19 Jul-19 Aug-19 Sep-19 Oct-19 Nov-19 Dec-19. i Un. v. 38 DOI:10.6814/NCCU202000683.

(48)