Bachelor Thesis Economics and Business Specialisation: Economics and Finance
Faculty of Economics and Business Academic year: 2017 – 2018
Influence of United States’ imposing
import tariff on Chinese products
Student Name: Yi Wang Student Number: 11089776 Supervisor: Péter Foldvari
Abstract
The United States and China are the top two economies in the world, and if they impose import taxes on each other, an adverse spillover effect on other countries will result. Hence, it is vital to verify whether an increase in tax duty is effective for the United States to reduce the trade deficit and unemployment rate. This paper thus uses time series data and employs OLS and Prais-Winsten estimators method to analyze this policy. The empirical results illustrate that an increase in import tax may improve the United States’ trade deficit, but it would aggravate the unemployment rate.
Key words: Import tariff, Trade deficit, Unemployment rate
Statement of Originality
This document is written by Yi Wang 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.
Table of Contents
1. Introduction ... 4
2. Literature review ... 6
2.1 Determinations of trade balance ... 6
2.2 Trade and unemployment... 7
2.3 Conclusion of the literature review ... 8
2.4 Hypotheses ... 9
3 Data description ... 10
3.1 Dependent variable and independent variable ... 11
3.2 Control variables ... 11
3.3 Model specification ... 12
4 Results and analysis ... 13
4.1 Variable statistics ... 13
4.2 Pearson correlation coefficient analysis ... 14
4.3 Linear regression analysis ... 16
4.3.1 Trade deficit analysis ... 17
4.3.2 Unemployment analysis ... 18
5 Discussion and conclusion ... 19
1. Introduction
How international trade contributes to a country’s economy is a rather complex matter. Exports are defined as domestic products being sold in a global market, favouring businesses. Imports beget fierce competition and add variety to local markets, which is advantageous to consumers yet harms the domestic industry. Schneider (2004) has asserted that trade offers firms the opportunity to refine the usage of the inputs, enjoy the economic scale and improve productivity, which ultimately leads to a positive outcome. A trade deficit is a significant macroeconomic indicator that evaluates the performance of a country’s economy (Hassan, Wajid & Kalim, 2017). Since the 1970s, the United States has had a trade deficit, and from 1980 onwards the trade deficit has been continuous (Liu & Yang, 2011). Hatsopoulos, Krugman, and Summers (1988) have alleged that the steady increase in the trade deficit has provoked misgivings about the competitiveness of the United States in the long-run. Trade deficits indicate an excessive expenditure, and investment outpaces saving, meaning that foreign money flows into the U.S. market (Cooper, 2008). These offshore funds finance the trade deficit at the cost of a higher interest rate because such rates draw capital inflow. As a result, the deficit increases with the rise in the interest rate (Batra & Beladi, 2018).
President Trump (The New York Times, 2018) has asserted that the trade deficit not only concerns losing money to other countries, it also impedes economic growth via the reduction of jobs. Hence, during his 2016 Presidential Campaign, he promised to reduce the trade deficit. However, more than one year has passed since he was elected. According to the latest census latest data, the trade deficit in goods was $796 billion in 2017 compared to $736 billion in 2016, an increase of 8.06%. Trade with China is the primary driver of this trade deficit, with approximately $375 billion (Thebalance, 2017), almost half of the total trade deficit. To keep his promise and improve the trade deficit, President Trump announced at the end of March (BBC, 2018) that the United States would impose an import tariff of up to 25% on specific Chinese products with a total value of approximately $50 to $60 billion. As a reaction to the import tariffs, the Chinese government announced on April 3 (CNBC, 2018) that 106 United States products might be taxed. Value is around $50 billion. Beijing threatened to introduce a 25% import tax on U.S. imports such as whiskey and cars. Although these announcements have not been put into action, intensive global concerns have risen of a tit-for-tat trade war between the two countries.
According to the principle of the WTO, all members must commit to abrogate protectionism and reduce trade barriers. Imposing high tariffs on each other violates
international law (Loridas, 2011). Moreover, Koopman, Wang, and Wei (2008) have shown that China, as a ‘world factory’, is integrated into a global supply chain. Many international companies outsource manufacturing and assembling to China, which abundantly uses imported inputs and exports globally. The United States is the top supplier, and it is thus detrimental to the U.S. if China mounts the proposed import tax. Additionally, China and the U.S., as the most significant economies and top two importers and exporters globally (World Integrated Trade Solution), together capture more than one-fifth of total world trade, which would affect other countries’ economies. Consequently, the entire global economy is directly and indirectly jeopardized. Furthermore, Cooper (2008) and Feldstein (1987) have shown that, if the United States tries to force a reduction in the trade deficit, it might lead to a financial crisis or recession caused by the devaluation of the dollar, since China holds U.S. government bonds worth billions. When China sells these bonds, their price declines, followed by a fall in the interest rate. Consequently, the value of the dollar decreases and a recession occurs.
Most researchers, based on qualitative analysis, have stated that a trade war between the United States and China is highly undesirable. As the two largest and most significant economies in the world, it is necessary to thoroughly study whether the United States will benefit by executing the intended import taxes on Chinese products. In this paper, based on quantities analysis, I elaborate on the following question: ‘Would this U.S. fiscal policy be effective based on its purpose of reducing the trade deficit and lowering the unemployment rate to benefit the U.S. economy?’ To make this paper more valuable, I consider that there will be a Chinese response; the most realistic case is a Chinese import tax on American goods.
To obtain a reliable answer to this research question, time series data from the years 1992 to 2016, are used in this paper due to the availability of all the variables. The variables were selected according to the literature review to build this study’s dataset. The datasets were selected from the World Bank, except for the deficit numbers between the United States and China, which came from the World Integrated Trade Solution (WITS).
This paper continues with a literature review, which discusses the determinants of the trade deficit and the unemployment rate. Two hypotheses are based on this review. The data and methodology section follows, in which all the variables are described in detail, and the regression models are also introduced. Next, it is the results and analysis part, where regression results are depicted and discussed. Finally, it ends with the discussion of the limitation of this analysis and conclusion.
2. Literature review
This chapter begins by discussing the factors which affect the trade deficit and the unemployment rate. Second, a more thorough explanation, including a figure, is provided to give a better overview of the relationship between the earlier mentioned variables. Finally, two hypotheses are proposed to answer the research question.
2.1 Determinations of trade balance
There are many factors that affect a trade balance in addition to import and export tariffs. The most frequently debated researched variable is the real exchange rate. Hassan, Wajid, and Kalim (2015) have demonstrated that there is a close linkage between the real exchange rate and trade deficit. In their empirical study, they found that, as the exchange rate depreciated, the trade deficit decreased sharply in Pakistan, India, and Bangladesh from 1972 to 2013. Yi (2013) has suggested that the Chinese People’s Bank of China interposes the exchange rate to try to fix their exchange rate to devaluate their currency, making exports more competitive. To some extent, this devaluated RMB attributes to the U.S. – China’s deficit. Feldstein (1987) has studied an example showing that, from 1980 to 1985, the dollar appreciated 70%, leading to the decrease of the U.S’s volume, which consequently caused a significant trade deficit. However, some studies disagree that depreciated currency ameliorates the trade deficit, arguing instead that depreciation deteriorates the short-term economic and short-term trade balance(Kwalingana et al., 2012; Khan, Ali & Ali, 2016). Chinn (2004) has noted that the change in the exchange rate has a limited effect on trade.
The above-discussed exchange rate is related to the make-up of the import price. However, some researchers have argued that economic growth is a crucial factor that influences trade. Rapid economic growth countries, such as India, Bangladesh, and Pakistan, require a high demand for capital import that causes their trade deficit (Hassan, Wajid & Kalim, 2015). On the other hand, Krugman and Baldwin (1987) have disclosed that, although developing countries grow faster than the United States, there is a divergence between the U.S. and foreign Countries. Hence, although foreign countries have a faster growth rate than the U.S., this does not mitigate the trade deficit. Houthakker and Magee (1969) and Chinn (2004) have found that the income elasticity plays a vital role in the trade deficit. Houthakker and Megaee (1969) have unearthed that there is asymmetrical income elasticity between the United States and other countries. First, they found that the U.S. and other developed nations have an equal income elasticity of the total demand for imports, whereas the income elasticity of exports is unequal. The income elasticity of the U.S.’s
exports is much higher than in other countries’. Second, with respect to the demand for importing final manufacture products, the United States’ income elasticity is higher than other countries. Thirdly, the price elasticity related to imports and exports was not significant. Income elasticity is thus the main cause of the U.S.’s trade deficit. Chinn (2004) has also claimed that price elasticity to import has an insignificant effect. Americans are not sensitive to price changes. Due to high-income elasticity, the trade deficit could be reduced by low-income growth or foreign high-low-income growth.
2.2 Trade and unemployment
Many previous studies have researched how trade influences unemployment, and most results have found that the influence is limited (Krause & Mathieson, 1971; Porto, 2008; Davidson & Matuz, 2011; Gaston & Rajaguru, 2013). Gaston and Rajaguru (2013) have asserted that the unemployment rate is more affected by the supply side. Their empirical evidence shows that in Australia’s trade with China, the unemployment rate declined. Porto (2008) has declared that international trade does not arouse the unemployment rate. He believes that unemployed people would be employed in the export industry, and employed individuals would lose their jobs in the importing industry. Hence, the net employment flow does not change. Davis, Haltiwanger, and Schuh (1996) have also supported this argumentation, asserting that trade does not influence job destruction and creation. Krause and Mathieson (1971) have argued that in a dynamic and efficient economy, except imports, there are many other reasons that cause job-destruction. Jobs can be destroyed by government expenditure and new technology as well. Most research advocates trade liberalization rather than trade barriers like an import tax. Francis and Zheng (2011) have shown in their regression result that, since the United States joined the North American Free Trade Agreement (NAFTA), the unemployment rate was reduced by 4.3 percent. Dutt, Mitra and Ranjan (2009) have used panel data to study how trade policy influences unemployment, and they have demonstrated that the openness of the trade has a negative impact on unemployment, but in the long-run, it is advantageous to dwindle the unemployment rate. Matusz and Tarr (2000) reached a similar conclusion that the openness of trade is positively related to employment. To sum up, the relationship between trade tariffs and unemployment is ambiguous.
Another debated factor that influences unemployment is the real exchange rate. Many studies, indeed a majority, agree on the relationship between the real exchange rate and unemployment. Empirical research in Latin America by Frenkel and Ros (2006), has shown that the hypothesis on the correlation between the real exchange rate and employment could
not be rejected. Nevertheless, it is difficult to conclude if the real exchange rate has a positive or negative impact on unemployment, according to other researchers. Ranjbar and Moazen (2009) and Behnamian (2012) have discovered a negative relationship between the two variables. This negative relationship means the value of the domestic currency increases and the unemployment decreases. Frenkel (2004) has suggested that the employment is affected by the real effective exchange rate through three parameters: the macroeconomic part, the developing part, and the labor intensity part. It might lead to a negative coefficient if the different parts work together. Frenkel (2013) also has further tested the correlation and found that, in Hong Kong, Singapore, and the Netherlands, as the real exchange rate increases, unemployment decreases. However, much other research has stated that the appreciation of domestic currency results in local goods losing price competitiveness. As a consequence, production and labor decrease, which leads to higher unemployment (Dutch disease). It can thus be argued that the real exchange rate has a positive relationship with unemployment (Feldmann, 2011; Chimanani et al. ,2012; Zhou, 2010).
However, the inflation rate is closely related to unemployment. The Philips Curve shows that there is a trade-off between inflation rate and unemployment; low unemployment comes at the cost of high inflation, whereas a low inflation rate has to tolerate the high unemployment rate (Treynor, 1975).
2.3 Conclusion of the literature review
As the figure 2-1 shows, this section discusses several factors that affect the trade deficit and unemployment. The trade deficit is not only influenced by imports tax but also by the exchange rate, economic growth and income elasticity. Most researchers agree that the exchange rate has a positive relationship with the trade deficit. Economic growth and income are positively related to the trade deficit.
As to the second goal of the reduction of unemployment, President Trump has averred that imports endanger local business, and destroy the imports industry’s work resulting in more unemployment. Nevertheless, groups of researchers have asserted that the liberation of trade does not damage jobs (Krause & Mathieson, 1971; Porto, 2008; Davidson & Matuz, 2011; Gaston & Rajaguru, 2013). They have shown some empirical evidence to attest that free trade decreases unemployment, and the openness of trade is thus negatively related to unemployment. Import tax is positively associated with unemployment. The real effective exchange rate is the second discussed factor; it is somewhat debatable, and it can both negatively and positively impinge on unemployment, depending on government intervention. The last element is inflation. The Philips Curve exposes that it has a negative relationship with the trade deficit, which indicates that higher inflation level is always concomitant with the lower unemployment rate.
2.4 Hypotheses
The primary research question in this paper is: Whether the U.S. fiscal policy, imposing the import tariff on Chinese imports, would be effective based on its purpose of reducing the trade deficit and lowering the unemployment rate to benefit the U.S. economy, when meanwhile Chinese government imposes a corresponding import tax on the U.S.. This research question can be divided in two sub-questions. The first concerns how import tax
Figure 2-0-1 Source: Summary of the Literature Review
Increase Imports Tax on American Goods Problem Policy Goals Determinants Increase Imports Tax on Chinese Goods Reduction of the Trade Deficit with China
Reduce Unemployment Exchange Rate (US and RMB) Economy Growth(Income) (Income Real Effective
Exchange Rate Inflation Reaction
influences the trade deficit and how it affects the unemployment. This article advances the following two hypotheses:
Hypothesis 1: There is no significant correlation between an import tariff on Chinese goods and the U.S.-China trade deficit; imposing a tariff will not improve the trade deficit.
According to the literature review, it is debunked that the exchange rate and income elasticity would impact the trade deficit. Hence, in this hypothesis, the real exchange rate between the United States and China, and the GDP per capita of the United States are introduced and treated as control variables. These two variables are expected to have a direct or indirect influence on the dependent variable imports tax and dependent variable trade deficit. At the end of the analysis, I assume the exchange rate and the GDP per capita exchange rate have the dominant effect on the trade deficit, and imposing duties will not reduce the trade deficit.
Hypothesis 2: There is a positive relationship between import tariffs and the unemployment rate of the United States; increasing import tariffs will increase the unemployment rate.
In the literature review, it has disclosed that the real effective exchange rate has either a positive or negative effect on the unemployment rate, and inflation negatively affects unemployment. Hence, these two variables are adopted and considered as the control variables that influence the dependent variable, the import tax, and the independent variable, unemployment rate, in this hypothesis. Furthermore, the literature review has discovered that import tax as a trade barrier curbs free trade. As a result, it arouses the unemployment. From another perspective, the import tax increases the import price that protects the imports industry. This protection contributes to preventing from jobs being destroyed. Hence, I expect the increase of the import tax will cause a rise in the unemployment rate.
3 Data description
In this section, it describes the obtained data that are used in the regression analysis. All the data are secondary data from official websites. The U.S.- China trade deficit is collected from the World Integrated Trade Solution (WITS) which is a collaboration of the World Bank collaborates with other organizations, including UNCTAD, IDB, UN COMTRADE, and TRAINS. Other variables are from the World Development Indicators in the World Bank database. According to the literature review, the data from the United States this article needs are the trade deficit with China, GDP per capita, the real effective exchange rate, CPI index, and the United States’ unemployment rate and inflation. From the Chinese side, this paper
requires official exchange rate and CPI index. Due to the availability of bilateral trade data in WITS, the period extends from 1992 to 2016 when the latest data was disclosed on the website. In the following subsections, it explains each variable in detail in light of the World Bank’s definition, and at the end, specifies the regression model and provide a summary of the variables.
3.1 Dependent variable and independent variable
In this paper, concerning the hypotheses, the dependent variables are the United States’ trade deficit with China and its unemployment rate. The value of the trade deficit is calculated by imports minus exports, which is a positive number since the U.S. imports more than it exports with China. The data concerning imports and exports are from WITS that indicates the trade of all products, except the trade in service. Data are measured in a million US dollar. In the regression, it is LGTRDF, which is the logarithmic scale of the trade deficit.
UNEM indicates the dependent variable unemployment rate that is the unemployment divided by the total labor force. Unemployment here refers to the people who are able to work and are searching for a job but currently is jobless. Individuals who receive offers but have not yet started are also counted in unemployment.
The independent variables are the United States import tariff, and China import tariff in the regression is shortened to ITU and ITC. This import tax is the weighted average of all effective tax rates by category times their import product share. If the effective rate applied is not found, the import tax rate could be replaced by its most favored nation rate. The data is projected by World Bank employees who adopt the World Integrated Trade Solution system. 3.2 Control variables
In this paper, there are four control variables. GDP per capita and the official Chinese exchange rate are related to the trade deficit, while inflation rate and the real effective exchange rate pertain to the unemployment rate. Detailed information on the variables is provided as below:
GDP per capita: GDP, the gross domestic product, is computed by the World Bank staff such that the gross value is added by the local manufacturer in the country minus the subsidiary that does not include in the production, followed by adding up product tax. Afterward, the GDP divided by the population mid-year is the GDP per capita. In the regression, the GDP per capita of the U.S. and China are in logarithmic form, and the abbreviations are LGGDPPUS and LGGDPCH.
Official exchange rate: the description of the official exchange rate in the World Bank refers to the exchange rate determined by the legally sanctioned exchange market or national authorities, and it is annualized by the average each month. The official exchange rate of the United States is 1 that acts as the benchmark. Hence, the Chinese official exchange rate is the exchange rate between the RMB and the USD. The definition of the exchange rate in the World Bank is how much one pays in a local currency in exchange for one USD, the price of the dollar. The Chinese official exchange rate is denoted by the abbreviation OER. An increased OER indicates that the U.S. dollar appreciates. This indicator is the nominal exchange rate.
Real effective exchange rate: By definition in the World Bank, this is the nominal effective exchange rate divided by a cost index or a price deflator. The nominal effective exchange rate is a ratio expressed in a base (taking 2010 as 100) compared with a weighted average of the primary foreign currency and the euro area. How much the weighted basket currency you can buy with one domestic currency. In the market, the choice made by individuals, manufacturers, households and the government is affected by the relative price. For this reason, this paper uses the real effective exchange rate rather than the nominal effective exchange rate. If the real effective exchange rate goes up, it means the U.S. dollar has depreciated. This relationship is opposite to that of the OER.
Inflation rate: the inflation rate measures the increase in the price of goods and services over the time. The World Bank adopts Laspeyres formula to compute the inflation, which is the price of the basket of products and services in the year divided by the price of the same basket of goods and services in the previous year. However, the basket of goods and services can change at a particular interval, and has a yearly basis. The inflation is shortened as INF in the regression.
Consumer Price Index (CPI): This index reflects the change of the cost by consumption of a basket of goods or service. This index is similar to inflation, and the two indications are firmly connected. Generally speaking, an increase of the price level of an economy is called inflation. It is mentioned separately here from inflation as later it is used to compute the real exchange rate between the United States and China.
3.3 Model specification
To test the relationships between import tax and trade deficits, as well as the import tax with unemployment, the following two models were built:
𝑈𝑁𝐸𝑀𝑡 = 𝛾0+ 𝛾1𝐼𝑇𝑈𝑡+ 𝛾2𝐼𝑇𝐶𝑡𝜋𝑟2+ 𝛾3𝑅𝐸𝐸𝑅𝑡+ 𝛾4𝐼𝑁𝐹𝑡+𝜖𝑡
Where ‘t’ refers to the years from 1992 to 2016. LGTRDF is the United States-China deficit in logarithm form. ITU and the ITC are the import taxes of America and China respectively. The RERUC is the real exchange rate of the USD and RMB. The formula is 𝑅𝑒𝑎𝑙 𝐸𝑥𝑐ℎ𝑎𝑛𝑔𝑒 𝑅𝑎𝑡𝑒𝑒𝑅𝑀𝐵/𝑈𝑆𝐷 = 𝑁𝑜𝑚𝑖𝑛𝑎𝑙 𝐸𝑥𝑐ℎ𝑎𝑛𝑔𝑒 𝑅𝑎𝑡𝑒𝑅𝑀𝐵
𝑈𝑆𝐷∗∗ 𝐶𝑃𝐼𝑈𝑆
𝐶𝑃𝐼𝐶ℎ𝑖𝑛𝑎 , where the
nominal exchange rate is the previously mentioned OER. CPI is the consumer price index. LGGDPP is the logarithm of the GDP per capita containing income information. REER is the real effective exchange rate, while INF is the inflation rate. α and δ are constants, and ε and ϵ are the standard errors. Table 3-1 below gives a clear overview of the variables and their descriptions.
Table 3-1 Variable Name and Descriptions Name Description
LGTRDF The United States’ trade deficit with China in logarithm UNEM The unemployment rate of the United States
ITU The United States’ import tariff on Chinese imports ITC Chinese import tariff on the U.S imports
RERUC Real exchange rate between RMB and USD LGGDPPUS The United States’ GDP per capital in logarithm LGGDPPCH The China’s GDP per capital in logarithm
REER The real effective exchange rate of the United States with respect to the basket currency INF The rate of inflation in the United States
4 Results and analysis
In this section, first it summarizes each variable and presents the correlation among them. Afterward, it displays and describes the regression results with regard to the trade deficit and unemployment rates of the United States.
4.1 Variable statistics
Table 4-1 presents the summary of each variable’s statistics. It displays data from 1992 to 2016, yet there is one variable missing in American import tax (ITU, the year 1994), and three variables are missing in the Chinese import tax (ITC, the years 1995, 2012 and 2013). As a result, only 21 years of data are regressed later. In addition, the average import tax of China is 9.3%, almost three times higher than the United States’ import tax at 3.3%. Concerning GDP per capita in log form, the lowest GDP per capita is approximately 40% higher than the highest GDP per capita of China. Despite the ITC, the means of other
variables are located approximately in the middle of the average of the minimum and maximum. The only negative number in the dataset is the inflation in 2009, -0.3556%, which is due to the reaction to the financial crisis in 2008.
Table 4-1 Variable statistics
Variables Mean Min Max Std.dev N
LGTRDF 2.112 1.301 2.589 0.413 25 UNEM 0.060 0.040 0.096 0.016 25 ITU 0.033 0.020 0.040 0.006 24 ITC 0.093 0.019 0.250 0.077 22 LGGDPPUS 4.535 4.328 4.697 0.113 25 LGGDPPCH 3.185 2.509 3.802 0.422 25 RERUC 7.454 5.889 9.222 0.964 25 REER 107.425 95.131 124.242 8.368 25 INF 0.023 -0.004 0.038 0.010 25
4.2 Pearson correlation coefficient analysis
The ordinary least squares (OLS) is one of the linear regression methods used in this paper. It minimizes the total of the square of the difference between dependent variables in the dataset and projected in the linear function (Stock & Watson, 2011). One of the essential assumptions about the best linear unbiased estimator (BLUE) under the Gauss-Markov Theorem is that there is no perfect correlation between dependent and control variables in a multi-linear regression, and the estimator will be more precise with fewer correlations (Stock & Watson, 2011). This section focuses on the Pearson correlation coefficient to check correlations between every two variables and ignoring the error term. The correlation with the error term will be discussed in the next subsection. The results are shown in the table below.
Table 4-2 Trade deficit correlation
Correlation Matrix for the U.S. Trade Deficit
LGTRDF ITU ITC RERUC LGGDPPUS LGGDPCH
LGTRDF 1 ITU 0.4117 1 ITC -0.8077*** -0.7789*** 1 RERUC -0.5592** -0.4167* 0.3402 1 LGGDPPUS 0.9904*** 0.4216** -0.7762*** 0.6240*** 1 LGGDPPCH 0.9519*** 0.4909 -0.7674*** 0.7604*** 0.9722*** 1 * =significance at 0.05 level, ** p=significance at 0.01 level, *** = significance level at 0.001 level
Table 4-2 and figure 4-1 illustrate the correlation between variables concerning the trade deficit. The trade deficit is negatively related with Chinese import tax 0.8077,p<0.001) and the real exchange rate between the dollar and the RMB
(ρ=-0.5592,p<0.01), and positively related with two countries’ GDP per capita (ρUS=0.9904 and ρCH=0.9519,p<0.0001). There is no strong evidence to prove the relationship with the United States import tax, which has a negative relationship with the Chinese import tax (ρ=-0.7789,p<0.001) and the real effective exchange rate between dollar and the RMB. ITU positively related with the GDP per capita of the United States, but there was not enough evidence to show that it has a relationship with China’s GDP per capita. The Chinese import tax influences both countries’ GDP per capita negatively, with ρ=-0.7762 and - 0.7674, p<0.001) respectively. The real effective exchange rate positively related with the two countries’ economies, which are positively related with each other (ρ=0.9722,p<0.001).
For the unemployment rate, the correlations are shown in table 4-3 and figure 4-2. In this correlation matrix, except for the negative correlation between the two import taxes discussed previously, only American import tax and real effective exchange rate affect unemployment positively (ρ=0.569,p<0.01) and negatively ρ=-0.6147, p<0.01 respectively. For the relationships to all other variables, there are not enough data to demonstrate the ρ are significantly different from 0.
Table 4-3 Unemployment rate correlations
Correlation Matrix for the Unemployment
UNEM ITU ITC REER INF
UNEM 1
ITU 0.5690** 1
ITC -0.4056 -0.7789*** 1
REER -0.6147** -0.4161 0.2245 1
INF -0.2932 -0.1610 0.2570 -0.1679 1
* =significance at 0.05 level, ** p=significance at 0.01 level, *** = significance level at 0.001 level
4.3 Linear regression analysis
Table 4-4 Regression Results
In this paper, the dataset is time series data, and thus the autocorrelation of the error term has to be considered. Autocorrelation, namely serial correlation, is where the residuals in the time series are correlated with each other. Take the first-order autocorrelation as an example, it can be expressed as 𝐞𝐭 = 𝛉𝐞𝐭−𝟏+ 𝛕𝐭 (Stock & Watson, 2011). This correlation calculates the relationship with its current value and previous value (Stock & Watson, 2011).
The Prais-Winsten estimation is a particular case of the feasible generalized least squares (FGLS) to address the autocorrelation. Generalized least squares (GLS) and ordinary least squares (OLS) are two basic techniques used in a linear regression for estimating the unknown parameters. The difference between them is that GLS consider a certain degree of the correlation between the error terms, while OLS does not take this into account(Stock &
OLS Prais-Winsten OLS Prais-Winsten
Dependent Variable LGTRDF (1) LGTRDF (2) UNEM (3) UNEM (4)
Independent Variable ITU -5.396 -3.530 1.469* 1.299* (3.411) (3.479) (0.728) (0.731) ITC -0.718* -0.582* 0.0440 0.0299 (0.370) (0.278) (0.0519) (0.0495) Control Variable RERUC 0.0885** -0.00528 (0.0410) (0.0354) LGGDPPUS 2.008** 3.896*** (0.891) (0.800) LGGDPPCH 0.477 -0.149 (0.302) (0.259) REER -0.00118*** -0.00110*** (0.000325) (0.000354) INF -0.564** -0.594*** (0.248) (0.176) Constant -8.927*** -14.88*** 0.149*** 0.148*** (2.860) (2.687) (0.0479) (0.0483) Observations 21 21 21 21 R-squared 0.990 0.988 0.639 0.765 F-test 2.73** 1.29 3.89* 3.05* p-value 0.01193 0.2737 0.0662 0.0998 D-W test 1.096787 1.129801 0.478859 0.943388
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Watson, 2011). Therefore, the Prais-Winsten estimator is relatively more efficient than the OLS estimator when there is a series correlation. Both regression methods are used in this paper and the results are shown in table 4-4.
4.3.1 Trade deficit analysis
In both regressions (1) and (2), from R-square, more than 98% of the variance is explained by this model. Also, both regressions show that there is not enough evidence to support that increasing the import tax on Chinese goods will reduce the bilateral trade deficit. However, both regression results provide some evidence that when China imposes an the import tax on American products, it contributes to abating the U.S. trade deficit, which seems unreasonable. But it can be demonstrated as follows:
The regression model is:
𝐿𝐺𝑇𝑅𝐷𝐹𝑡 = 𝛽0+ 𝛽1𝐼𝑇𝑈𝑡+ 𝛽2𝐼𝑇𝐶𝑡+ 𝛽3𝑅𝐸𝑅𝑈𝐶𝑡+ 𝛽4𝐿𝐺𝐺𝐷𝑃𝑃𝑈𝑆𝑡+ 𝛽5𝐿𝐺𝐺𝐷𝑃𝑃𝐶𝐻𝑡+ 𝜀𝑡 (1) Due to the correlation and autocorrelation issues, equation (1) is combined with the below three equations (2), (3),(4)
𝐷𝑢𝑠 = 𝑎
0+ 𝑎1𝐼𝑇𝑈 + 𝑎2𝐼𝑇𝐶 + 𝑎3𝑅𝐸𝑅𝑈𝐶 + 𝑎4𝐿𝐺𝐺𝐷𝑃𝑃𝑈𝑆 + 𝑎5𝐿𝐺𝐺𝐷𝑃𝐶𝐻 + 𝜇 (2)
Here, D refers to the trade deficit. And the signs of the coefficient are:
𝑎1 < 0, the United States’ import tax increase, imports decrease, 𝐷𝑢𝑠 decreases.
𝑎2 > 0, China increase import tax, the U.S. exports decrease, 𝐷𝑢𝑠 increases.
𝑎3 > 0, the U.S. dollar appreciates that is to the disadvantage to the trade, 𝐷𝑢𝑠 increases.
𝑎4 > 0, when the U.S.’s GDP per capita increase, people consumes more, and hence import more, 𝐷𝑢𝑠 increases.
𝑎5 < 0, opposite sign with 𝑎4, when China becomes wealthier, they consume more imports from the United States, so 𝐷𝑢𝑠 decreases.
With respect to import tax, it can be assume that the import tax is positively related to the trade deficit and the other country’s import tax. This assumption can be supported by the background of this paper. Due to the high trade deficit of the United States, President Trump wants to increase the import tax to reduce trade deficit. This action implies that the import tax is affected by the level of the trade deficit. Through the Chinese reaction to the increase of the U.S. import tax on Chinese goods, China will increase the import tax on American goods. This reaction hints that the import tax is positively related to other countries’ import tax. So, I create the following equations (3) and (4):
𝐼𝑇𝑈 = 𝑏0+ 𝑏1𝐷𝑈𝑆+ 𝑏
𝐼𝑇𝐶 = 𝑐0+ 𝑐1𝐷𝐶𝐻+ 𝑐2𝐼𝑇𝑈 + 𝜔 (4)
Where 𝑏1, 𝑐1 > 0, and 𝑏2, 𝑐2 > 0.
In the bilateral trade, 𝐷𝐶𝐻 = −𝐷𝑈𝑆, and together with (2),(3),and (4), I arrive at equation (5):
𝐷𝑢𝑠 = 𝑎
0+ 𝑎1(𝑏0+ 𝑏1𝐷𝑈𝑆+ 𝑏2𝐼𝑇𝐶 + 𝜕 ) + 𝑎2(𝑐0+ 𝑐1𝐷𝐶𝐻+ 𝑐2𝐼𝑇𝐶 + 𝜔) + 𝑎3𝑅𝐸𝑅𝑈𝐶 +
𝑎4𝐿𝐺𝐺𝐷𝑃𝑃𝑈𝑆 + 𝑎5𝐿𝐺𝐺𝐷𝑃𝐶𝐻 + 𝜇 (5)
Simplified (5) results in equation (6): 𝐷𝑢𝑠 =(𝑎0+𝑎1𝑏0+𝑎2𝑐0) 1−𝑎1𝑏1+𝑎2𝑐1 + 𝑎2𝑐2 1−𝑎1𝑏1+𝑎2𝑐1𝐼𝑇𝑈 + 𝑎1𝑏2 1−𝑎1𝑏1+𝑎2𝑐1𝐼𝑇𝐶 + 𝑎3 1−𝑎1𝑏1+𝑎2𝑐1𝑅𝐸𝑅𝑈𝐶 + 𝑎4 1−𝑎1𝑏1+𝑎2𝑐1𝐿𝐺𝐺𝐷𝑃𝑃𝑈𝑆 + 𝑎5 1−𝑎1𝑏1+𝑎2𝑐1𝐿𝐺𝐺𝐷𝑃𝐶𝐻 + (𝜇+𝑎1 𝜕+𝑎2𝜔) 1−𝑎1𝑏1+𝑎2𝑐1 (6)
Comparing (1) and (6), the difference is that one has log form, and (6) is without the log, but signs of the coefficients are not affected according to the mathematics, this infers that 𝛽2, the coefficient of ITU has same sign with 1−𝑎𝑎1𝑏2
1𝑏1+𝑎2𝑐1 which is negative. This
demonstration provides a reason why China’s import tax increase might lead to a decrease in the United States trade deficit.
Despite China’s import tax, regression (1) gives the significant evidence that when the real exchange rate between the USD and RMB increases by 1%, the trade deficit increases by 0.0885% (p<0.05). Although this is not supported by regression (2), it makes sense that as RERUC increase, the dollar becomes expensive, exports decrease, and the trade deficit increases, matching the β3 > 0. Both (1) and (2) display that if the U.S. GDP per capita increases by 1%, the trade deficit increases by 2.008% (p<0.01) and 3.896% (p<0.01), respectively. However, there is not enough evidence to prove that the Chinese GDP per capita affects the trade deficit.
The F test analyses the effect when both countries increase import taxes that tests if 𝛽𝐼𝑇𝑈+ 𝛽𝐼𝑇𝐶 = 0. From the results, it manifests that in (1) under a significant level of 5%, there
is some evidence to conclude that when both countries increase import duties, the trade deficit of United Stated decreases. However, with a higher significant level or regression (2), this conclusion is not supported.
4.3.2 Unemployment analysis
In regression 3 and 4, 63.9% and 76.5% of variances are explained by the model. Under a significance level of 10%, both results display that unemployment rate is positively related with the United States’ import tax with beta 1.469 and 1.299 respectively. This means that
when the United States increases a 1% import tariff, the unemployment rate increases more than 1.29%, which runs counter to the purpose of increasing the duty. This can be explained by the liberalization trade in the literature review. The real effective exchange rate negatively influences the unemployment with beta -0.0118 (p<0.01) and beta -0.0011(p<0.01). Apart from that, both (3) and (4) support the claim that inflation helps reduce unemployment, with beta -0.564 (p<0.05) and beta -0.594 (p<0.01) respectively. The value of the D-W test improves from 0.48 in (3) to 0.94 in (4), almost doubled, but the signs of the betas are same, and the value and the significance levels are similar. This strengthens that the regression results are relatively reliable.
In the F test, it also tests the co-effect when both countries impose higher tariffs. It discovers that the unemployment rate increases if the policy is conducted with F value 3.89 (p<0.1) and 3.05 (p<0.1) respectively. The evidence is weak under the higher significance level. There is no evidence to show that the unemployment rate of the United States is improved under this scenario.
5 Discussion and conclusion
In this paper, it investigates the research question: whether the U.S. fiscal policy, imposing import tariff on Chinese imports, would be effective based on its purpose of reducing the trade deficit and lowering the unemployment rate to benefit the U.S economy, when meanwhile the Chinese government imposes the corresponding import tax on the U.S.. To answer this question, in the literature review section, control variables were identified that affect the dependent variables. The nations’ income levels and the real exchange rate between the dollar and the RMB impact the trade deficit, and the inflation and real effective exchange rate influence the unemployment rate. The data used for the regression was provided by the World Bank and WITS. Through the Pearson correlation analysis, it was discovered that there exist many correlations between independent variables and control variables. Additionally, due to the inherent problem of the time series data, this paper employed both OLS regression and the Prais-Winsten estimator to answer the research question.
In the results section, it was uncovered that an increase in the import tax on Chinse imports does not reduce the trade deficit, but when China increases the import tax on American products, this does contribute to reducing the trade deficit. The explanation of this phenomenon is that the import tax is a reaction of the trade imbalance and the counterparty’s import tax strategy. If the United States and China both increase the import tariff, the trade deficit would decrease. However, this conclusion is only supported in the OLS estimator with
a 10% significance level. Other higher significance levels and the Prais-Winsten estimation conclude that raising the import tax will not affect the trade deficit. Concerning the unemployment rate, an increase of the United States’ import tax leads to a higher unemployment, whereas the import tax of China has no effect on the unemployment rate. If both countries increase the import tax reciprocally, the unemployment rate increases with the significance level of 10%.
Therefore, it is not advisable to impose the reciprocal import tax. The goal of reducing the trade deficit might be achieved with weak evidence. However, the purpose of the reduction of the unemployment rate would not be reached. On the contrary, the unemployment would be exacerbated. Hence, it is not wise for President Trump to impose an import tax on China. Instead, liberalization trade should be advocated to reduce the unemployment.
There are several limitations in this paper. First, the time series dataset is limited (1992 to 2016) due to the availability of the import tax of China. Furthermore, within that 25 year period, there are still some data are missing. Ultimately, the number of years with available data is only 23. This dataset is therefore not large enough to draw convincing conclusions. Second, as no specific model was advanced by the previous study fitting this research, the model was built by the author via the combination of literature reviews. Third, because the dataset is imitated, it could only contain few degrees of freedom. Otherwise, an overfitting problem would result, which is a modeling error. To avoid this modeling error the number of control variables was constrained. If not, too many control variables would lead to the overfitting problem, making the results biased and unable to thoroughly explain the relationship between the import taxes and the dependent variables. Fourth, the import taxes used in the analysis are derived from the weighted averaged mean of the all effective tax rates. This tax rate is generally a countries’ tax rate, but not the specific tax rate between the United States and China. In an optimal case, the import taxes would be calculated by the weighted average tax rate of the United States exporting to China, and China exporting to the U.S., by categories multiplied by their import product share. However, this information is not accessible.
Reference
Amadeo, K.(2018, March 20). U.S. Trade Deficit With China and Why It's So High. The Balance. . Retrieved from https://www.thebalance.com/u-s-china-trade-deficit-causes-effects-and- solutions-3306277.
Anderson, J., & van Wincoop, E. (2004). Trade costs. Journal of Economic Literature, 42(3), 691–751.
BBC. (2018, March 22). Trump Announces Tariffs on $60 bn in Chinese imports. Retrieved from http://www.bbc.com/news/business-43494001.
Bloomberg. (2018, May30). China Cuts Tariffs on Wide Range of Consumer Goods From July. Retrieved from https://www.bloomberg.com/news/articles/2018-05-30/china-cuts-tariffs-on-wide-range-of-consumer-goods-from-july-1.
Behnamian, M. (2012). Effects of real exchange rate on unemployment in Iran. Economic Journal:
Bimonthly of Investigation into Economic Policies and Affairs, 11(12), 23–40.
Cooper, R. N. (2008). Global Imbalances: Globalization, Demography, and Sustainability
Journal of Economic Perspectives, 22 (3), 93-112.
Chimanani, H., Bhutto, N. A., Butt, F., Sheikh, S. A., & Devi, W. (2012). The effect of exchange rate on unemployment rate in Asian countries. Proceedings of 2nd International Conference on Business Management.
Chinn, M.D.( 2004). Incomes, exchange rates and the U.S. trade deficit, once again. International
Finance . 7(3), 451-469.
Davidson, C., Martin, L., & Matusz, S. (1999). Trade and Search Generated Unemployment. Journal
of International Economics. 48(2), 271-299.
Davis, S.J., Haltiwanger,J., & Schuh, S. (1996). Small business and job creation: Dissecting the myth and reassessing the facts. Small Business Economics, 8(4), 297-315.
Dutt, P., Mitra, D., & Ranjan, P. (2009). International trade and unemployment: Theory and cross- national evidence. Journal of International Economics, 78(1), 32–44.
Feldmann, H. (2011). The unemployment effect of exchange rate volatility in industrial countries.
Economic letters, 111(3), 268–271.
Feldstein,M. (1987). Correcting the Trade Deficit. Foreign Affairs, 65, 791-806.
Francis, J., & Zhen, Y.Q. (2011). Trade liberalization, unemployment, and adjustment: Evidence from NAFTA using state level data. Applied Economics, 43(13), 1657-1671.
Frenkel , R. (2004). Real exchange rate and employment in Argentina, Brazil, Chile and Mexico.
Frenkel, R., & Ros, J. (2006). Unemployment and the real exchange rate in Latin America. World
Development, 34(4), 631–646.
Hassan, S.H., Wajid, A., & Kalim, R. (2017). Factors affecting trade deficit in Pakistan, India and Bangladesh. Econ Polit, 34, 283-304.
He, X. (2013). Real effective exchange rate and unemployment rate: The difference between re- exporting and non-re-exporting countries (Unpublished master's thesis). Clemson University, United States.
Houthakker, H.S., & Magee, S.P. (1969). Income and Price Elasticities in World Trade. The Review of
Economics and Statistics, 51(2), 111-125.
Khan, Z., Ali, A., & Ali, S. (2016). Impact of Devaluation on Balance of Trade: A Case Study of Pakistan Economy. Asian Journal of Economic Modelling, 4(2), 90-94.
Krugman, P., & Baldwin, P. (1987). The Persistence of the U.S. Trade Deficit. Brookings Papers on
Economic Activity, 1, 1–44.
Kwalingana, S., K. Simwaka., Munthali. T., & Chiumia, A. (2012). The short-run and long-run trade balance response to exchange rate changes in Malawi. Journal of Development and
Agricultural Economics, 4(8): 221-232.
Gaston, N.,& Rajaguru, G. (2013). How an export boom affects unemployment. Economic Modelling. 30, 343–355.
Guo, M., Lu, L., Sheng, L. G., & Yu, M. J. (2018). The Day After Tomorrow: Evaluating the Burden of Trump's Trade War. Asian Economic Papers, 17 (1), 101-120.
Guo, R, X. (2004). How Culture Influences Foreign Trade: Evidence from the U.S. and China. Journal
of Socio-Economics, 33, 785–812.
Hummels, D. (1999). Toward a Geography of Trade Costs. Krannert School of Management. Purdue University.
Koopman, R., Wang, Z.,& Wei, S.J. (2008) How much of Chinese exports is really made in China? Assessing Domestic Value-Added When Processing Trade is Pervasive. NBER Working Paper, 14109.
Krause, L., & Mathieson, J.A. (1971). How much of Current Unemployment did We Import.
Brookings Papers on Economic Activity, 2, 417-428.
Loridas, K. (2011). United States-China Trade War: Signs of Protectionism in a Globalized Economy.
34 Suffolk Transnat'l L, (304).
Matusz, S. J. & Tarr, D. (2000). Adjusting to trade policy reform, in Economic Policy Reform: The Second Stage, The University of Chicago Press, Chicago and London.
and Chemicals. CNBC. Retrieved from https://www.cnbc.com/2018/04/04/china-new-us- tariffs-including-soy-cars-and-chemicals.html.
Porto, G.G., 2008. Agro-manufactured export prices, wages and unemployment. American Journal of
Agricultural Economics, 90(3), 748–764.
Ranjbar, H., & Moazen, S. (2009). An investigation into the relationship between exchange rate and other factors affecting unemployment in countries of the Organization of Islamic
Cooperation (OIC). The Economic Journal, 3(6), 43–63.
Stock, J.H., & Watson, M.W. (2011). Introduction to economics (3th ed.). Boston, MA: Pearson.
Treynor, J.L. (1975). Unemployment and Inflation. Financial Analysts Journal, 5, 21-28.
Yao, K., & Coghill, K. (2018, April 9). China Unwilling to Fight Trade War, but Not Afraid of One-Vice Commerce Minister. The New York Times. Retrieved from
https://www.nytimes.com/reuters/2018/04/09/business/09reuters-usa-trade-china-tradewar.html.
Yi, G. (2013). RMB Exchange Rate and Trade Balance. Business Hortzons, 56, 421-432.
Zhou, Y. (2010). Transmitted unemployment and exchange rate effect on labor market (Unpublished master's thesis). Lingnan University, Hong Kong.
