The Effect of Trade Barriers on Sectorial Wages in Ecuador Between 2007 and 2019: Is Protectionism a Curse Word?

(1)

The Effect of Trade Barriers on Sectorial Wages in Ecuador Between

2007 and 2019:

Is Protectionism a Curse Word?

Achic Pacari Lema Muenala -114524409

Faculty of Economics and Business, University of Amsterdam BSc Economics and Business Economics: Major in Economics

MSc C.W. Haasnoot 2019-2020

(2)

Statement of Originality

This document is written by the student Achic Pacari Lema Muenala, who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are 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 the completion of the work, not for the contents.

(3)

Abstract

This thesis investigates the response of sectoral wages to the introduction of international trade policy in Ecuador between 2007 and 2019. Governments tend to protect some industries making a worker industry association a key channel in the way trade policy may affect the labor market.

The author exploits a period of intense protectionism(2009-2012) in Ecuador and analyses trade policy effects on an extensive and intensive margin. The evidence suggests that wages and wage growths in protected industries did not show any significant differences relative to unprotected sectors at the extensive scope. At the intensive margin and controlling for labor market features, however, industries targeted by protectionism showed lower-wage growths relative to unprotected sectors. It concludes that the volume of trade barriers play an important role in the determination of trade barrier effects.

(4)

Acknowledgments

I render thanks to the Secretaría de Educación Superior, Ciencia, Tecnología e Innovación (SENESCYT), the sponsor of my scholarship, that has enabled me to access an undergraduate education of quality.

To my family and the indigenous peoples, they have been my motivation for learning. Eventually, I will be able to contribute to finding solutions for the corollaries that have tormented their welfare for centuries.

My sincere gratitude to my thesis supervisor, P.h.D. Kees Haasnoot, for his exceptional guidance and encouragement through this thesis and my bachelor’s career.

(5)

In 2008, the newly elected president at the time, Rafael Correa, established drastic changes in the production model of the country. It attempted to overcome the stated of oil-dependent of the economy, flagging remittances from workers abroad, and the drying-up of foreign investment. The constitution of the Republic of Ecuador was adjusted, pursuing the reconstitution of the state and the government as its regulatory role. Among the new features in which the government had a major executive role were the regulation of markets and the economy (Government of the Republic of Ecuador, 2013). Ecuador turned to protectionism and, by 2015, the country was positioned as the fifth economy with more protectionist measures (Global Trade Alert, 2015).

In 2009, the discussion about free trade and protectionism intensified after the announcement of the National Plan for the Good Living in the Ecuadorian context. On the one hand, the private sector and right-party politicians have been argued that releasing trade barriers will improve the quality of life of all countries, and thus, it was a better option for the country. On the other, the president at the time and his cabinet affirmed that protectionist policies are essential to advance towards economic development, especially for developing countries that have recurred to protectionist policies to shield their workers’ wages against the called unfair competition when free trade. Less liberal trade policies are aimed to promote the development of the national industry, favoring local production against foreign competition, secure domestic jobs, and wages.

(7)

In this regard, this work seeks to highlight the effect of the protectionist measures taken in Ecuador during 2008 and 2012 on its sectorial wages. In order to answer whether workers’ wages suffer from positive deviation as a result of protectionist policies.

On the contrary, economists and international organizations advocate countries be open to trade liberalization. The International Monetary Fund (2001), claims trade liberalization as a powerful instrument in economic growth, development, and reduction of poverty for all countries. This consensus is based on the idea that less developed countries would experience large income inflows if the removal of trade barriers are implemented. The liberalization of agriculture would favor low- and middle-income countries provided the relative importance that primary sectors have within their economies. The structure of Ecuadorian foreign trade has not changed since initial times. It has been characterized by the prevalence of primary products such as oil, bananas, flowers, cocoa, tuna, among which represent the greatest weight of exports. Ecuador’s major exports are crude petroleum and bananas, about 36% and 13% of the total exports, respectively (OEC, 2018). Surprisingly, refined petroleum comprehends the top import good (ibid).

A historical review reminds us that the integration of a unique economy was unequal during the last decades. Some Asian economies experienced incredible advances, with less intensity seen in Latin America. Yet, the distribution of those gains is put into question in developing countries. It is argued that trade is not always beneficial to all parties involved, like labor, capital, and landowners (Krugman et al., 2018). The classical justifications of free trade do not agree with the reality of the political, social, and economic structure of countries (Edwards, 1988).

The investigation will develop a Difference-in-Difference research method to encounter with the inquiry. Where the nominal wage per sector and its growth are our variables of interest

(8)

and our treatment are trade policies like tariffs and quotas implemented by the Ecuadorian government from 2009 to 2012.

Data and information are gathered from Ecuadorian government official websites, international organizations, and empirical studies. Accordingly, the research has the following structure: main text, conclusions, discussion and limitations, and recommendations.

The main text comprehends four main sections. The first is the literature review, where the main concepts and definitions, theory, and previous research are explained. In the methodology section, the actual research is conducted. Previous to the conclusions and limitation section, results from the previous section are presented.

(9)

2. Literature Review

This section contains general theory and previous empirical literature on international trade and wages. First, the theory on trade policy, its types, and trends over time will be presented. Next, the effect of protectionist policies on welfare is discussed. Following this subsection, it is presented theory and empirical analysis on the effect of trade barriers on wages. Furthermore, brief literature on some other sources of wage determination is considered. Finally, based on the literature provided, conclusions and hypothesis are derived.

2.1.International Trade Policy

The dynamic process of globalization has caused an incessant flow of goods and services, despite the curtails in periods of crisis. Nations engage measures over their economy that influence their trade and growth (see figure 1).

(10)

Next to the boosted cross-border trade, there are trade policies that regulate global trade. Its definition varies across institutions and economists. Some define trade policies as the framework of global trade, redistribution perspective, or the enabler of trade.

According to the Dictionary of Trade Policy Terms (Goode, 2003), trade policy embodies a set of instruments, norms, regulations implemented by governments to permit legal market access for firms to trade. It seeks to draw the playing field, securing predictability and fairness. Trade policies are domestic policies with international dimensions. Since government actions are viewed as capable of influence and redistribution, the design of trade policy is supported and influenced by certain considerations like political, social, and cultural nature(Gaisford & Kerr, 2007).

Andreas Dür (2015) considers commercial policy as the governmental segment aimed at either facilitating or constraining the amount of international trade. He brings a broader idea of a commercial policy that encompasses a strong element of negotiation since trade policies do affect trading partners apart from unilateral decisions. Once a country engages in international trade, it is constrained to agreements among the participants, bilateral or multilateral. He summarizes trade policy into liberalizing and protectionist policies that are applied unilaterally or by negotiations, and in a preferential or nonpreferential way. The following table describes these dimensions of trade policies.

Source. Andreas Dür (2015) TABLE 1.

(11)

2.2.Type of International Trade Policies

The implementation of a restrictive trade policy is specified in a diverse group of measures and instruments that countries can put into practice. The most commonly used protectionist policies are discussed below.

There exist two main groups of protectionist policies: tariffs and non-tariff restrictions or NTBs (Ferrara, 2013). The most conventional protectionist barrier is tariff usually used to protect agricultural goods, textiles, apparel, and footwear are the most protected goods(GTA, n.d.). This barrier is the amount of tax or levy imposed on foreign goods paid to the government. Given the trade of goods between two countries, imported products pay a certain amount as fee by the receiving country. Tariffs are added to the cost of imported goods, which turns goods more expensive (Krugman et al., 2018). It includes specific or ad-valorem tariffs. A specific duty is a tax charged on each unit of an imported product. The expression of ad valorem comes from Latin, which stands for ‘according to value.’ Tariffs are also taxed as a percentage of a good value, which is the ad valorem tariffs. Due to inflation, ad valorem tariffs are more implemented relative to the specific duty (Grimwade, 1996).

Previously, countries used to get funding from tariff revenue. Nowadays, developed countries have different sources of income, while tariff revenues are still a way of financing developing countries' spending (Chambers & Sampson, 2008). After 2008, countries largely developed new tools to shield their domestic economy as a response to crises, called non-tariff barriers (Ferrara, 2013). The second type of barrier is any government mechanisms that challenge or impede the access of products to a country, other than tariffs (Ferrara, 2013). The Dictionary of Trade Policy Terms (2007) uses two interchangeably terms in practice with slightly different concepts. It refers to NTBs or Non-Tariff Barriers when there is a mechanism

(12)

placing a foreign good at a disadvantage, while, NTMs or Non-Tariff Measures because it restricts good and services, rather than only goods.

NTBs shows a wide form of trade barriers. Some of them can be drastic, meaning that it could lead to the complete prohibition of importations like import bans, or smoother like the adoption of standards of quality. The following are the most common in international trade (Ferrara, 2013).

Licenses are granted to businesses before they can import a specific good into a country. Given a food product, for example, only companies granted with licenses can act as importers of that product. Governments tend to use licenses to less competition for domestic products. Import quotas are another popular barrier associated with international trade licensing agreements. This NTB limits the quantity of importation of a particular good. In 2012, for example, Ecuador introduced an import quota on CKD parts for motor vehicles, which established a maximum volume of those imports (GTA, 2009-2012).

Voluntary Export Restraint (VER) is a special case of trade restriction because it is quota implemented by the exporting country, rather than the importing one. Generally, it is called voluntarily because the introduction of VERs results from a mutual agreement among the countries (Krugman, Obstfeld, & Melitz, 2018). These restrictions are typically based on availability and political ties. According to Wang (2011), VERs are more costly for the importing country than for the exporting. Local Content Requirement is another NTB that requires the exporting country to produce a specific percentage of a good to be from the importing side (Krugman, Obstfeld, & Melitz, 2018). In 2012, the Ecuadorian government used localization incentive for motor vehicles. There was a tariff discount for products containing between 5% and 20% domestic pieces (GTA, 2009-2012). Apart from the already mentioned NTBs, there are many other forms that governments make use to restrict trade like export

(13)

subsidies, intellectual property, or red-tape barriers. Trade barriers presented in the below graph were among the most used during the intense neo-protectionist phase in Ecuador (GTA, 2009-2012).

FIGURE 2

TYPE OF PROTECTIONIST TRADE BARRIERS IN ECUADOR (2009-2012)

Source. Global Trade Alter(2009-2012)

2.3.Regional trade negotiations and world trading patterns

After World War II, the world economy experienced an enormous and rapid economic expansion as a consequence of trade liberalization. By 1947, the creation of the General Agreement on Trade and Tariffs(GATT) led global exports to grow by an average of around 10% by 1973 (Ray, 1998).

Years after, the global economy continued experiencing a similar trend, but at lower rates. Although some of the less-developed countries like South Korea reached similar income levels as industrialized countries, most of the developing countries did not show improvements as much as developed ones, especially Africa, East Asia, and Latin America (see Figure 3).

5%

16%

67% 7%

3% 2%

Import licence requirement Import quota

Import tariff

Intellectual property protection Local Content Requirement Others

(14)

According to Krugman, Obstfeld, and Melitz (2018), various developing countries applied protectionist trade policies based on the idea that a robust manufacturing sector would lead to economic development, while industrialized countries went towards free trade.

Trade negotiations became an essential instrument of trade policy over time (AndreasDür, 2015). In 1993, significant developments in international trade policy reached after the conclusion of the Uruguay Round in 1993 that led to the creation of the World Trade Organization(WTO) (Grimwade, 1996). The WTO replaced the GATT and absorbed its responsibilities: the regulation of trade between nations through multilateral trade negotiations promoting and removing trade barriers (WTO, n.d.).

In recent years, countries commenced opting for Preferential Trade Agreements(PTA). The WTO considers PTAs as unilateral trade preferences and incorporates the Generalized System of Preferences schemes in which developed countries grant to the less industrialized economies reduced tariffs to their imports (Chambers & Sampson, 2008).

(15)

Regional Trade Agreements are also a way to enforce relations with trade partners. These agreements refer to signed pacts among two or more countries. In the case of Ecuador, it joined the WTO in 1996 and is part of most Latin American commerce agreements. Table 2 provides a list of which agreements Ecuador has membership.

TABLE 2.

ECUADOR MEMBERSHIP IN REGIONAL TRADE AGREEMENTS

Year Regional Trade Agreements

1988 Andean Community (CAN)

1983 Ecuador-Mexico

2017 El Salvador-Ecuador

2017 EU-Colombia and Peru

1989 Global System of Trade Preferences among

Developing countries (GSTP)

1981 Latin American Integration Association (LAIA)

Source. World Trade Organization

The majority of members of the WTO are developing countries. The Organization highlights the participation of developing economies into the global market and the importance of trade in economic development for those economies. For many years, developing countries, especially the smallest and most vulnerable, were reluctant to engage in the world trading system. The structural and economic problems, as well as the fear competition, are reasons they argue. Nowadays, WTO trade agreements include special facilities in developing countries, such as technical and legal assistance or extra time (Grimwade, 1996).

(16)

2.4. Effects of Protectionism

This subsection examines the most common protectionist policies used by the Ecuadorian government from 2009 to 2012. Partial equilibrium welfare effects are analyzed, taking this nation as the home country for three sides: consumers, producers, and the government.

The main effect of an import tariff is a decline in trade and higher domestic prices for the protected good (Krugman et al., 2018). Consider that Ecuador introduces an import tariff on red onions from Peru(GTA, 2010). Firms import less red onions as their cost raised by the amount of the tariff. In turn, they have to increase the price of the red onions to payout the reduction in the quantity and the higher importing expenses. In the domestic market, the quantity supplied by domestic producers increases while lowering the foreign supply of red onion at a higher price. Generally, this effect is the tariff objective to protect domestic producers from foreign competition. The cost and benefits vary across participants—domestic producers gains from higher prices. The government is benefited from the collection of tariff revenues. However, consumers lose welfare because they pay more for the product than before.

Similarly, import quotas cause the domestic good price to rise and limit the volume of importations (Krugman et al., 2018). In 2012, Ecuador imposed import quotas on motor vehicles (GTA, 2012). With the idea of giving protection to infant industries, the Ecuadorian manufacturing sector. In this case, when imports reach the limit of the quota, the demand for motor vehicles surpasses the total supply with the country, domestic plus foreign supply. The price of motors increases within the country as a consequence. Regarding welfare effects, import quotas have similar effects as tariffs, but without government revenues from the barrier.

(17)

Suppose Ecuadorian firms import vehicle parts to assembly autos. Local content requirements protect parts produced domestically, like import quotas. However, this barrier does not limit imports largely because firms are allowed to import, but with a condition. Importing firms must purchase some local inputs for their production. Thus firms face average prices on car pieces resulted from the imported and locally purchased vehicle parts. On the other hand, consumer’s losses are palpable when they buy the final product only.

At first sight, the net effect seems positive. Producers receive more revenue, which also increases wages for the labor input. In case government revenues are present, they can be invested in education to improve human capital. However, the boundaries among the groups are blurred. Firm workers are also consumers. Then, even the income gain from the change in wages may be banished by the losses in consumption, an offsetting effect. Additionally, asymmetric power within a firm may reduce wage gains, even further preventing resources from being transfer as they should be.

Furthermore, there is no guarantee that trade barrier revenues will be allocated properly. The spillover of higher prices leads to lower overall demand that may cause wage reductions and job losses in other industries. As production decreases because of lower demand, intermediate inputs prices for protected sectors may be subject to the protection leading to higher production costs and lower wages as well.

2.5.Effects on wages

2.5.1. Free trade effect on wages

Another notorious argument concerning trade liberalization is its strong distributional impact within countries. There exists a wide range of theoretical literature regarding the effects on wages. The specific factors model and Heckscher-Ohlin model shaped some theoretical bases for the link among trade openness and distributional effect on wages.

(18)

The factor-specific model predicts an ambiguous income effect for workers (Krugman, Obstfeld, & Melitz, 2018). Trade alters the price of goods, causing production decisions to adjust. Nominal wages rise due to production decisions resulted from positive price changes. Nonetheless, real wages may either increase or decrease based on changes in workers’ purchasing power.

The Heckscher-Ohlin(HO) model predicts that economies with relatively abundant unskilled labor will specialize and, thus, export goods that are unskilled-labor intensive (Goldberg & Pavcnik, 2007). The Stolper-Samuelson theorem of the model affirms increases in the return of the intensive-factor in the production of a good (Krugman, Obstfeld, & Melitz, 2018). In this case, labor wages rise when the good that is exported is labor-intensive. For developing countries, this argument is valid as they are abundant with labor, especially unskilled labor (Goldberg & Pavcnik, 2007). It suggests that changes generated by trade in developing countries favor unskilled workers. Nevertheless, is this accurate?

Both models conclude that the losing sectors are less than the winners causing an overall welfare increase. However, those distributional effects occur under some particular assumptions that are violated in reality (Edwards, 1988). They neglect the fact that countries show differences in technology and that labor is not perfectly mobile within an economy. The later brings out the belief of all workers being equal and that once trade effects adjust market labor, wages are equalized in the economy.

According to Sebastian Edwards (1988), distributional trade effects rely on the structure of the labor market in a country. He introduces a sector-specific wage rigidity in the OH model analysis where trade liberalization generates significant short-run unemployment problems.

To sum up, trade increases the overall wage level, but it creates winners and losers. Workers in industries that face competition from importations may find their demand falling,

(19)

and their wages declining with a rise in international trade. On the contrary, workers in exporting find their demand and wages increase (OpenStax Economics, 2016).

2.5.2. Protectionist policy effect on wages

In 2016, Thompson analyses tariff effects on factor prices in small open economies. From this angle, tariffs boost wages only under a limited group of assumptions. Assuming two production factors and two goods, tariffs increase wages if the importing sector is labor-intensive. He, nonetheless, states that this finding should not be assumed because once the model is relaxed, the effect weakens. Yet, what do empirical studies say about trade barriers? Does it help to protect wages from international trade?

The idea of unskilled workers competing against very low-wages from the rest of the world in manufacturing products like clothing has been making room in the U.S. policymaking in favor of protectionism for several decades(OpenStax Economics, 2016). Gaston and Trefler (1994) conducted a study to isolate the effect of international trade policy on wages of the manufacturing sector in the U.S.. They found that wages in protected import industries were lower to those export-oriented industries that were unprotected. Tariff barriers showed a significant negative correlation with industry wage. Non-tariff barriers insignificant positive relationship with wages. In 1983, increases in protectionist policies reduced industry protected wages by around 3% compared to those unprotected. Moreover, there was no evidence that tariffs do increase unskilled labor wages.

If these results applied to Ecuador, a developing country, it is worth pointing out that low-wage workers would suffer due to protectionism in industries. Commonly protected industries are food and clothing. Again, low-wage workers would find these necessities more expensive, and as such, their income stretches in real terms. While in the US, low-wage and unskilled workers comprehend a very small portion of the population, and, therefore, the impact

(20)

of tariffs is small in this regard (Gaston & Trefler, 1994). In Ecuador, low-income households, poverty, is around a quarter (INEC, 2019).

Contrary to these conclusions, Goldberg and Pavcnik(2005) pointed out that such results do not apply to developing countries. Studies like the one of Gaston and Trefler focused on cross-sectional data, which refuses the effect of time changes on wages. Goldberg and Pavcnik(2005) affirm that wages across industries in developing countries are more volatile than those in developed countries, in essence, in the US.

During the 1980s, Colombia went through absolute protection for some industries with high tariff levels. Goldberg and Pavcnik(2005) researched Colombia and found that trade protection improves sectorial relative wages in the short and medium-term when there exist labor market rigidities. Although results without controlling for industry-specific factors seem that there was a negative correlation between tariffs and wages. Once controlling for these differences, tariffs reverse the effect. Moreover, the authors found that tariff reductions in industries with less-skilled workers led to a decrease in their wages.

2.6.Determinants of wages:

Many factors can influence wage determination like education, employment rate, working hours, GDP per capita, or informality in the case of developing countries.

Wages are increased by education once a worker completes a level of schooling (Borjas, 2016). The rate of return to schooling is estimated to be around 9 percent for decades in the US (ibid). For example, a worker that has graduated from university earns more salary than a worker that has only completed primary school. The employment rate is and indicator of the behavior for the demand or supply of labor(ibid). Generally, wages are paid based on the hours a person works. A person may choose to work more or fewer hours, depending on his/her utility.

(21)

Positive economic growth phases, usually, lead to higher wages. Theoretical background claims that GDP per capita growth is enhanced by the higher productivity levels making wages to increase (Ray, 1998). On the other hand, a recession lowers GDP per capita and, so, wages.

In the case of developing countries, the structure of their labor market, as Edwards (1988) states, it is important. Approximately 50 percent of the workers are part of informal jobs in Latin American economies (ILO, n.d.). Ecuador’s labor market shows informality at 45%(INEC, 2019). Informal jobs have detrimental effects on the labor market. It does not comply with labor market laws like minimum wages and is more likely to function under poor working conditions like extremely low wages (Meghir et al., 2015). Strengthening enforcement over informality towards formality leaves unemployment unchanged and increases wages (ibid).

2.7. Hypotheses:

The discussion up to this point, trade benefits the overall economy and creates winners and losers. Countries implement different trade policies that have been developing over time for political and economic reasons. The World Trade Organization and its agreements regulate countries' decisions on trade policies to increase international trade and their welfare.

General theory literature and studies focused on developed countries, the US, concludes that there is a negative impact of protectionist policies on wages. By contrast, there are some valuable arguments that those finding do not apply to developing nations because their labor market is different from developed ones. The research on Colombia trade policy reforms has concluded a positive effect of tariffs on sectorial wages. Following this idea, theory assumptions would not hold with the reality of Ecuador, and the empirical evidence from Colombia seems more promising. Moreover, Colombia and Ecuador are similar in terms of economic structure

(22)

and labor markets. Hence, my hypothesis lies in the expectation of a positive effect of trade barriers on wages. Nonetheless, it does not neglect the opposite to happen.

3. Methodology

This section discusses how the research proceeds. First, the regression model, assumptions are presented. Next, it goes through the data used: its sources, operationalization, and descriptive analysis. Finally, a brief discussion and conclusions concerning the model are elaborated.

3.1.Models and Assumptions:

Many macroeconomic studies encounter with particular challenges. Most data on the field contains a small number of observations that are dispersed. Also, there exist structural differences among observations. Furthermore, finding randomness across variables make it difficult to implement experiments

The present study deals with the same difficulties. There are data on 23 economic sectors across 13 years, and every sector has its specific characteristics. We are interested in the effect of trade barriers, independent variable, on sectoral wage and sectoral wage growth, the dependent variables, in Ecuador during 2007 and 2019. This empirical research assumes a linear relationship between the variables. However, as mentioned before, linear regressions may face the problem of omitted variable bias or reverse causation. Despite pursuing the effect of protectionist restrictions on sectorial wages, there may be another variable apart from trade barriers influencing wages and trade barriers, for example. As a result, filtering the effect of the treatment becomes complicated.

(23)

I considered a Difference in Difference(DiD) approach. Difference-in-Difference research design is commonly practiced to assess the effectiveness of policies (Bertrand et al., 2004). This methodology is a quasi-experimental design and uses observational data from treatment and control samples to draw the causal effect of a treatment (Angrist & Pischke, 2015).

The present work contemplates the last indicated approach as suitable aiming the effect of restrictive trade policies on wages. The implementation of trade barriers is labeled as the treatment which is applied to one out of two groups. A set of industries that is subject to trade restrictions while another set that is not. The DiD method compares the average changes in wages and wage growths, the interesting outcomes, across the two groups of economic sectors over time. The approach uses panel data to estimate the causal effect and also exploits the time variation in the data.

The design requires some key elements: a dummy variable capturing the general differences between industries, dummy variables capturing the period after the implementation of barriers, and interaction variables among the previously mentioned dummies. Despite my target is to know whether trade policies successfully protected sectoral wages after the four-year of implementation. I have considered realistic to generate another dummy to prove the effectiveness of the trade policy in Ecuador from 2007 to 2019. There are trade policies that may have influenced wages within 2009 and 2012, like those that were implemented in the first years. Therefore, there is a dummy for periods during and after the implementation. The coefficient of the interaction variables provide the effect of interest; there is a positive effect of protectionist policies on sectorial wages.

Likewise, the thesis ponders to research two different levels: an extensive and intensive assessment. The extensive analysis tries to discover whether the sectorial wage level or growth has been influenced given industries were subject to protectionist policies or not, regardless of

(24)

the volume of restrictions they were subject. The intensive analysis, on the other hand, attempts to reveal the impact of trade barriers on sectorial wages conditional on the number of restrictions industries were exposed to. It is expected that more restrictions would lead to a stronger effect on industries.

Thus, in order to test the hypotheses derived in the literature review section, the thesis considers two models to conduct the research:

3.1.1. Model 1: Extensive Regression EQUATION 1

𝑊𝑖𝑡 𝒐𝒓 𝑊̂ = 𝛼 + 𝛽𝑖𝑡 1𝑇𝑟𝑒𝑎𝑡𝑒𝑑𝑖 + 𝛽2𝐷𝑢𝑟𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛𝑡 + 𝛽2𝑃𝑜𝑠𝑡𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛𝑡

+ 𝛿₀𝐷𝑢𝑟𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡_𝑖𝑡+ 𝛿₁𝑃𝑜𝑠𝑡𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡_𝑖𝑡+ 𝜃𝑋_𝑖𝑡+ 𝑒_𝑖𝑡

Where i is equal to 1, 2, …,23, and indicates the industry variable. The index t is 1, 2, …, 13, and denotes the years from 2007-2019.

The endogenous variables are 𝑊_𝑖𝑡 and 𝑊̂ . 𝑊_𝑖𝑡 _𝑖𝑡 represents annual average wage level per month of industry 𝑖 at time 𝑡 . 𝑊̂ is the growth of the variable 𝑊𝑖𝑡 𝑖𝑡. The

constant-coefficient is 𝛼 and denotes the autonomous level or growth of sectoral wages. The error terms that have been used for the models is 𝑒_𝑖𝑡 .

It is defined as the treated group as the industries targeted by the implementation of trade barriers and control groups otherwise. Then, 𝑇𝑟𝑒𝑎𝑡𝑒𝑑𝑖 variable is the industry-specific

dummy variable, where 1 represents directly affected industries by trade barriers among 2009 and 2012, and 0 otherwise. This analysis wants to find out whether the sectorial wage level or growth has been influenced given industries were subject to protectionist policies or not. Here, it is completely ignored the volume of policies implemented. For example, industries may have been affected by multiple restrictive measures to trade and others with only one barrier. Both

(25)

industries are assumed to be affected in the same way. Then, 𝑇𝑟𝑒𝑎𝑡𝑒𝑑_𝑖 = 1 only means that industries where subject to trade barriers.

𝐷𝑢𝑟𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛_𝑡 is a time dummy variable where 1 represents the phase during the implementation of trade barriers (2009-2012) and 0 otherwise. 𝑃𝑜𝑠𝑡𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛_𝑡 is another time dummy variable where 1 is the phase after the implementation of trade restrictions(2013-2019) and 0 otherwise. 𝐷𝑢𝑟𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛𝑡 and 𝑃𝑜𝑠𝑡𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛𝑡

shows the pattern that the treated and control groups follow, during, and after trade barriers are implemented. During the four years of implementation, both groups would present different paths if there was any effect. It is expected the same result in the period after the implementation.

𝐷𝑢𝑟𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡𝑖𝑡 and 𝑃𝑜𝑠𝑡𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡𝑖𝑡 are the interaction variables resulted from the

multiplication of the industry dummy and the former time dummy variables. 𝐷𝑢𝑟𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡_𝑖𝑡 is equal to 𝑇𝑟𝑒𝑎𝑡𝑒𝑑_𝑖∗ 𝐷𝑢𝑟𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛_𝑡. Its coefficient, 𝛿₀, tells the effect of trade barriers on sectorial wages during the treatment introduction phase of sectors that were targeted by the barriers relative to those that were not. 𝑃𝑜𝑠𝑡𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡𝑖𝑡 is equal to 𝑇𝑟𝑒𝑎𝑡𝑒𝑑𝑖∗

𝑃𝑜𝑠𝑡𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛_𝑡. Its coefficients, 𝛿₁, shows the same effect as 𝛿₀ estimate, but in phase after the introduction of the barriers.

𝑋𝑖𝑡 represents a set of explanatory control variables: monthly Hours worked, GDP per

capita growth, education, labor formality ratio, and the labor rate.

3.1.2. Model 2: Intensive Analysis

𝑾𝒊𝒕𝒐𝒓𝑾̂ = 𝜶 + 𝜷𝒊𝒕 𝟎𝑰𝒏𝒕𝒆𝒓𝒗𝒆𝒏𝒕𝒊𝒐𝒏𝒊+ 𝜷𝟏𝑫𝒖𝒓𝑰𝒎𝒑𝒍𝒆𝒎𝒆𝒏𝒕𝒂𝒕𝒊𝒐𝒏𝒕+

𝜷𝟐𝑷𝒐𝒔𝒕𝑰𝒎𝒑𝒍𝒆𝒎𝒆𝒏𝒕𝒂𝒕𝒊𝒐𝒏𝒕+ 𝜹𝟎𝑰𝒏𝒕𝒆𝒓𝒗𝒆𝒏𝒕𝒊𝒐𝒏𝑫𝒖𝒓𝒊𝒕+ 𝜹𝟎𝑰𝒏𝒕𝒆𝒓𝒗𝒆𝒏𝒕𝒊𝒐𝒏𝑷𝒐𝒔𝒕𝒊𝒕+

𝜽𝟎𝑿𝒊𝒕+ 𝒆𝒊𝒕EQUATION 2

The 𝑖 and 𝑡 indexes show industry and year items, respectively. The dependent variables, error term, and the time dummy variables have the same connotation as in the extensive analysis.

(26)

Instead of zero and one values of the treatment variable as in the extensive analysis, the intensive analysis applies the treatment variable as the number of interventions each industry was faced to. If the manufacturing industry, for example, was exposed to 19 restrictions, the treatment variable is 19 instead of 1, and if the restrictions did not target an industry, then the value of the treatment becomes 0. Then, Model 2 defines the 𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛_𝑖 variable as the number of interventions per sector between 2009 and 2012.

𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛𝐷𝑢𝑟𝑖𝑡 and 𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛𝑃𝑜𝑠𝑡𝑖𝑡 are interaction variables from the

multiplication of the previously time dummy variables and the new generated treatment variable. 𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛𝐷𝑢𝑟_𝑖𝑡 is equal to 𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛_𝑖 ∗ 𝐷𝑢𝑟𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛_𝑡 𝑃𝑜𝑠𝑡𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡_𝑖𝑡 is the result of 𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛_𝑖 ∗ 𝐷𝑢𝑟𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛_𝑡. The coefficients 𝛿0 and 𝛿1 are the effects of interest in the intensive analysis. They check whether the number

of trade restrictions does matter in the impact of them on sectorial wages during and after the policy implementation based on its number.

3.2.Data

The data is collected for 21 economic sectors in Ecuador, from 2007 until 2019. Moreover, it is included two more sections, one for the non-specified economic activities and one for an aggregate division. Different sources have been used to gather all data for this thesis. The resulting panel data present missing data for some industries in different years, which makes the panel data unbalanced.

3.2.1. Dependent Variable

Sectors or industries are classified according to the International Standard Industrial Classification of All Economic Activities or ISIC. Its classification encompasses mutually exclusive economic activities within the margins of the System of National Accounts. This classification is widely accepted by countries, including Ecuador, and organizations globally. The International Labor Organization(ILO), the International Monetary Fund(IMF) or the

(27)

World Bank are some of the UN bodies making use of the ISIC (Department of Economic and Social Affairs, 2008). The following table shows the aggregate classification that is used in this thesis.

Source.International Standard Industrial Classification of All Economic Activities

The International Labor Organization is a prominent and reliable labor data source worldwide. Its empirical data collection method and reproduction are gathered from national statistical offices around the world (ILO, n.d.). Therefore, this thesis chose Ecuadorian official sources to conduct the research. The 2019 Official Gazette includes the Consumer and Producer Price Index and labor market data. Data on historical series on wages per sector, the dependent variable, is taken from the Central Bank of Ecuador. Wages are published as nominal wage earned per month in US dollars. Gross Wage is the value paid monthly by the establishment or employer to its employees or workers. It includes monthly values of wages(before deducting various taxes and discounts), overtime, commissions, production bonuses, subsidies, transportation, among others (BCE, 2019).

TABLE 3.

(28)

3.2.2. Treatment

For the treatment variable, data on protectionist policies are collected from the Global Trade Alert(GTA) website. The GTA database is a complete and comprehensive monitor on policy changes influencing international trade (Evenett, 2019). The used database provides information about the type of intervention and the industries directly affected by those restrictions (GTA, 2009-2012).

During 2009 and 2012, the Ecuadorian government established 36 protectionist policies. By 2019, there were 22 still in force. Note that one intervention may affect more than a sector. Also, the data only consider Red Alters. The Global Trade Alters sorts the level of harmfulness with Red as the highest on impact. Those interventions certainly discriminate against foreign commercial interests according to this source (Evenett & Fritz, 2020)

3.2.3. Control Variables

Data on the control variable of Hours worked per sector is collected from the same source as the endogenous variable, the Central Bank of Ecuador. It defines Hours worked as the number of ordinary and extra hours worked in the month for a year that is reported. Labor structure market indicators like labor formality, employment rate, and education of the employed population is retrieved from the Ecuadorian National Institute of Statistics and Censuses or INEC(2019). The National Survey of Employment and Unemployment contains historical series on those indicators in percentage terms. The censuses considered employed as all those that are at working age who, during the reference week, were engaged in some activity to produce goods or provide services in exchange for remuneration or benefits (INEC, 2019).

Ecuadorian employment has four differentiated levels of education. Some of the employed have not received any formal education, called None_it. 𝑃𝑟𝑖𝑚𝑎𝑟𝑦 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛_𝑖𝑡, and 𝑆𝑒𝑐𝑜𝑛𝑑𝑎𝑟𝑦 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛_𝑖𝑡 are those employed workers that have attended school and high

(29)

school, respectively. Workers who received college or further education are defined as 𝐻𝑖𝑔ℎ𝑒𝑟 𝑒𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑖𝑡 (INEC, 2019).

Furthermore, Ecuador's data on its labor market provides information about its informality. The INEC(2019) defines employment in the informal sector as the set of people who work in production units of less than 100 workers and do not pay taxes or are not registered by the government.

Furthermore, data on GDP per capita growth was downloaded from the World Bank Databases. The World Bank (n.d.) defines this indicator as to the percentage growth on local currency per year.

3.2.4. Procedure

For the dependent variables wage growth, 𝑊̂ , is calculated from the level wage, _𝑖𝑡 𝑊𝑖𝑡, per month using its natural logarithm, and deducting the natural logarithm of wage level at

𝑡 − 1 from the natural logarithm of wage level at 𝑡. Real wage may be a better variable; however, it did not alter the results. Additionally, reported nominal wages did not variate once inflation was considered.

To generate the treatment variables, the raw data on trade barriers is manipulated as follows. First, data on restrictions are matched with each economic sector, given the type of intervention and the code of sectors they affected. Next, each intervention type is given the value of 1 and is added up as it is seen in the next table. Sector E, for example, was intervened with an import tariff, import license requirement, and intellectual property protection barriers (GTA, 2009-2012). All these restrictions are summed up and give a value of 3.

(30)

TABLE 4

For the extensive model, it is generated a [1,0] variable, Treatedi , to distinguish

targeted sectors by the intervention. Indeed, the primary and secondary sectors of the economy are the most protected.

Sectors A, B, C, D, E, F, G, H, I, Q, and R received the treatment. While for the intensive model, 𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛_𝑖 is generated, and sectors A, B, C, D, E, F, G, H, I, Q, are given their respective number of trade barriers.

Industry Total

Interventions

A. Agriculture, farming, forestry, and fishing 14

B. Mining and quarrying 9

C. Manufacturing 19

D. Electricity, gas, steam and air conditioning supply 7 E. Water supply; sewerage, waste management and remediation

activities

3

F. Construction 8

G. Wholesale and retail trade; repair of motor vehicles, motorcycles and personal and household goods

22

H. Transport and storage 5

I. Accommodation and food service activities 11

J. Information and comunication 10

K. Financial and insurance activities 0

L. Real estate activities 0

M. Professional, scientific and technical activities 0

N.Administrative and support service activities 0

0. Public administration and defence; compulsory social security 0

P. Education 0

Q. Human health and social work activities 3

R. Arts, entertainment and recreation 9

S. Other service activities 0

T. Activities of households as employers; undifferentiated goods and services producing activities of households for own use

0 U. Activities of extraterritorial organizations and bodies 0

(31)

In both models, the research ponders 13 years period, and it is divided into three phases: a pre-implementation phase for all years before the treatment, 2007, and 2008. This phase is used as a reference in the regressions for which it is not included. 𝐷𝑢𝑟𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛_𝑡 is created for the phase during the implementation of protectionist policies for 2009, 2010, 2011, and 2012. 𝑃𝑜𝑠𝑡𝐼𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛𝑡 is generated for years after or equal to 2010.

Given control variables in percentages, new variables out of the education rate, and formality structure rates are created.

The four levels of education are transformed into a category variable, 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑖.It

provides flexibility and inclusion of all four levels rates from the employed population. Where 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛_𝑖𝑡 is equal to 1 ∗ 𝑁𝑜𝑛𝑒_𝑖𝑡+, 2 ∗ 𝑃𝑟𝑖𝑚𝑎𝑟𝑦 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛_𝑖𝑡 + 3 ∗ 𝑆𝑒𝑐𝑜𝑛𝑑𝑎𝑟𝑦 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑖𝑡+, 4 ∗ 𝐻𝑖𝑔ℎ𝑒𝑟 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛𝑖𝑡. Also, 𝐹𝑜𝑟𝑚𝑎𝑙 𝑟𝑎𝑡𝑖𝑜𝑖𝑡 is produced from

𝐹𝑜𝑟𝑚𝑎𝑙𝑟𝑎𝑡𝑒_𝑖𝑡⁄𝐼𝑛𝑓𝑜𝑟𝑚𝑎𝑙𝑟𝑎𝑡𝑒_𝑖𝑡. Weekly worked hours are transformed into monthly worked hours, 𝑀𝑜𝑛𝑡ℎ 𝐻𝑜𝑢𝑟𝑠_𝑖𝑡 = 𝑊𝑒𝑒𝑘𝑙𝑦 𝐻𝑜𝑢𝑟𝑠_𝑖𝑡∗ 4.345.

3.3.Descriptive analysis

The table below provides a summary of statistics on the principal variables. The number of observations is not equal for all variables. Given panel data, there are industry and/or years that do not have data on some of the rest variables, which means that the panel data is unbalanced. For example, 𝑊̂ shows fewer observations because of its construction used data _𝑖𝑡 on wages from previous years. Data only comprehend years above 2007; then, calculating the wage growth of a particular industry in 2007 is not possible. Across industries and years, the average sectorial mean is 502.19 USD, while its growth has been increasing by around 4%. Interestingly, wages variance is indicating that wage observations points are very spread out from the mean of 502.19 USD, and one another. This can be seen in the minimum and maximum values of wages. There may be outliers and heteroskedasticity.

(32)

TABLE 5.SUMMARY STATISTICS OF VARIABLES

Wage distributions tend to be skewed to the right, as in the graph below. Average wages exceed the median ones. Latin America is the most unequal region globally. The graph confirms that top percentiles of wages account for a really small share of the total.

TABLE 6 0 2 0 4 0 6 0 8 0 1 0 0 F re q u e n c y 0 500 1000 1500 2000 2500 Wages Ecuador, 2007-2019

Distribution of Sectorial Wages

Variable N Mean Variance Standard Deviation Minimum Maximum

Sectorial Wage(𝑾𝒊𝒕 ) 290 502.1908 80824.5 284.2965 72.5 2567 Sectorial-WageGrowth( 𝑾̂) 264 .0376681 𝒊𝒕 0.039239 .1980878 -1.237875 1.346579 Treated 286 .5909091 0.242583732 .4925279 0 1 Total Intervention 295 1.627119 18.84008 4.340516 0 22 Labour rate 208 .0625 0.002073 .0455288 .0078953 .1935439 Formal ratio 299 1.079665 0.010361 .1017908 .9075736 1.247064 Education 299 2.969069 0.003056 .0552798 2.860134 3.045592 GDP per capita 299 1.354886 6.563875 2.562006 -2.898025 6.221052

(33)

There are outliers around 2500 USD, and results may be sensitive to them. Results did not change when outliers were dropped. Wages that were above 1500US and bellow 100US were eliminated.

Over time all sectorial wages show positive slopes, treated and non-treated. Surprisingly, sectorial wages were equal on average at the end of the implementation phase. It is expected that it was because of the treatment.

FIGURE 4

Besides, Difference-in-Difference regressions have an underlying key assumption, parallel trends. This means that treated sectoral wages would have followed the same trend as the control ones if the treatment would not exist. Tha parallel tren assumption is extremely difficult to prove as there is no other reality in which the treated were not treated. Thus, it assumed to be true for our data.

3 0 0 4 0 0 5 0 0 6 0 0 7 0 0 8 0 0 W a g e 2006 2008 2010 2012 2014 2016 2018 2020 Year

Period of Implementation Post-Implementation Period

Control Treated

Ecuador, 2007-2019 Sectorial Average Wages

(34)

-5 0 0 0 5 0 0 1 0 0 0 1 5 0 0 2 0 0 0 R e s id u a ls 300 400 500 600 700 Fitted values -5 0 0 0 5 0 0 1 0 0 0 1 5 0 0 R e s id u a ls 0 500 1000 1500 Fitted values -5 0 0 0 5 0 0 1 0 0 0 R e s id u a ls 0 200 400 600 800 Fitted values -5 0 0 0 5 0 0 1 0 0 0 1 5 0 0 2 0 0 0 R e s id u a ls 300 400 500 600 Fitted values -5 0 0 0 5 0 0 1 0 0 0 1 5 0 0 R e s id u a ls 0 500 1000 1500 Fitted values -5 0 0 0 5 0 0 1 0 0 0 R e s id u a ls 0 200 400 600 800 Fitted values -2 -1 0 1 2 R e s id u a ls 0 .05 .1 .15 Fitted values -1 -. 5 0 .5 1 1 .5 R e s id u a ls -1.5 -1 -.5 0 .5 Fitted values -. 6 -. 4 -. 2 0 .2 .4 R e s id u a ls -.05 0 .05 .1 .15 Fitted values

Sectorial Wages Growth

Basic DiD With Fixed Effects With Control Variables

Sectorial Wages Level

Basic DiD With Fixed Effects With Control Variables FIGURE 5EXTENSIVE ANALYSIS

Sectorial Wages Level

Sectorial Wages Growth

FIGURE 6INTENSIVE ANALYSIS -.6

-. 4 -. 2 0 .2 .4 R e s id u a ls -.05 0 .05 .1 .15 Fitted values -1 -. 5 0 .5 1 1 .5 R e s id u a ls -.1 -.05 0 .05 .1 .15 Fitted values -1 -. 5 0 .5 1 1 .5 R e s id u a ls 0 .02 .04 .06 .08 .1 Fitted values

(35)

Cross-sectional studies show large differences between its small or large observations; thus, they are likely to have heteroskedasticity (Greene, 2003). It is regressed the fitted values on the error term to test for this. Graphs 5 and 6 show the presence of heteroskedasticity on some of the regressions. As mentioned before, Ecuador has wages for people that work in the mining sector, as well as those that work as doctors that earn high wages. Furthermore, wages have changed a lot since the start of the study period. It seems that this is a case of pure heteroskedasticity due to the nature of the dependent variable.

Robustness command for the regression will be used. It checks for robust variance estimators.

4. Results:

This section provides the main results of the research and some insights from them, given the methodology. Extensive and intensive regression results are broken down. After that, a robustness check is mentioned.

Tables 5 and 6 contain the results for the first defined regression model. Regressions 1.1, 1.2, 1.3 and 2.1, 2.2, and 2.3 shows the estimates for the Sectorial Wage level, 𝑊_𝑖𝑡, while 1.4, 1.5, 1.6 and 2.4, 2., 2.6 are coefficients corresponding to the Sectorial Wage growth, 𝑊̂ . _𝑖𝑡 Wage growth estimates should be multiplied by 100 before giving any inferential meaning.

Regressions 1.1, 2.1, and 1.4, 2.4 only contain the basic Difference-in-Difference estimates: the constant, its industry and time dummies, and interaction variables. In 1.2, 2.2 and 1.5, 2.5 estimates are controlled for time and industry-specific features. This was done by creating dummies for all year and sector. They are not listed because they do not provide much information. Next, 1.3 and 1.6 regression are controlled for control variables: labor rate, the growth of GDP-per-Capita, education, and formality ratio.

(36)

4.1.Extensive regression’s results

According to 1.1, 1.2, and 1.3 regression estimates, the coefficients of interest are negative for both interaction terms, DurTreatment and PostTreatment. However, the estimates are not statistically significant. It does not appear any causal effect of trade barriers on wages or wage growth during or after the implementation. Likewise, Treated and DurImplementation coefficients are insignificant. Then, it is difficult to attach them to any inferential meaning. Trade barriers neither increase nor decrease protected sectorial wages.

However, if the coefficient of the interaction terms were significant for 𝑊_𝑖𝑡 and 𝑊̂ _𝑖𝑡 regressions. These coefficients show the additional effect on wages that were protected. Industries targeted by trade restrictions may be affected by reductions on the average wage level per month relative to those that were not targeted. In 1.1, relative to unprotected sectors, protected wages would show an additional average decrease of around 17 and 69 US dollars during and after the introduction of barriers, respectively. In 1.3, the additional effect on wages would be an increase of only 2.29US dollars during the implementation and a fall of 29.95US dollars after, compared to industries without protection. Given a mean of around 500, a reduction of 69 US dollars, for example, would represent 14% of worker wage, which is sizeable in a country like Ecuador. Following the idea, 1.6 regression would show that wage growth of protected industries fell by 2% and 6% additionally during and after the treatment, relative to unprotected sectoral wages.

By contrary, in 1.1, the estimate of the PostImplementation variable seems positive and statistically significant at a 1% significance level. The phase after the implementation of the protectionist policies, there was a sizable uniform effect of 269.20 US dollar increase on wages for all economic sectors, on average. Every industry performed better during 2013 and 2019. In 1.3, the post-period effect becomes weak as control variables are included. The constant,

(37)

also, is significant at a 1% significance level. As the treatment coefficients are not significant, increases in wages may be a matter of inflation.

In 1.3, the labor rate negatively affects the level of wages at a significance level of 1%. In reality, its size is smaller than the coefficient appears to be. A 1 standard deviation increase in labor rate causes wage to decline by 22.2 standard deviations (1/0.045). Then, an increase in the labor rate by 22 standard deviations lead to a decrease in sectorial wages by 8 standard deviations(2153.3/284.2964). Also, education appears significant at a 5% significance level in 1.3. The same basics as for labor rate applies here. A 1 standard deviation increase in makes wages to increase by 1.81 (1/0.553) standard deviations. Thus, if education rise by 1.81 standard deviations, sectorial wages rise by 6 standard deviations (1590.3/284.2965).

In 1.6, formality ratio shows some positive effect on wages at 10% p-value. If formality rises by 1 relative to informality in labor markets, sectoral wage growth rises by 35%.

(38)

4.2.Intensive regression’s results

The number of trade barriers shows insignificant impacts on sectorial wages without and with control variables in 2.1, 2.2, and 2.3. For sectorial wage growths, after controlling for explanatory variables in 2.6, the number of trade restrictions do affect negatively protected sectorial wages.

In the table below, its seen that all sectorial wages improved during and after the implementation of protectionist policies by 96 and 241 US dollars, respectively. The 2.3 regression shows that those effects have vanished when adding control variables. The positive effects may be due to education, while some negative effects come from the labor rate. Those impacts are similar to the previous table.

After controlling for industries and year fixed effects, in 2.2, heavily protected industries showed lower average wages relative to those that were not targeted by trade restrictions, the difference was about 63 US dollars. The effect weakens once it is controlled for the other variables in 2.3. It is important to say that most of the protected industries belong to the primary economic sector, and unprotected ones are mostly focused on services like financial or medical attention.

In 2.6, the number of trade barriers does affect sectorial wages on protected industries at 10 percent significance level. After the period of implementation, adding one more intervention to a protected industry reduces its wages growth by 0.5% extra relative to unprotected wage growths. The size of the estimate seems large as it represents approximately 13 percent of the average sectoral wage growths(around 4%). It represents 12.5% of the growth. In the same regression, formal ratio coefficient appears to be strongly significant (1% p-value). Sectorial wage growths increase by 35% when formality rate increase by 1 percent in relation with informality rate.

(39)

The research findings do not support the alternative hypothesis in most of the cases. Thus, the null hypothesis is not rejected and there is no enough evidence that sectorial wages deviate from the introduction of restrictive trade policies. The significant result, at 10 percent significance level, on wage growth rejects the null hypothesis and accepts that, indeed, there is an effect on wages from protectionism, a negative one.

(40)

5. Conclusion

The last section will deliver a review of the main findings of this investigation. As well, it highlights some of the limitations and recommendation for further future research on the topic.

The use of trade policy by countries has been controversial for developing countries in particular. Protectionism seeks to favor a country’s products and services by imposing limitations on imports (AndreasDür, 2015). However, the provided literature and empirical evidence neglect that protectionism improves distributional effect within a country and that workers are hurt by the fall in wages (Gaston & Trefler, Grimwade, Krugman et al., 1994,1996 & 2018). They explain that because of higher prices paid by consumers and distortion in markets, the gains are diminished. Indeed, these theories and studies advocate for relaxing trade.

On the other hand, international organizations recognize that less developed economies are in a situation of vulnerability and the specific structural obstacles they face in the global economy. The reasons for this marginalization are complex and encompass deep-rooted structural problems, flimsy regulatory institutions and frameworks, and internal and external protectionist structures (Chambers & Sampson, Edwards, Goldberg & Pavcniks; 2008, 1988 & 2007). Unlike the conventional findings, Goldberg and Pavcniks (2007) reflected the positive influence of restrictive trade policies on sectorial wages in a developing economy.

To my knowledge, there has not been much empirical analysis on this matter for developing countries and, even more, the effects of trade policies within an economy. This research aimed to contribute to the field by isolating the trade policy effects on income distribution across protected industries and unprotected industries. Considering the results, sectorial wage levels of protected industries were not influenced in either direction by trade

(41)

policies, even though wages increased at the time and after the introduction of protectionist policies. Those boosting effects seem to lie in other variables, like education or formality ratios.

This research presented some limitations while being conducted. Most of the data on sectorial levels were not available, which left the author the present study with limited and general data. It was considered that labor productivity, for example, plays a crucial role in the determination of wages. Furthermore, sectorial data on this measure was only available for years after 2016, for example. It is recommended that it would be more relevant to investigate at a firm-level because of more detailed and existing data. Finally, a research design like fixed effects, following the path of the Colombian case, may have the potential to differ from the results of this thesis. Yet, the developed model is solid as its methodology can be applied to all different country contexts. Thus, further research can be conducted in other countries and preferably focused on wage growth rather than its levels.

Although the volume of trade barriers did not appear to influence the sectorial wages under the used control variables. As the volume of trade restrictions rose in number, wage growths in industries targeted by many trade restrictions significantly fell by an extra 0.5% relative to unprotected wages.

To sum up, most of the outcomes showed that the Ecuadorian protectionist phase during 2009 and 2012 did not influence in sectorial wages and its growth once an industry is protected. Nonetheless, the insignificant estimates did show some direction, and the effect may be really small that the model is unable to detect such effect. Another reason for those results may be that there was actually no effect of the protectionist program in Ecuadorian sectorial wages. And, it is probably that “Protectionism” is neither a curse-word as international organizations claim nor a blessing-word as the previous Ecuadorian government emphasized.

(42)

References

AndreasDür. (2015). International Trade: Commercial Policy and Trade Negotiations. In J. D. Wright, International Encyclopedia of the Social & Behavioral Sciences (pp. 568-573). Orlando: Elsevier. doi:978-0-08-097087-5

Angrist, J., & Pischke, J.-S. (2015). Mastering Metrics. The Path from Cause to Effect. Princeton; Oxford: Princeton University Press.

BCE. (2019). Price Index for the Consumer, Producer and Labour Market. Quito: Ecuador

Central Bank. Opgehaald van

https://www.bce.fin.ec/index.php/component/k2/item/315-indice-de-precios-al-consumidor-y-productor-salarios-empleo-y-mercado-laboral

Bertrand, M., Duflo, E., & Mullainathan, S. (2004). How Much Should We Trust Differences-in-Differences Estimates? Quarterly Journal of Economics, 119(1), 249-275. Opgehaald van https://www-jstor-org.proxy.uba.uva.nl:2443/stable/25098683

Borjas, G. J. (2016). Labour Economics (7 ed.). New York: McGraw-Hill Education.

Chambers , B. W., & Sampson, G. P. (2008). Developing Countries and the WTO : Policy Approaches. Tokyo: United Nations University Press.

Department of Economic and Social Affairs. (2008). International. New York: United Nations. doi:978-92-1-161518-0

Edwards, S. (1988). Terms of Trade, Tariffs, and Labor Market Adjustment in Developing Countries. The World Bank Economic Review, 2(2), 165-185.

Evenett, S. J. (2019). Protectionism, state discrimination, and international business since the onset of the Global Financial Crisis. Journal of International Business Policy.

Evenett, S. J., & Fritz, J. (2020). The Global Trade Alert Database Handbook. Manuscript. Ferrara, M. H. (2013). Gale Business Insights Handbook of Global Business Law. Gale:

(43)

Gaston, N., & Trefler, D. (1994). Protection, Trade, and Wages: Evidence from U.S. Manufacturing. ILR, 47(4), 575-593. Opgeroepen op 03 30, 2020, van Stable URL: https://www.jstor.org/stable/2524659

Giavazzi, F., & Tabellini, G. (2005). Economic and political liberalizations. Journal of Monetary Economics. Journal of Monetary Economics, 52(7), 1297-1330.

Goldberg, P. K., & Pavcnik, N. (2005). Trade, wages, and the political economy of trade protection: evidence from the Colombian trade reforms. Journal of International Economics, 66, 75-105.

Goldberg, P. K., & Pavcnik, N. (2007). Distributional effects of globalization in developing countries. Journal of Economic Literature, 45(1), 39-82.

Goode, W. (2003). Dictionary of Trade Policy Terms (Fourth Edition ed.). Cambridge

University Press. Opgehaald van

http://ctrc.sice.oas.org/trc/WTO/Documents/Dictionary%20of%20trade%20%20policy %20terms.pdf

Goode, W., Adelaide, U. o., & WTO. (2007). Dictionary of Trade Policy Terms. Cambrige University Press. Opgehaald van https://web-b-ebscohost- com.proxy.uba.uva.nl:2443/ehost/detail/detail?vid=0&sid=22f279fb-fb83-4158-aa00-

4cb74dbe7da4%40pdc-v-sessmgr05&bdata=JnNpdGU9ZWhvc3QtbGl2ZSZzY29wZT1zaXRl#AN=335152&d b=e000tww

Government of the Republic of Ecuador. (2013, 02). National Plan for the Buen Vivir 2009-2013. Opgehaald van Technical Secretariat : Plan Ecuador: https://www.planificacion.gob.ec/plan-nacional-para-el-buen-vivir-2009-2013/

The Effect of Trade Barriers on Sectorial Wages in Ecuador Between 2007 and 2019: Is Protectionism a Curse Word?