BACHELOR’S THESIS
THE EFFECT OF INCREASING MINIMUM WAGES ON
UNEMPLOYMENT RATES:
EVIDENCE FROM UNEMPLOYMENT TRENDS IN SOUTH KOREA
FROM 1989 TO 2016
Written by
YeJin Lee
10706321
Supervised by
Péter Foldvari
Abstract
The minimum wage system in South Korea was first implemented in 1989. An increase in minimum wage has been deemed having positive effects on decreasing unemployment rates, and this can be seen from the constant increases in minimum wages since its introduction in 1989. However, the effect of minimum wage on unemployment rates has been debated for the recent few decades and a concrete conclusion regarding this debate is yet to exist. In order to empirically investigate the effect of increasing minimum wages on unemployment rates in South Korea, this paper employed a panel regression model with aids of a fixed effect estimator and a first difference estimator in the analysis. The empirical results demonstrate two conflicting outcomes: when the fixed effect estimator is used, minimum wage has a positive relationship with unemployment rates, and when the first difference estimator is used, minimum wage has a negative relationship on unemployment rates.
Statement of Originality
This document is written by Student [YeJin Lee] 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……….……….….5
2 Literature Review……….………...6
2.1 Theoretical Framework………....6
2.1.1 Competitive Labor Market Model……….6
2.1.2 Monopsony Model………....8
2.2 Empirical Studies………...12
2.2.1 Trends in the Previous Studies………12
2.2.2 Conflicts Between the Studies……….13
3 Methodology and Data………15
3.1 Methodology………..15 3.1.1 Regression Model………...15 3.1.2 Hypotheses Set Up ……….16 3.2 Panel Data………..16 3.2.1 Data Sources………...16 4 Results………..17 4.1 Descriptive Statistics……….17
4.2 Research Findings and Analysis………...18
4.2.1 Test Descriptions………18
4.2.2 Hypothesis Outcomes………19
5 Conclusion and Limitations………..21
1 Introduction
The new government of South Korea has implemented a progressive law of increasing minimum wage by 16.3% from the national minimum wage in 2017 (Minimum Wage Commission, 2018). This is so far the highest hike the country has reached in 17 years, since 2001 (Minimum Wage Commission, 2018). Moreover, the new government has proposed to support a continuous escalation of the minimum wage by average of 15.7% per year, so as to attain the minimum wage of 10,000 won by 2020 from the current minimum wage of 7,530 won (Minimum Wage Commission, 2018). The minimum wage law was first implemented in South Korea in 1988 (Minimum Wage Commission, 2018). Thenceforth, it has been
continuously on its upturn (Minimum Wage Commission, 2018). The increment of minimum wage in South Korea seems to be a positive sign for its economy, and this must be one of the reasons for the government’s audacious movement (Jeong et al, 2011). However, this cannot be taken as granted, because the effect of minimum wage on economy has been very
controversial (Neumark & Wascher, 2007). There exist multitudinous studies about the relationship between minimum wages and unemployment rate (Card & Krueger, 1994; Neumark & Wascher, 2000; and Brown et al, 1982). In addition to the existing studies, unemployment rates will perform in this paper as the main indicator to study the economic effect of minimum wage in South Korea. Hence, the paper will be directed to answer the following: How has consistent increase in minimum wage affected unemployment rates in South Korea for the past 28 years?
Four of the factors that affect rate of unemployment, namely minimum wage, GDP per capita, inflation rate and size of labor force, were identified to be analyzed on their effect on unemployment rate in South Korea over the period of 28 years, with the help of fixed effect and first difference estimators in panel data regression.
The remainder of this research paper is structured as followed. Section 2, under literature review, discusses the theoretical framework of this paper, about the two important economic models to be considered for studies on labor markets. Kicking off with competitive labor market model, or also known as simple supply and demand model, it then moves onto monopsony model with only one buyer in the market. In the second part of literature review, this paper studies prior literature on the relationship between unemployment rates and minimum wages. Section 3 consists of research design which consists of descriptions of the
model that is employed for this paper and data collection that is used to estimate the
relationship between unemployment rates and rising minimum wage. Section 4 presents the results of the tests, and makes an analysis of each outcome. Last but not least, section 5 provides a conclusion regarding the relationship between unemployment rates and changes in minimum wages in South Korea for the period of 28 years, limitations of this paper and suggestions for possible further research.
2 Literature Review
2.1 Theoretical Framework
2.1.1 Competitive labor market model (Simple supply and demand model)
The first model to be considered is simple supply and demand model of labor, which focuses on a competitive labor market that consists of homogeneous workers (Brown et al, 1982). There are a few assumptions in this market: first, employers minimize their costs regardless of the existence of minimum wage law (Brown et al, 1982). Secondly, workers have identical skills and effort, and they are given exogenously (Brown et al, 1982). Lastly, workers are protected by minimum wage from receiving an unjustifiable amount of wage from their employers (Brown et al, 1982). In the simple supply and demand model in figure 2.1, the
initial employment Ea is determined at the equilibrium point between the labor demand and
labor supply curves (Brown et al, 1982). When a minimum wage higher than the initial wage
Wa is introduced, employment level decreases to Eb, and the demand will decrease by Eb-Ea
and the supply will increase by the amount of Sb-Ea (Brown et al, 1982). This leads to
employment level of Eb and unemployment level of Sb-Eb at the wage level of Wb (Brown et
Figure 2.1. Simple Supply and Demand Model of Labor Market
Figure 2.1 Simple supply and demand model of labor market. Adapted from “The Effect of
Minimum Wage on Employment and Unemployment,” by C. Brown, C. Gilroy, and A.
Kohen, 1982, Journal of Economic Literature, 20(2), p. 488. Copyright 1982 by American
Economic Association.
The figure 2.2 below describes the relationship between a minimum wage and an employment level in a competitive labor market (Jeong et al, 2011).
Figure 2.2. Relationship Between Minimum Wage and Employment in a Competitive
Labor Market
Figure 2.2. Relationship between a minimum wage and an employment in a competitive
labor market. Adapted from “Analysis of the Effect of Minimum wage” by J. Jeong, J. Nam, J. Kim, and Y. Jun, 2011, p. 26. Copyright 2011 by Korea Labor Institute.
As it can be shown in the figure 2.2, implying a minimum wage at a lower level than
Wa will not have any effect on employment level (Jeong et al, 2011). However, this is a
meaningless action, because equilibrium point is then equal to the one in a situation without floor price system (Jeong et al, 2011). However, if minimum wage is set above the
equilibrium wage level at Wa, there is an excess supply of labor. As level of employment
depends on the demand of labor, the wage level above equilibrium leads to drop in labor demand, hence lower employment level (Jeong et al, 2011).
The decrease in an employment level is an outcome of two different effects, scale effects and substitution effects (Jeong et al, 2011). First of all, the scale effects occur when there is a rise in the price of goods due to higher wages (Jeong et al, 2011). In other words, along with the increase in the minimum wages that an employer has to pay their employees, higher prices will be charged for the goods that firms produce and sell (Jeong et al, 2011). Consequently, there will be a drop in the demand of the goods with higher prices, which will then lead to the decrease in production of the goods (Jeong et al, 2011). As the level of production decreases, there are less inputs to be used for the production (Jeong et al, 2011). Labor is considered one of the inputs to be used for the production of most goods, so if there is a decrease in input usage, it also refers to a decrease in labor requirement, and hence the employment level (Jeong et al, 2011). Secondly, considering all factors of production including labor, if the price of labor increases, there will be a substitution effect where the other factors of production with relatively lower prices substitute labor in production of goods (Slaughter, 2001). This is another effect that results in a decrease in labor demand (Slaughter, 2011).
The size of the effect of implying a minimum wage depends largely on the elasticity of labor demand of a firm (Jeong et al, 2011). This is because labor demand is negatively related to the price elasticity of demand for related goods, the fraction of labor in the
production cost, and substitutability between labor and other factors of the production (Jeong et al, 2011).
2.1.2 Monopsony model
Monopsony refers to a situation where there is only one buyer in a market (Brown et al, 1982). In a labor market, the buyer would be a firm, and suppliers would be workers who provide labor (Brown et al, 1982). Monopsony model in a labor market shows that if
minimum wage is set at a too high level, there will be a decrease in an employment level, just as in the competitive labor market (Brown et al, 1982). However, if there is an appropriate level of adjustment to a minimum wage, it is plausible that the minimum wage higher than at equilibrium level is applied with a result of a higher employment rate (Jeong et al, 2011). If the firm in the labor market aims to maximize its profit, the profit function is as following:
otal Revenue otal Cost
π = T − T
(L) (L)L
= f − w
aximization with respect to L f (L) (L) (L)L
M : ′ − w − w′ = 0
f (L)′ = w′(L)L+ w(L)
From this equation, f (L) refers to marginal revenue product (MRP), while the terms on the′
right, w (L)L + w(L) refers to marginal labor cost (MC) of the firm (Jeong et al, 2011). The′
equation can then be expressed in terms of elasticity of labor supply ∊:
(L) (1 )
f′ = w + ∊1
In a monopsony labor market as in figure 2.3, production level of the firm is determined at the intersection between MRP and MC, and here the employment level is also determined. Wage is set at a corresponding level of an employment level and the labor supply curve (Boal & Ransom, 1997).
Figure 2.3. Monopsony Labor Market
Figure 2.3. Monopsony Labor Market. Adapted from “Monopsony in the Labor Market,” by
W. M. Boan, and M. R. Ranson, 1997, Journal of Economic Literature, 35(1), p. 88.
Copyright 1997 by American Economic Association.
The introduction of a minimum wage in monopsonistic market has different effects from the competitive market on an employment level (Jeong et al, 2011). When the minimum wage introduced is at a lower level than the monopsony wage, this will not have any new effect on the employment level (Boal & Ransom, 1997). However, if the minimum wage rises above the monopsony wage, the upward sloping labor supply curve will partially become horizontal (Boal & Ransom, 1997). Also, the originally upward sloping MC curve also becomes horizontal until it meets the labor supply curve, where it starts to slope upwards again, hence creating a kink (Boal & Ransom, 1997).
As shown from the figure 2.3, if the minimum wage is settled between the monopsony wage and the equilibrium wage in the perfectly competitive labor market, the employment level will move along the labor supply curve, leading to an increase in the employment level (Jeong et al, 2011). However, if the minimum wage introduced exceeds the competitive wage, the employment level will retrogress along the MRP and therefore result in a drop in the employment rate (Jeong et al, 2011). Figure 2.4 shows a simpler movement of changes in the employment level along with the changes in wages.
Figure 2.4. Movement of employment changes along with the changes in wages.
Figure 2.4. Movement of changes in the employment level along with the changes in wages.
Adapted from “Analysis of the Effect of Minimum wage” by J. Jeong, J. Nam, J. Kim, and Y. Jun, 2011, p. 28. Copyright 2011 by Korea Labor Institute.
The range of a minimum wage that does not bring along a drop in an employment rate depends on elasticity of labor supply, and the higher the elasticity, the smaller the range is (Nam et al, 2009, p. 41). Within this range, the elasticity of labor supply has two
interpretations (Nam et al, 2009, p. 41). Firstly, one unit increase in the minimum wage has higher effects on a rise in the employment rate under a higher labor supply elasticity (Nam et al, 2009, p. 41). This is because a less elastic labor supply has a less reaction on wage rise from increased minimum wage (Nam et al, 2009, p. 41). Secondly, a more elastic labor supply has an effect on narrowing down the range that is mentioned at the beginning of this paragraph (Nam et al, 2009, p. 41). The reason is that the monopolistic strength of firms shrinks with the labor supply elasticity, and hence the ability of the firms to discount on wages also declines (Nam et al, 2009, p. 41).
The problem with a monopsony model is whether a low wage labor market, which is expected to be highly affected by minimum wages, can still be regarded as a monopsonistic labor market in reality (Manning, 2003, p.18). Manning (2003) pointed out that even if a firm is a monopsonist, taking into account an interplay and a variety of different firms to model the entire labor market, the assumption of monopsony does not maintain. However, the term monopsony can be interpreted as a more general term when applied to the reality, where there
are multiple buyers (Nam et al, 2009, p.41). In this case, whether a labor market is monopsonistic depends on the wage setting power of individual firms that demand labor (Nam et al, 2009, p.41). In other words, if each firm is able to purchase any amount of labor for a given price, it can be seen as a perfectly competitive labor market (Nam et al, 2009, p.41). If the firms have to provide higher wages for extra labor, the labor market can be seen monopsonistic (Nam et al, 2009, p.41). An example of such situation is when the employer is not able to perfectly monitor the employees, providing higher wages for workers prevent possible laziness of them (Nam et al, 2009, p.41). Here, even with multiple buyers in the labor market, employers can have an upward sloping labor supply curve (Nam et al, 2009, p.41).
2.2 Empirical Studies
2.2.1 Trends in the Previous Studies
As can be conjectured from the dispute over the efficacy of minimum wages on the
employment level, there has been a transition in the flow of trends in the beginning of 1990s (Brown et al, 1982; and Card & Krueger, 1994). The majority of and also the winning arguments until the early 1980s contended that an increase in the minimum wage has a negative impact on the employment level (Brown et al, 1982). Brown, Gilroy and Kohen (1982) have found a negative correlation between the minimum wage and the employment rate, that a rise in the minimum wage actually brings down the employment rate.
On the contrary, the studies of 1990s exhibited contrasting outcomes from the ones in the 1980s (Card, 1992; Card & Krueger, 1994). The researchers in the 1990s apprehended the imperfection in the research model of the past, and to surmount these shortcomings of the 1980s, the newly set up model consisted mainly of empirical studies that comprise of cross-sectional or panel data (Nam et al, 2009, p.35). A number of renowned studies in this field targeted mainly of the changes in the employment rate due to the minimum wage adjustment, usually an increase in minimum wages, in specific sectors, such as fast food restaurants in Card & Krueger (1994). This is due to the fact that these usually weigh the most in teenager employment rates (Card & Krueger, 1994). The employment rates of teenagers are also a notable factor, because a large proportion of teenagers in the labor force work for the minimum wages (Card & Krueger, 1994). It had been revealed in the study of Card & Krueger (1994) that an increase in the minimum wage either has a positive effect on
the employment rates, or does not have any significant effect on them. Card (1992) had noted that the rise in the minimum wage in California from $3.35 to $4.45 in 1988 did not prove any negative effect on the employment rate (Card, 1992).
2.2.2 Conflicts Between the Studies
The above mentioned Card & Krueger (1994) still serves as one of the most persuasive studies that deals with the negative relationship between unemployment rates and a rise in minimum wages. The experimentation was based on two states in the United States,
Pennsylvania and New Jersey, and difference in differences method was adopted to carry out the experiment regarding the effect of a rise in the minimum wage on the employment rates (Card & Krueger, 1994). A group of fast food restaurants in eastern Pennsylvania had been chosen as a control group, as periodic patterns of the employment rates were analogous in New Jersey and eastern Pennsylvania (Card & Krueger, 1994). In 1992, the minimum wage in New Jersey had risen from $4.25 to $5.05, whereas in east Pennsylvania, which is located right next to New Jersey, the minimum wage stayed constant (Card & Krueger, 1994). This way, it was possible for them to arrive at a less biased result from analyzing the effect of increasing minimum wages on the employment rates during identical time periods (Card & Krueger, 1994). According to the telephone surveys targeted at fast-food restaurants that were located in between the two states, surprisingly it was displayed that the rise in employment rate in New Jersey overtook the one in east Pennsylvania (Card & Krueger, 1994). This refutes the opposing arguments regarding the effect of the increasing minimum wage on the unemployment rates (Brown et al, 1982).
Notwithstanding, this argument was again refuted by Neumark & Wascher (2000). They had performed a study based on the same chain of fast-food restaurants in the same geographic areas, New Jersey and east Pennsylvania, but they used a different research method (Neumark & Wascher, 2000). They had revealed that in the test results of Card & Krueger, standard deviations of changes in the employment rates were rather substantial, which could have been caused by a data distortion (Neumark & Wascher, 2000). Pointing out a few plausible problems in the data collected by Card & Krueger using telephone survey method, Neumark & Wascher (2000) persuaded that a new data and evaluation method were necessary for a more precise and correctly oriented test results. For their research, they had utilized administrative payroll data of the fast-food restaurants, and this demonstrated a
conflicting impact of rising minimum wages on employment rates compared to the observations of Card and Krueger (Neumark & Wascher, 2000). Neumark and Wascher (2000) had replicated the identical difference in differences evaluation method of Card and Krueger using the different set of data they had collected. The outcome was such that there was a 3.9 to 4 percent decrease in the employment rate in New Jersey after the rise in the minimum wage, as to the control group in Pennsylvania (Neumark & Wascher, 2000).
Nevertheless, this contention was again rebutted in a new study by Card and Krueger (2000), of which the administrative employment data were analyzed. Moreover, they had strengthened their stand by using the Bureau of Labor Statistics’ (BLS’s) data to investigate the employment status in fast-food restaurants in New Jersey and Pennsylvania (Card & Krueger, 2000). They assert that the reliability of their sample has been reinforced, as for the fact that the BLS sample contains the knowledge of every employers in a certain number of restaurant chains (Card & Krueger, 2000). The new test using the BLS sample demonstrated that the pace of employment growth in New Jersey after the rise in the minimum wage is higher than in Pennsylvania in a small scale (Card & Krueger, 2000). The BLS sample was also used to investigate long run effects of the minimum wage increase in New Jersey, and the effect of the federal minimum wage increment in 1996 which influenced the minimum wage in Pennsylvania but not in New Jersey, where the minimum wage had already
outstripped the federal minimum wage (Card & Krueger, 2000). This further analysis by Card and Krueger (2000) undermined the argument of Neumark and Wascher by disclosing the fact that a moderate incline of the minimum wage only has insignificant effects on the employment rate.
In relation to the social disputes and the unbounded interest about the effect of minimum wages on employment rates, this has also been one of the most desired content of studies in South Korea (Jeong et al, 2011, p.34). However, there has been high limitations in performing such researches, and many of them were highly dependent on studies done in major foreign countries, such as the United States (Jeong et al, 2011, p.34). This was
essentially due to the identical minimum wages across the country, which made it technically arduous to find the right variables for the test (Jeong et al, 2011, p.35). Despite this, recently there has been frequent endeavors to increase in the number of studies that analyze the effect of minimum wages on the employment rate, which nonetheless still does not seem to be of a large help to reach a coherent outcome (Jeong et al, 2011, p.35).
Not only in the the studies mentioned above, the dispute over the employment effect of increasing minimum wage is still an extensive debate. Although there have been countless studies on this issue, the researchers have still not been able to agree upon one corresponding conclusion (Card & Krueger, 1994; Neumark & Wascher, 2000; Brown et al, 1982; and Jeong et al, 2011, p.35).
3 Methodology and data
3.1 Methodology
3.1.1 Regression Model
In order to test the hypothesis between unemployment rates and minimum wages in South Korea over the period of 1989 and 2016, a panel data regression model consisting of several variables was set up. The regression model that has been used for the test is built as follows:
(Minimum wage) nβ (GDP per capita)
Unemployment ratei,t = β0+ lnβ1 t+ l 2 t+
(Inflation rate) (Labor force)
β3 t+ lnβ4 i,t+ εi,t
In the table 3.1, a summary of interpretations and measurements of each variables of the model above is provided.
Table 3.1. Dependent and Independent Variables
Variable Measurement
Dependent variable
Unemp Number of unemployed people as a percentage of total labor force
Explanatory variables
Minwage The national minimum wage of South Korea
Inflation Inflation rate measured by the annual percentage change in the cost to the average consumer of acquiring the same amount of goods and services
lnGDPpercap Log of GDP per capita of South Korea
lnLaborforce Log of changes in the number of labor force of South Korea
3.1.2 Hypotheses Set Up
In contemplation of right conclusion from this test, there is a need for good hypotheses regarding the effect of a constant increment in minimum wage on unemployment rates in South Korea during the period between 1989 and 2016. The hypotheses that had been constructed for the analysis are as followed:
H0: The constant increase of the national minimum wage has induced an escalation of the
unemployment rate in South Korea for the period of 28 years.
H1: The constant increase of the national minimum wage has not induced an escalation of the
unemployment rate in South Korea for the period of 28 years.
3.2 Panel Data
A panel data set, also known as a longitudinal data set, concerns a given sample of individuals across a time period, and therefore yields a collective observation of the
mentioned individuals (Hsiao, 2003, p.1). A panel data set has several advantages over time series data set or cross-sectional data set (Hsiao, 2003, p.3). One example is that using a panel data set gives an access to estimate with higher efficiency, because an extensive number of data points employed in the research allows for a higher degrees of freedom and lower collinearity between independent variables (Hsiao, 2003, p.3). In this research, a balanced panel data set is adopted, where ‘balanced’ refers to a status of a data set that does not incorporate any time gap within the given time period.
3.2.1 Data Sources
This paper analyzes the changes in unemployment rates in South Korea over the period of 28 years, from the year 1989 to the year 2016. The ‘Minimum Wages Law’ was first enacted and proclaimed on the 31st of December in 1986, and was implemented from 1st of January in 1988 (Minimum Wage Commission, 2018). However, the present system of single minimum wage across the country was only enforced from the year 1989, which is why the data
adopted in this test commences from 1989 (Minimum Wage Commission, 2018). The year 2016 is the latest year of which every data used for the test was available during the period of data collection, hence this research is based on the years from 1989 to 2016.
To formulate a panel data, the data containing unemployment rates of 16 different provinces of South Korea over the period of 28 years were collected. The similar method was
used to manipulate the data for the size of the labor force in South Korea, for which the log of size of labor force represents one of the independent variables in the regression model
employed in the test. Other independent variables used in the model are national based, hence they are not segregated by regions, but they also represent the time period of 28 years. The data for regional unemployment rates and regional sizes of labor force were collected from Statistics Korea (2017), for the years 1989-2016. The remaining independent variables used for the test are log of minimum wage, log of GDP per capita and inflation rates of South Korea, also for the identical time periods. Statistics Korea (2017) provided annual data for unemployment rates and sizes of labor force in the 16 provinces of South Korea. The data for minimum wages of each year were obtained from Minimum Wage Commission of South Korea (2017), and the data for GDP per capita and inflation rates over the whole country were obtained from World Bank (2017).
4 Results
4.1 Descriptive Statistics
The table 4.1 provides a summary of descriptive statistics of the sample used in the data. From the table, it is possible to confirm that the average unemployment rate in South Korea is 3.10%. Log was applied to the original minimum wage values in South Korea won before being utilized in the test. Started from the minimum wage of 600 won in 1989, South Korea, after 28 years, had reached the minimum wage of approximately ten times higher than the beginning. These values in log are 6.40 for minimum and 8.70 for maximum. It is also
interesting to know that the speed of inflation has been very volatile for the past 28 years. The highest inflation rate that the country has come across is 9.3%, and the lowest is only 0.706%. The log values of GDP per capita and the size of the labor force were manipulated to be used in the tests. For the period of 28 years, South Korea had seen an average log GDP per capita of 9.58. The log size of the labor force also has an increasing trend, of which the minimum and maximum values are 5.42 and 8.82 respectively.
Table 4.1. Descriptive Statistics for the years 1989-2016
Variable Obs Mean Std. Dev. Min Max
Unemp 448 3.1005 1.4278 0.4 9.1
lnMinwage 448 7.6705 0.6786 6.3969 8.7045
Inflation 448 3.8438 2.2297 0.7062 9.3
lnGDPpc 448 9.5787 0.4695 8.6547 10.2332
lnlaborforce 448 6.9763 0.7338 5.4161 8.8225
Source: the table is constructed by author based on the complete dataset
4.2 Research Findings and Analysis
4.2.1 Test descriptions
The main action taken in order for this research was to regress the three independent variables minimum wage, inflation rate and log of labor force against the dependent variable,
unemployment rate. To study and estimate the effect in the 16 provinces in South Korea across the period of 28 years, a panel data regression model was in the concern of this research. The Hausman specification test was employed in order to decide between random effects and fixed effects specification. In the fixed effect model, it is assumed that there is only a single valid effect and that every other discrepancies in the effects are due to sampling errors (Borenstein et al, 2010). In the random effect model on the other hand, the possibility of contrasting effect sizes are taken into account (Borenstein et al, 2010). It is crucial that the appropriate model is chosen, because not only the estimation of the data, but also the
interpretation of the results of the tests are affected (Borenstein et al, 2010). From this test on the unemployment rate in South Korea, the null hypothesis was rejected with a p-value of 0.9992 which is higher than 0.05. Accordingly, the random effect model was taken into account. The result of Hausman test is exhibited in table 4.2.1.
Table 4.2.1 Hausman test
Chi2 p value
0.08 0.9992
*p<0.10, **p<0.05, ***p<0.01
4.2.2 Hypothesis Outcomes
In this research, two estimators, fixed effects estimator and first difference estimators are employed in order to examine the effect of minimum wage increment on unemployment rates in South Korea.
A fixed effects estimator, also known as a within effects estimator, refers to an estimator for the coefficients in a regression model that contains fixed effects. In the case of random effect model, it is plausible that both fixed effect and random effect estimators are utilized. Yet, random effect estimator is more preferred in the situation where the individual specific effect meets the unrelatedness assumption. Unrelated assumption is translated as an assumption where the random-specific effect is a random variable that has no correlations with independent variables of all past, present and future time periods of the identical type. However, the unrelated assumption is an acutely powerful assumption and the fixed effect estimator can be a more reasonable estimator if this assumption is exquisitely met. Therefore in this test, the fixed effect estimator will be employed. Following equations illustrate the mechanism of the fixed effect estimator.
X
Yi,t = β0 + β1 i,t+ ξi+ ui,t
X Yi= 1 T ∑ T t=1Yi,t = β0+ β1 i+ ξi+ ui Y ) (X )
Ỹi,t = ( i,t− Yi = β1 i,t− Xi + ui,t− ui
On the panel data collected, a test was run using fixed effect estimators. Referring to this test, it is observed that the R-squared value is 0.4443, which indicates that the model explains 44.43% of the variability of the data around its mean. The coefficient of lnMinwage is 3.5280, and this reveals that 1% change in minimum wage changes the unemployment rate by 3.5280*0.01= 0.0353. This implies a positive relationship between the unemployment rate and the minimum wage. It is important to consider the p-value in order to decide whether or
and even 0.01, the H0 is rejected at all three significance levels. The output of the test is summarized and provided in the table 4.2.1, under columns labelled as (1).
In econometrics, first difference estimator is used as an approach to settle the issue of omitted variables (Wooldridge, 2002, p.279). By eliminating time invariant omitted variable
ξi through repeated observation over time, the first difference estimator averts bias which
could have been caused by ξi (Wooldridge, 2002, p.279). The corresponding equations which
demonstrate the effect of first difference estimator are as followed:
X
Yi,t = β0 + βi i,t+ ξi+ u1,t
X
Yi,t−1 = β0+ βi i,t−1+ ξi+ ui,t−1
Y X u
Δ i,t = Yi,t− Yi,t−1= β1· Δ i,t+ Δ i,t
A test was run on stata using first difference estimators. Based on the test run on the sample data, it is important to note that the R-squared value is 0.4605, and when converted to percentage, 46.05%. This indicates that 46.05% of the dependent variable, unemployment rate, has been explained by the four independent variables listed: lnMinwage, inflationrate, lnlaborforce and lnGDPpc. The coefficient of lnMinwage is found to be -0.5938. The interpretation for this finding is that 1% change in minimum wage changes the
unemployment rate by -0.5938*0.01=-0.0059. This indicates that unemployment rates and log of minimum wage have negative relationship, where a rise in minimum wage will lead to a fall in the unemployment rate. However, in order to test the validity of the hypothesis,
p-value has to be considered. lnMinwage has a p-value higher than 0.1, hence the H0 built for
this analysis is not rejected. The information of the output of the test is provided in the table 4.2.1, under columns labelled as (2).
Table 4.2.1. Fixed effect estimator and first difference estimator output summary
Independent variable
Coefficients (1) Std Error (1) Coefficients (2) Robust Std.
Error (2) lnMinwage 3.5280*** .2354 -.5938 1.2028 inflationrate -.2167*** .0282 -.0154 .0196 lnlaborforce .5690 .4152 -4.1709*** 1.2482 lnGDPpc -5.7668*** .3380 -4.7043*** .4934 constant 28.1284*** 3.0385 .4136 .1241 R2 0.4443 0.4605
Note: *significant at 10% level; **significant at 5% level; ***significant at 1% level.
Source: the table is constructed by author based on the complete dataset
5 Conclusion and Limitation
T
his paper is aimed at analyzing the effect of increasing minimum wages on unemploymentrates, specifically in South Korea, for the period of 28 years using panel data regression. To formulate the panel data, the unemployment data of 16 provinces of South Korea had been collected. Besides minimum wages, the data for inflation rates, size of labor forces and GDP per capita were employed to build a regression model.
Consistent with the prior research and the current debates about the relationship between the minimum wage and the unemployment rate, the two test results based on the fixed effect estimator and the first difference estimator illustrate different results. According to the output from the fixed effect estimator, a rise in the minimum wage leads to an increase in the unemployment rate, and this effect is significant. From the output of the first difference estimator, minimum wages and unemployment rates have a negative relationship, which indicates that an increase in the minimum wage leads to a lower unemployment rate.
However, this effect is not significant. Due to the insignificance of the outcome from the first difference estimator, the overall conclusion leans towards the positive relationship between the minimum wage and the unemployment rate, which explains that increasing minimum wage does not help in lowering the unemployment rates.
Not surprisingly, there is a number of limitations in this. To begin with the panel regression model that has been used for the analysis, only four factors had been employed as dependent variables. This could lead to a omitted variables error, because in reality there are countless number of variables that affect the unemployment rate in South Korea. In order to minimize the error regarding this matter, one solution could be taking into account as many variables as possible in the setup of the panel regression model. However this too, is not the perfect clarification, because in reality, there are countless variables of whose data cannot be collected that have influences on the unemployment rate in South Korea. Moreover, in time series data, it is inevitable that the outcomes of previous period affects those of the next. In the usage of the panel data regression in this research, it was assumed that each period is independent and products of previous periods have no effect on the next period. Therefore, an error regarding this matter is foreseen. To resolve this problem, it is wise to consider dynamic panel model.
6 References
Boal, W. M., & Ranson, M. R. (1997). Monopsony in the Labor Market. Journal of
Economic Literature, 35(1), 86-112.
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2010). A Basic Introduction to Fixed-Effect and Random-Effects Models for Meta-Analysis. Research Synthesis Methods, 1, 97-111.
Brown, C., Gilroy, C., & Kohen, A. (1982). The Effect of the Minimum Wage on
Employment and Unemployment: A Survey. Journal of Economic Literature, 20(2),
487-528.
Card, D. (1992). Using Regional Variation in Wages to Measure the Effects of the Federal
Minimum Wage. Industrial and Labor Relations Review, 46(1), 22-37.
Card, D., & Krueger, A. B. (1994). Minimum Wages and Employment: A Case Study of the
Fast-Food Industry in New Jersey and Pennsylvania. The American Economic Review,
84(4), 772-793.
Card, D., & Krueger, A. B. (2000). Minimum Wages and Employment: A Case Study of the
Fast-Food Industry in New Jersey and Pennsylvania: Reply. The American Economic
Review, 90(5), 1397-1427.
Hsiao, C. (2003). Analysis of Panel Data (2nd ed.). Cambridge: Cambridge University Press.
Jeong, J., Nam, J., Kim, J., & Jun, Y. (2011). Analysis of the Effect of Minimum wage. Seoul:
Korea Labor Institute.
Manning, A. (2003). Monopsony in Motion: Imperfect Competition in Labor Markets.
Princeton: Princeton University Press.
Minimum Wage Commission (2018). Current Minimum Wage. Retrieved from
http://www.minimumwage.go.kr/stat/statMiniStat.jsp
Nam, J., Ahn, T., Ahn, J., & Jeon, Y. (2009). Study of Labor Poor Solutions. Seoul: Korea
Labor Institute.
Neumark, D., & Wascher, W. (2000). Minimum Wages and Employment: A Case Study of
the Fast-Food Industry in New Jersey and Pennsylvania: Comment. The American
Economic Review, 90(5), 1362-1402.
Neumark, D., & Wascher, W. L. (2007). Minimum Wages and Employment. Foundations and Trends in Microeconomics, 3(1), 1-182.
Slaughter, M. J. (2001). International Trade and Labor-Demand Elasticities. Journal of
International Economics, 54, 27-56.
Statistics Korea, Survey Statistics (2014). Regional Unemployment Rate 1989-2014 [Data
file]. Retrieved from http://kosis.kr/statHtml/statHtml.do?orgId=101&tblId=DT_1D0 704&conn_path=I2
Statistics Korea, Survey Statistics (2017). Regional Unemployment Rate 1999-2017 [Data
file]. Retrieved from http://kosis.kr/statHtml/statHtml.do?orgId=101&tblId=DT_1D A7104&conn_path=I2
Statistics Korea, Survey Statistics (2014). Regional Labor Force Size 1989-2014 [Data file]. Retrieved from http://kosis.kr/statHtml/statHtml.do?orgId=101&tblId=DT_1D070
04&conn_path=I2
Statistics Korea, Survey Statistics (2017). Regional Labor Force Size 1999-2017 [Data file]. Retrieved from http://kosis.kr/statHtml/statHtml.do?orgId=101&tblId=DT_1DA70 04&conn_path=I2
The World Bank, World Development Indicators (2017). GDP per capita, current US$ [Data
file]. Retrieved from https://data.worldbank.org/indicator/NY.GDP.PCAP.CD
The World Bank, World Development Indicators (2017). Inflation, consumer prices (annual
%) [Data file]. Retrieved from https://data.worldbank.org/indicator/FP.CPI.TOTL.ZG
Wooldridge, J. M. (2002). Econometric Analysis of Cross Section and Panel Data.
