Professional Documents
Culture Documents
SSRN Id2782708
SSRN Id2782708
SSRN Id2782708
Abstract
Due to an aging population and falling birthrates, the labor force in Japan will continue to
decline in the next decades, and per capita income is also likely to decrease. To address this problem,
the Japanese government is considering relaxing restrictions on foreign labor, and there have been
debates regarding accepting foreign workers in Japan. Current Japanese law imposes strong
restrictions on unskilled foreign labor; thus, relaxing these restrictions will lead to an inflow of
unskilled foreign workers into Japan. However, few studies have analyzed the possible quantitative
effects of such measures.
Using a simulation based on a computable general equilibrium model, this paper analyzes the
quantitative impacts of the foreign labor inflow on the Japanese economy. Our model is a recursive
dynamic model from 2010 to 2030 with 32 sectors. We categorize labor into the following two types:
skilled and unskilled. In the main scenario, we assume an inflow of 5.32 million unskilled foreign
workers, which is equal to half of the labor force decrease in Japan.
The results of the main scenario are summarized as follows. First, foreign labor inflow lowers
the wage rate of unskilled labor but raises that of skilled labor and the rental price of capital. Second,
the inflow increases the GDP, income, and consumption, but the increases in income and
consumption are very small compared to the increase in GDP. Third, the inflow has very different
impacts on the different sectors. Fourth, we show that a policy limiting the sectors that accept
foreign labor does not necessarily increase the output of these sectors. Finally, we show that if we
accept not only unskilled but also skilled labor, the impacts on wage rates change significantly.
Keywords: foreign labor inflow, immigration, Japan, computable general equilibrium analysis
Email: saitou@narapu.ac.jp. We would like to acknowledge Kenji Kondo’s (Chukyo University) and Kazushi
Ugami’s (Kobe University) helpful comments and suggestions. We also thank the participants of the JSIE and JEA
annual meetings for fruitful discussions.
Recently, there has been debate regarding whether to accept foreign workers in Japan. This
is because the labor force has been decreasing due to low fertility and a greater number of
retired workers, while the recent economic recovery has increased labor demand. However, this
discussion is not new. In the 1980s, labor demand had increased due to the economic bubble,
and the government implicitly allowed unskilled foreign workers in the construction industry
and others. However, in the 1990s, the bubble burst, and public opinion was against foreign
labor inflow as it reduced the employment opportunities of Japanese workers. Due to
globalization and regional economic integration in the 2000s, this issue was raised in EPA
negotiations once more. The acceptance of foreign workers has often been a primary policy
issue in Japan within the past few decades.
Additionally, from the viewpoints of fiscal deficit and low fertility, the acceptance of
foreign workers has recently attracted attention. As such, Table 1 shows a population projection
estimate by the National Institute of Population and Social Security Research (NIPSSR) in
Japan. In this estimate, the labor force is expected to decrease by 10.19% from 2010 to 2020,
7.74% from 2020 to 2030, and 14.56% from 2030 to 2040. It also predicts that the ratio of the
elderly in the total population will rise from 25.1% in 2013 to 33.4% in 2035. As these estimates
show, Japan has been facing rapid aging because of low birthrates. However, the budget deficit
continues to exist in Japan, and government debt has already exceeded 1,000 trillion yen. It can
be argued that to address these problems, Japan should accept foreign workers. The Japanese
government takes a positive stand regarding the acceptance of highly skilled workers in these
discussions, but there are pros and cons according to public opinion. For these discussions to
yield results, detailed information on the impact of foreign labor acceptance in Japan is
necessary. This issue has been analyzed in numerous countries, and there is theoretical and
empirical literature on immigration or emigration. For example, Berry and Soligo (1969) show
that immigration increases the welfare of natives in the host country in theory. Subsequently,
numerous studies analyzed the effects of immigration on a host country using various models.
In contrast, Borjas (1995) empirically estimates that the economic benefits of immigration in the
United States reach around $7~25 billion per year.1
Nakamura et al. (2009) empirically analyze the effects of immigration in Japan. Using
econometric methods, their research examines the impacts of immigration on Japanese labor
wages, labor allocation across industries and industrial structure, among others. Although their
research offers important suggestions regarding the influence of foreign labor acceptance in
1
For recent empirical research on immigration, see Kerr and Kerr (2010).
2. Data
CGE analysis assumes that the economy is at equilibrium under the benchmark data. For
the benchmark data, we use the 2010 IO table of the JIP Database 2013 (Japan Industrial
Productivity Database 2013) provided by the Research Institute of Economy, Trade and Industry
(RIETI). 5 Although the original IO table of the JIP Database 2013 has 108 sectors, we
aggregate it into 32 sectors as shown in Table 2.6 We assume that each sector produces one
good. However, the “MIN” sector produces multiple goods due to by-products obtained.
Furthermore, as primary factors, we assume labor (skilled and unskilled) and capital.7 We
use the “Compensation of Employees” data from the IO table as the payment data for labor in
4
A multi-region model requires a global dataset. However, including multiple regions reduces the accuracy of the
data for individual countries.
5
http://www.rieti.go.jp/jp/database/JIP2013/
6
The correspondence table of the integrated 32 sectors to the original 108 sectors is available on request.
7
We consider the specific factor (resource) in the “MIN” sector.
In the analysis of foreign labor inflow, we divide labor into “skilled workers” and
“unskilled workers.” Of course, classifying labor into only two types is not ideal because of its
diversity in many different aspects (e.g., gender, habitation, education level, qualification/ability,
age). However, it is difficult to consider a wide variety of labor categories in a general
equilibrium model. Thus, we divide labor into only the two types.
To consider two types of labor, we need to divide the payments for labor, the number of
laborers, and wage rates in each sector according to the types of labor. In this paper, we divide
skilled and unskilled labor from the viewpoint of job categories. The JIP data provide
information regarding the number of workers according to the following seven job categories:
“Technicians and associate professionals,” “Officials and managers,” “Clerks,” “Sales workers,”
“Services workers,” “Plant and machine operators, and assemblers,” and “Others.” We regard
labor in the first two categories as skilled and labor in other categories as unskilled. The same
classification is adopted by, for example, the GTAP data and Kiyota (2014).
In Japan, there are already foreign workers. Therefore, the labor data will contain not only
Japanese but also foreign workers. Ideally, it is better to distinguish Japanese workers and
foreign workers who were already working in Japan. In simulating a general equilibrium model,
however, it is very difficult to divide workers in the existing data into Japanese and foreign
workers because of the difficulty in constructing the data and the model.8 Consequently, we do
not explicitly separate foreign workers who have already been in Japan (instead, we regard them
as Japanese). Thus, we analyze only the newly accepted inflow of foreign labor.
3. The Model
The model is a multi-sector recursive dynamic model of Japan over 20 years, from 2010
(the benchmark year) to 20309. All markets are considered perfectly competitive, and all
economic agents (firms, households, and governments) act as price takers. Firms maximize their
profits by producing goods using intermediate inputs and primary factors. We assume a
8
For instance, we need to divide sectoral data into Japanese and foreign labor.
9
The appendix provides an algebraic representation of the model.
First, regarding the behavior of firms, they have a “constant returns to scale” technology
with regard to deciding the amount of input and output to maximize profit. Figure 1 shows the
production function, which is a multi-stage CES function. Variables, such as E_KL and E_LL,
represent the elasticity of substitution. Firms use intermediate inputs and primary factors
(capital stock and labor). As previously explained in section 2, we divide labor into skilled and
unskilled.
In the production function in Figure 1, “skilled labor” and “unskilled labor” are first
aggregated into a labor composite, and the labor composite and capital are aggregated into a
primary factor composite. Finally, the output is determined using a fixed coefficient aggregation
of the primary factor composite and other intermediate inputs. When we analyze the impacts of
foreign labor inflow, substitution between “skilled” and “unskilled” labor and substitution
between labor and capital can play an important role. The value of the elasticity of substitution
is provided in section 3.6. Therefore, our model is a standard CGE model, except that labor is
divided into two types (our model is similar to, for example, the GTAP model).10 The produced
goods are allocated to foreign and domestic markets through a constant elasticity of
transformation (CET) function.
In a general equilibrium model, it is often assumed that the allocation of capital and labor
across sectors is determined such that factor prices are equalized in all sectors. In this paper,
however, we consider the differences in wage rates in different sectors that are observed in the
actual economy. Therefore, we assume that the production factors in each sector are imperfect
substitutes, and they are thus allocated through a CET function. This assumption limits the
mobility of primary factors across sectors. A sensitivity analysis of this assumption is performed
subsequently.
3.2. Household
To represent the demand side, we assume a representative household. The household earns
its income by providing production factors to firms and uses its income for consumption and
10
For details on the GTAP model, see Hertel (1999).
We assume that Japan is a small country (terms of trade are fixed). Similar to other CGE
analyses, we use the Armington assumption, which implies that domestically produced and
imported goods are imperfect substitutes. Domestic and imported goods are aggregated through
a CE function.
In this model, the current account is expressed as follows:
Current account = Trade balance – Remittance abroad
In other words, the current account is the trade balance minus remittances made by foreign
workers. If there is no remittance, the current account is equal to the trade balance. We assume
that the exchange rate (the price of foreign currencies) is determined such that the current
account is equal to the benchmark value.
The IO table of the JIP Database does not have tariff data (it has only the sum of imports
and tariffs). Goto (2010) indicates that the existence of tariffs can have an important role when
analyzing the effects of foreign labor inflows.12 Therefore, we introduce tariffs by separating
tariff data from imports that include the tariff. To do so, we use information of the GTAP
database.
3.4. Government
11
To devide the households according to the types of labor, we need to devide consumption of individual goods,
capital income and tax payments according to the types of labor. To do this division with accuracy, we need the
detailed data of consumption, income and tax data which are divided according to labor type. However, there are only
incomplete and insufficient datasets as of now.
12
The effect is called the “Uzawa effect” in Goto (2010).
This model is a recursive dynamic model used to solve a statics model repeatedly. Capital
stock is accumulated through investments made by the household and evolves according to the
following relation:
𝐾𝑡+1 = (1 − 𝛿)𝐾𝑡 + 𝐼𝑡 ,
where 𝐾𝑡 is the capital stock in period 𝑡, 𝐼𝑡 is investment in period 𝑡, and 𝛿 is the capital
depreciation rate. We adjust 𝛿 so that GDP in 2030 in the business as usual (BAU) scenario
with no foreign labor inflow is equal to GDP in 2010.
In the dynamic model, foreign labor inflow affects capital accumulation. That is, foreign
labor inflow increases production and thereby income. The increase in income leads to
investment and promotes capital stock accumulation. This generates further production
increases. To check the size of the capital accumulation effect, we consider the static model in
the sensitivity analysis.
There are various approaches to acceptance of foreign labor. For example, we can assume
13
E_N is the elasticity of substitution between a specific factor and other inputs in the MIN sector.
5. Policy Scenarios
We analyze the scenarios in Table 4. As the BAU scenario is a reference case in which
there is no foreign labor inflow, we first solve this scenario and derive the dynamic equilibrium
without foreign labor inflow. Subsequently, we solve the other scenarios with foreign labor
inflow and examine how the equilibrium changes from the BAU scenario.Scenario FL assumes
that Japan accepts unskilled foreign workers. As previously explained in section 4, the ongoing
debate about foreign labor relates to the acceptance of labor that does not require high skills and
abilities. Thus, we focus mainly on the inflow of unskilled labor (workers). Additionally, the
number of foreign workers that will flow into Japan is important, but there is no clear number
available. Therefore, we assume that the number of foreign workers flowing into Japan is equal
to half of the decrease in the Japanese labor force during 2010–2030, which is 5.32 million
(266,000 a year). According to the NIPSSR estimate, the working-age population decreases by
17.1% over the two decades from 2010 to 2030. This corresponds to a decrease of 10.62 million
in the labor force in our model. Thus, 5.32 million foreign laborers equal half of the
working-age population reduction. There are no definite criteria for accepting the number of
foreign workers; however, the “Choice for the Future” model developed by the Council on
Economic and Fiscal Policy of the Japanese government argues for the acceptance of 200,000
foreign workers a year, which is sufficiently similar to our scenario.15
The increase in foreign labor in scenario FL is introduced into this model as an increase in
the total labor supply, which is allocated across each sector endogenously (allocated according
to the CET function). However, there is the opinion that we should limit the sectors that accept
foreign workers to sectors with a labor shortage. FL_A and FL_B are scenarios that represent
the acceptance of foreign workers only in certain sectors. In FL_A, we assume that only “AGR,”
“CNS,” and “MED,” which are sectors with serious labor shortages, accept foreign labor.
However, FL_B is a scenario in which “CHM,” “PAC,” “ELC,” “TRN,” “NFM,” “PRE,” and
“GMA,” which are manufacturing sectors with high wage rates, accept foreign labor. In FL_A
and FL_B, foreign labor increases only in specific sectors, which does not necessarily mean that
14
From savings, capital income occurs. However, we assume there is no the capital income as it is difficult to
separate payment to a foreigner from capital income. This issue requires further study.
15
“Choice for the Future” Committee: http://www5.cao.go.jp/keizai-shimon/kaigi/special/future/
10
Before examining the simulation results, it is worth noting that the benchmark data reveal
the situation in Japan. Figure 3 shows wage rates in the benchmark year16. Wage rates are
normalized such that the wage rate of unskilled labor in the AGR sector is equal to unity. This
figure shows that there is a large difference in wage rates across sectors. While wage rates in
“PAC,” “EGW,” “PUB,” and “WAW” are high, the wage rate in “AGR” is extremely low.
Figure 4 represents the share (%) of payments to unskilled labor in production costs.17 If
the payment share to unskilled labor is high, the sector is regarded as unskilled labor-intensive.
The shares of unskilled labor payments in service sectors such as “COM,” “OPS,” “TRS,”
“BSE,” “WAW,” and “PUB” are relatively high. Additionally, the values for CNS and some
manufacturing sectors are high as well. According to the Rybczynski theorem, when the
endowment of a factor increases, the production of sectors using the factor intensively increases.
Therefore, we expect that the production of unskilled labor-intensive sectors to increase with
increases in unskilled foreign labor.
7. Simulation Results
In this section, we present the simulation results.18 In the following section, we focus on
the values of the variables in the last time period of the simulation (specifically, the year 2030).
Table 5 reports the values of various variables in 2030, and Table 6 reports the percentage
change from the BAU values in 2030. We explain the simulation results using graphs.
16
Appendix A-1 explains how these wage rates are calculated.
17
For example, it is approximately 20% in sector AGR. This means that the payment to unskilled labor accounts for
20% of the total cost in AGR.
18
We solve the model with GAMS (The General Algebraic Modeling System). All simulation results and the
program for the simulation are available from authors upon request.
11
Figure 5 shows income, tax payment, remittance, and consumption of foreign labor (all
values are deflated using a consumer price index). In scenario FL, unskilled foreign labor earns
an income of 24.7 trillion yen, pays income taxes of 2.5 trillion yen, remits 4.9 trillion yen, and
uses the rest for consumption and saving.
In scenario FL_A, only sectors AGR, CNS, and MED accept foreign labor. In this scenario,
the income of foreign labor is much smaller than that in the FL scenario. This is because wage
rates in the three sectors are low and the inflow of foreign labor lowers the wage rates of these
sectors significantly. The smaller income in FL_A leads to smaller tax payments and remittances
and less consumption. Similarly, in scenario FL_B, in which manufacturing sectors with high
wage rates accept foreign labor, the income of foreign labor is smaller than in FL.
Scenario FL_SK assumes that 20% of foreign labor is skilled labor. The number of
unskilled foreign labor decreases; therefore, the incomes, tax payments, and remittances of
unskilled foreign labor decrease as well. However, the consumption of total foreign labor
increases relative to FL because skilled labor receives higher income; thus total income of
foreign labor increases.
In section 7.1, we examined the impacts of foreign labor inflows specifically on foreign
labor. However, the impacts on Japan are more important to Japanese policy making. Figure 6 is
the percentage change in wage rates and rental prices (2030, %). As explained in Section 3,
wage rates are different in different sectors, and Figure 6 shows averages. Because foreign and
Japanese labor are perfect substitutes, wage rates for foreign labor and Japanese labor are the
same. Additionally, because this model assumes a fixed supply of Japanese labor, the changes in
wage rate and labor income for Japanese labor are the same.
In scenario FL, the wage rate for unskilled labor falls by 4.6% and the wage rate for skilled
labor and the rental price increase by 1.5% and 3.6%, respectively. That is, in this scenario, an
increase in foreign labor generates losses for unskilled Japanese labor and gains for skilled
Japanese labor and capital owners. The direction of impacts in scenario FL_A are the same as
those in FL. However, the size of the impacts becomes smaller. In FL_B, the negative impact on
the wage rate for unskilled labor shrinks and the positive impact on wage rate for skilled labor
expands in comparison to FL. Therefore, Japanese workers receive more favorable results in
scenario FL_B than in scenario FL.
In scenario FL_SK, the wage rate for skilled labor falls substantially as well. This is an
12
13
Next, we examine the impacts of foreign labor inflow on sectoral output. Table 7 reports
the percentage changes in output by sector (year 2030, %). Figure 8 depicts scenario FL, in
which the sector with the highest rate of increase in output is ELC (12.5% increase), and the
sector with the lowest rate is PUB (1.0%). This shows that rates of increase in output are
different across sectors.
In Figure 4, we see which sectors are relatively unskilled labor-intensive. Figure 8 shows
that the rates of increase in the outputs of unskilled labor-intensive sectors are not necessarily
large. For example, the outputs of ELC and NFM increases substantially; however, the share of
unskilled labor in these sectors is not high. The results reported in Figure 8 include not only the
direct effects of foreign labor inflow but also the indirect effects such as price changes.
Therefore, these results indicate that the indirect effects are strong.
Table 7 reports changes in outputs in the other scenarios. In scenario FL_A, only sectors
AGR, CNS, and MED accept foreign labor. Thus, it is expected that the output of these three
sectors will largely increase. However, the output changes of the three sectors in FL_A are only
slightly larger than those in FL. The foreign labor inflow into these sectors lowers their wage
rates; thus, Japanese labor, which is mobile across sectors, moves to other sectors. Therefore,
the output of these sectors does not increase much. Recently, there has been serious concerns
regarding labor shortages in sectors such as agriculture, construction, and nursing services, and
it has been argued that Japan should accept foreign labor in these sectors. However, our analysis
shows that such a policy will not necessarily have the expected results.
In contrast, in scenario FL_B, the output of sectors that accept foreign labor, with the
exception of CHM and PAC, increases significantly; this differs from scenario FL_A. This
shows that even limiting the sectors that accept foreign labor does not necessarily imply that the
output of these sectors will increase substantially.
In the previous section, we examined the impacts of foreign labor inflow from various
viewpoints. However, the results of the simulation can vary according to the initial assumptions.
To check for the robustness of simulation results, we conduct a sensitivity analysis. Table 8
describes the scenarios of the sensitivity analysis. The “benchmark” is the scenario analyzed in
14
15
16
In scenario “FLAB”, the size of the changes in the variables changes significantly. This
means that the impacts of foreign labor inflow depend strongly on the number of foreign
workers flowing into Japan. In all the scenarios, the wage rate of unskilled labor falls and the
rental price of capital rises. In addition, except for in one case, the wage rate of skilled labor
rises. These results indicate that an inflow of (unskilled) foreign labor is likely to generate losses
for unskilled Japanese labor and gains for skilled Japanese labor and capital owners. However,
the size of losses and gains depends strongly on the assumptions regarding the elasticity
parameters. Thus, if we want to analyze the impacts on factor prices, we should pay attention to
the elasticity parameters. In all the scenarios, GDP increases, and its rate of increase does not
change significantly (except for in the “FLAB” scenario). The impacts on the income and
consumption of the Japanese change slightly across scenarios. In particular, the tax rate on the
income of foreign labor affects the income of the Japanese.
8. Concluding Remarks
This paper quantitatively analyzes the effects of foreign labor inflow (mainly unskilled
labor) in Japan. We use a 32-sector CGE model with an analysis period from 2010 to 2030. The
main scenario (scenario FL) assumes that there is an unskilled foreign labor inflow of 5.32
million for 20 years (266,000 per year). In addition, we also consider a scenario that limits the
sectors that accept foreign labor and a scenario where not only unskilled but also skilled labor
come into Japan.
We examined impacts on factor prices, macroeconomic variables, and sectoral output. Our
results are summarized as follows. First, as a result of foreign labor inflow, the wage rate of
unskilled labor falls but the wage rate of skilled labor and the rental price of capital rise. This
means that foreign labor inflow is likely to be detrimental to unskilled Japanese workers and
beneficial to skilled Japanese workers and capital owners. In the main scenario (FL), the rate of
the decrease in the wage rate of unskilled workers is 4.6%, which is a non-negligible negative
impact. When we consider an inflow of skilled labor in addition to unskilled, skilled Japanese
workers also suffer a large loss.
Second, in the main scenario, GDP increases by 6.5%. This means that the foreign labor
inflow assumed in scenario FL has very large positive impacts on GDP. Because of the GDP
increase, Japanese income and consumption increase, too. However, the rates of increase in
17
18
References
19
20
Table 1:Estimate and transition of the future labor force in Japan (medium mortality variant
projection)
2010 2020 2030 2040
Low fertility variant 81,735 63.8 73,408 59.7 67,328 59.0 56,056 54.1
Medium fertility variant 81,735 63.8 73,408 59.2 67,730 58.1 57,866 53.9
High fertility variant 81,735 63.8 73,408 58.6 68,073 57.1 59,770 53.7
Data source: NIPSSR, “Population Projection for Japan: 2010–2060 (January 2012)”
21
Table 4: Scenarios
Scenarios Explanation
BAU Reference equilibrium (no foreign labor inflow)
FL Increase in unskilled labor
FL_A Increase in unskilled labor only in AGR, CNS, MED sectors.
FL_B Increase in unskilled labor only in CHM, PAC, ELC, TRN, NFM, PRE, GMA.
FL_SK Increase in unskilled and skilled labor (20% is skilled labor).
22
23
24
All the values except the trade surplus and current account values indicate the percentage change from BAU values (2030, %).
Trade surplus and current account are level values (2030, trillion yen).
Goods
Production
Leontief
Intermediate inputs KL
E_KL
Capital stock
E_LL
Utility
E_CC
26
14.0
12.0
10.0
8.0
6.0
4.0
2.0
0.0
PPW
PUB
CHM
MET
AGR
EGW
COM
EDR
MED
MIN
PAC
NFM
ICE
FAI
PSE
OTH
FOO
TEX
GMA
WAW
BSE
IAS
ELE
ELC
TRN
TRS
CSC
CAB
PRE
CNS
RES
OPS
Skilled Unskilled
45
40
35
30
25
20
15
10
0
PPW
EDR
AGR
CHM
COM
MIN
NFM
MET
ICE
PAC
FAI
MED
EGW
PUB
PSE
FOO
WAW
TEX
CSC
IAS
GMA
ELE
ELC
OTH
CAB
BSE
TRN
RES
TRS
PRE
CNS
OPS
27
Income of foreign
unskilled workers
Remittance by foregin
unskilled workers
Consumption of
all foreing workers
6.00
4.00
2.00
0.00
FL FL_A FL_B FL_SK
-2.00
-4.00
-6.00
Wage rate for unskilled workers Wage rate for skilled workers Rental price
28
8.00
7.00
6.00
5.00
4.00
3.00
2.00
1.00
0.00
FL FL_A FL_B FL_SK
14.0
12.5
12.0 11.3 11.6
10.9
10.5 10.4
9.9 9.9 9.6
10.0 9.1
8.4 8.6 8.4 8.7
8.3 8.4 8.3 8.2 8.3
8.0 7.3 7.3 7.4
7.0
6.76.7 6.7
6.2
5.7
6.0
4.6
4.0
2.42.4
2.0
1.0
0.0
PPW
EDR
AGR
CHM
COM
MIN
NFM
MET
ICE
PAC
FAI
MED
EGW
PUB
PSE
FOO
ELC
WAW
OTH
TEX
CSC
IAS
GMA
ELE
CAB
BSE
TRN
RES
TRS
PRE
CNS
OPS
29
0.00
2.00
3.00
4.00
6.00
7.00
0.00
1.00
5.00
Benchmark Benchmark
FLAB (more) FLAB (more)
FLAB (less) FLAB (less)
Static Static
Variable CA Variable CA
EOS_KL (large) EOS_KL (large)
EOS_KL (small) EOS_KL (small)
30
EOS_LL (large) EOS_LL (large)
EOS_LL (small) EOS_LL (small)
EOT_LA (large) EOT_LA (large)
EOT_LA (small) EOT_LA (small)
Remittance (high) Remittance (high)
Remittance (low) Remittance (low)
Tax rate (high) Tax rate (high)
Tax rate (low) Tax rate (low)
Figure 10: Percentage change in the wage rate of skilled labor (2030, %).
Figure 9: Percentage change in the wage rate of unskilled labor (2030, %).
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
10.00
12.00
0.00
2.00
6.00
8.00
4.00
Benchmark Benchmark
Static Static
Variable CA Variable CA
31
EOS_LL (large) EOS_LL (large)
Benchmark Benchmark
FLAB (more) FLAB (more)
FLAB (less) FLAB (less)
Static Static
Variable CA Variable CA
EOS_KL (large) EOS_KL (large)
EOS_KL (small) EOS_KL (small)
32
EOS_LL (large) EOS_LL (large)
EOS_LL (small) EOS_LL (small)
EOT_LA (large) EOT_LA (large)
EOT_LA (small) EOT_LA (small)
Remittance (high) Remittance (high)
Remittance (low) Remittance (low)
Tax rate (high) Tax rate (high)
Figure 13: Percentage change in Japanese income (2030, %).
In this appendix, we explain how we determined wage rates in the initial equilibrium. JIP data includes
“the number of employees” by occupation and sectors. According to this data, the total number of
employees in all sectors is 62,512,899 and the number of skilled and unskilled workers is 11,495,845 and
51,017,054 respectively.
Let 𝑉𝑖 denote payment to labor in sector 𝑖. The, we have the following relation:
𝑉𝑖 = 𝑤𝑖𝑆 𝐿𝑆𝑖 + 𝑤𝑖𝑈 𝐿𝑈𝑖 (A − 1)
where
𝑤𝑖𝑆 and 𝑤𝑖𝑈 are wage rates for skilled and unskilled workers in sector 𝑖.
𝐿𝑆𝑖 and 𝐿𝑈𝑖 are the number of skilled and unskilled workers in sector 𝑖.
We assume that 𝑉𝑖 is equal to the value of “Compensation of Employees” in JIP data. In addition, we
determine the value of 𝐿𝑆𝑖 and 𝐿𝑈𝑖 by “the number of employees” of JIP data. Then Eq. (A-1) means that
if we give the value of one wage rate (for example, 𝑤𝑖𝑆 ), the other (𝑤𝑖𝑈 ) is determined by Eq. (A-1).
Namely, we cannot determine wage rates of skilled and unskilled workers independently.
To determine wage rates, we assume that “wage rate of skilled worker in sector 𝑖 is 100 × 𝑋𝑖 % higer
than that of unskilled workers”. This means 𝑤𝑖𝑆 = (1 + 𝑋𝑖 )𝑤𝑖𝑈 . Substitute this relation into (A-1), we have
𝑉𝑖 = (1 + 𝑋𝑖 )𝑤𝑖𝑈 𝐿𝑆𝑖 + 𝑤𝑖𝑈 𝐿𝑈𝑖
By rewriting this, we have
𝑉𝑖
𝑤𝑖𝑈 =
(1 + 𝑋𝑖 )𝐿𝑆𝑖 + 𝐿𝑈𝑖
We determine the initial wage rates in each sector by this relation.
To derive the value of 𝑋𝑖 , we use “Basic Survey on Wage Structure (Chingin Kozo Kihon Tokei
Chosa)”20. According to this survey, wage rate for “Plant and machine operators, and assemblers” is
267,400 yen per month and wage rate for “Technicians and associate professionals,” “Officials and
Mangers,” “Clerks” is 387,400 per month. We regard the former as wage rate for unskilled workers and
the latter as wage rate for skilled workers and calculate 𝑋𝑖 as follows
387400
𝑋𝑖 = − 1 = 0.45
267400
Note that we assume 𝑋𝑖 is common for all sectors.
20
http://www.mhlw.go.jp/english/database/db-l/wage-structure.html (accessed May 20th, 2016)
33
A-2.1. Notation
Index
Notation Explanation
𝑖, 𝑗 Index of goods and sectors
𝑙 Index of labor type {SKL, USK}
𝑁𝑆 Set of sectors which do not use specific factors
Activity variables.
Notation Explanation
𝑄𝑖 Production level
𝑋𝐷𝐴𝑖 Allocation to domestic and export supply
𝐾𝐿𝑖 Aggregation of capital and labor
𝐶 Aggregate consumption (= utility)
𝐼𝑉 Investment
𝐺 Government expenditure
𝐸𝑋𝑖 Export activity
𝐼𝑀𝑖 Import activity
𝐴𝑖 Armington aggregation
𝐾𝐴 Capital allocation activity
𝐿𝐴𝑙 Labor allocation activity
𝐹
𝐶 Consumption by foreign labor
34
35
Exogenous variables
Notation Explanation
𝑒𝑛𝑑𝑙𝐿 Endowment of domestic labor.
𝑒𝑛𝑑 𝐾 Endowment of capital
𝑒𝑛𝑑𝑖𝑁 Endowment of specific factors
𝑒𝑛𝑑𝑙𝐿𝐹 Supply of foreign labor
𝐶𝐴 Current account surplus
𝑒𝑛𝑑𝑖𝐻𝐻 Endowment of commodities owned by household
𝑡𝑖𝑄 Tax rate on output
𝑡𝑖𝑀 Tariff rates
𝐿𝐹
𝑡 Tax rate on income of foreign labor.
𝜃𝑅 Remittance rate
𝑠 𝐻𝐻 Saving rate of the Japanese household
𝑠𝐹 Saving rate of the foreign labor
36
𝑐 𝐼𝑉 = ∑ 𝜃𝑖𝐼𝑉 𝑝𝑖𝐴 {𝑐 𝐼𝑉 }
𝑖
37
38
39
40
A-2.7. Constraints
𝑒𝑛𝑑𝑙𝐿𝐹 = 0
𝑒𝑛𝑑𝑙𝑖𝐿𝐹𝑆 ≠ 0
41
Remittance.
42