Abstract: This paper presents the results of our experiments to assess average labor quality and labor force utilization in different regions of China using slack-based inefficiency measurement (SBI). We found that there is a discrepancy between different regions’ labor resources and their stages’ of economic development. In central and western regions, the average quality of labor is significantly higher than in eastern regions, but labor force utilization is less efficient. Slow in economic growth and laggard in industrial upgrading, central and western regions have failed to provide their high-quality labor forces with adequate and suitable job opportunities, leading to the discrepancy between labor resource quality and economic development. Resolving this discrepancy might help coordinate economic development across different regions in China.
Key words: labor quality, labor force utilization efficiency, SBI, Tobit Model
JEL: J01, R11
1. Introduction
In China, different regions are at different stages of economic development. Although in recent years the Chinese government has worked to guide regional planning and strongly promoted coordinated economic development among different regions, unbalanced regional development is still salient both economically and socially. Scholars have taken different interpretations to this imbalance. Examples include the differences in each region’s preferential policies and foreign direct investment (Wang Xiaolu, Fan Gang, 2004), the degree of globalization and urbanization (He Canfei, Liang Jinshe, 2004), and productive inputs and technical efficiency improvement (Guo Yuqing, Jiang Lei, 2010). However, existing studies have failed to address the significance of utilizing different labor qualities with respect to coordinated regional development.
After more than 30 years of reform, the biggest restraint on China’s economic growth is no longer capital, but skilled labor. If a region has a large number of high-quality workers and it fully leverages those workers, then it will be much more possible for that region to attain sustainable economic growth. So far in the 21st century, China has enjoyed rising labor quality and improving levels of education. China now has the world’s largest number of higher education graduates, with the annual number rising from 1.776 million in 2000 to 7.071 million in 20081. However, this high-quality labor is not distributed evenly across different regions, and, more importantly, the efficiency of the utilization of this high-skill labor force varies greatly. Only by understanding each region’s different labor resources and labor force utilization situation can more targeted development strategies be implemented. Based on the SBI technique, this paper will try to analyze each region’s labor force utilization efficiency in a way existing studies have failed to do.
Using labor quality data from national economic censuses from 2004 and 2008, we first appraised the average labor quality of different provinces, cities, and autonomous regions. Then we used the SBI technique, incorporating average labor quality, to calculate each region’s labor force utilization efficiency.
The paper is organized as follows: Section 2 provides a brief sketch of the analytical framework, including the method used for measuring labor quality and labor force utilization efficiency (slack-based inefficiency measurement (SBI)); Section 3 reports the results, including regional differences in labor quality, quality-adjusted labor force utilization efficiency, and the utilization efficiency of labor of different qualities; Section 4 discusses how labor force utilization efficiency relates to labor quality and puts forth the regional distribution assets of labor resources; Section 5 uses the Tobit model to discuss the factors affecting labor force utilization efficiency with the aim of finding ways to enhance utilization efficiency; and Section 6 rounds up with major findings and policy implications.
2. Analytical Framework
2.1. Measurement of labor quality
Labor quality is a comprehensive index comprised of education, skill, health, mental state, professional ethics, and more. Among these traits, skill and education are the most important for consideration of a worker’s contribution to technological progress, industrial upgrading, and economic growth. A worker’s education is relevant to, but still different from, his skill. Education is the basis of labor quality, but a worker still needs to have skills and expertise corresponding to his profession. A worker’s skill level is measured by his technical rank, technical titles, and other relevant data2.
Due to limited data availability, most existing studies measure labor quality by education levels alone. However, more details are available from China’s first and second national economic censuses, conducted in 2004 and 2008. The economic censuses provide researchers with provincial data on technical ranks and technical titles, making it possible to appraise labor quality and labor force utilization efficiency by skill level.
When measuring labor quality based on education level, the average education level of the labor force is the most commonly used index. A region’s average labor force education level can be defined as follows:
(1)
Here denotes average education level of labor force in province i, is the proportion of labor of education level j, and is the weight of education level j. Education level j = 1, 2, 3, 4, 5, refers to five different educational attainment levels: less than a junior high school diploma, a high school diploma, an Associate’s degree, a Bachelor’s degree, and a Master’s degree and above. The education levels are weighted as 6, 12, 15, 16, 20 respectively, according to the time required to reach each level.
To measure labor quality by skill level, an index for average skill level must first be constructed. This can be done by assigning weights to different technical ranks and technical titles, in reference to the average education level index. The index for average skill level can be defined as follows:
(2)
Here denotes average skill level of labor force in province i, is the proportion of labor of skill level j in province i, and is the weight of skill level j. Skill level j = 1, 2,……8, refers to 8 different technical ranks or technical titles, 1 being high and 8 being low (i.e., senior professional, medium professional, junior professional, senior technician, technician, senior skilled worker, medium skilled worker, junior skilled worker and ordinary worker (without any title)). Each skill level j is given a weight with values 26, 23, 21,18,15,12, 9, and 6, respectively3.
2.2. Slack-based inefficiency measurement
Data Envelopment Analysis (DEA) is one of the most popular approaches used to appraise efficiency because it is not necessary to impose any explicit functional form on production data. Take two groups of input and output as an example. In Figure 1, the horizontal axis represents input x, and the vertical axis represents output y. Suppose there are three production units ? B, C and D. For any production unit i, xi and yi represent its input and output, respectively. Given the assumption that the production possibility set is monotonic and convex, the possibility set is represented by the area between the envelope curve OBCDE and the x axis, and the envelope curve OBCDE is the production possibility frontier.
It is obvious that B, C, and D are efficient, but how to measure the inefficiency value of unit A? There are four common approaches: the traditional Shephard’s output distance function (SDF), the directional distance function (DDF), the slack-based measure of efficiency (SBM), and the slack-based inefficiency measurement (SBI). This paper adopts the fourth approach. In reference to the work of Fukuyama & Weber (2009) and Wang Bin (2010), the function that calculates the inefficiency value of both input and output can be defined as follows:
(3)
represents the inefficiency value of k’ among k production units during timeframe t; xt,k’ and yt,k’ represent the input vector and output vector of production unit k’ during timeframe t; N and M denote the amount of different kinds of input and output vector, respectively; and are the slack vectors of input n and output m, respectively, given the production unit k’ during timeframe t; zt,k’ is the weight. All slack vectors take a nonnegative value: if a vector takes a value of zero, it means that there is no excessive input (=0) or insufficient output ( =0); if its value is greater than zero, it means that there is either excessive input (>0) or insufficient output
(>0). The constraint
indicates that return to scale is variable (VRS); removing it means constant return to scale (CRS). This paper calculates under the assumption of variable return to scale (VRS).
The inefficiency value function can be further broken down into input inefficiency value and output inefficiency value . For production unit k’ during timeframe t, and can be defined as follows:
(4)
(5)
and
(6)
We use the GDP of each province, municipality and autonomous region as the output value, so M=1. For input, we consider two components, namely stock of fixed capital and quality-adjusted labor input, so N=2. If we change x in into L and K then we get the value of labor inefficiency and that of capital inefficiency , respectively. The lower the inefficiency value, the higher the utilization efficiency, and vice versa.
2.3. The measure of quality-adjusted labor force utilization inefficiency
Many studies use the quantity of labor to measure labor input. This paper also takes labor quality into account using two different methods. The first method calculates the quality-adjusted inefficiency value of total labor force utilization. First, we use the average skill level or education level of the labor force as the adjustment coefficient to appraise the quality-adjusted labor input defined as:
(7)
denotes the quality-adjusted labor input, refers to non-quality-adjusted labor input, and e is the average skill level or education level of the labor force. Coefficient e can be specified as tech when adjusting for average skill level or edu when by average education level. Then we enter the quality-adjusted labor input value into Function (4) to get the quality-adjusted inefficiency value of total labor force utilization.
The second method calculates the inefficiency value for each type of labor. To do this, the quantity of a type of labor, categorized by different technical titles or educational attainments, is incorporated into the SBI model. Therefore, the aggregated labor utilization inefficiency value is defined as:
(8)
and each type of labor’s utilization inefficiency value is defined as:
(9)
Where denotes the labor force utilization inefficiency value, refers to the inefficiency value of labor utilization given a certain technical title or educational attainment j, and J is the amount of different types of technical title or educational attainment.
Each province, municipality, and autonomous region is defined as a production unit except for Hong Kong, Macao, Taiwan and Tibet. For output value, we use the GDP of each province, municipality, and autonomous region given constant prices from the year 2000. For input, there are two components: stock of fixed capital and quality-adjusted labor input. The stock of fixed capital is measured according to the method developed by Zhang Jun et al (2004) and then converted into 2000 constant prices. This paper combines the stock of fixed capital of both Chongqing and Sichuan as “Sichuan” because Chongqing and Sichuan used to be accounted for as a single province, and only combined price indices for investment in fixed assets are available from before 1997. The data we use are from the China Statistical Yearbook and the China Compendium of Statistics 1949-2008.
3. Regional Differences in Labor Quality and Labor Force Utilization Efficiency
3.1. Regional differences in labor quality
In China, the distribution and allocation of labor quality is highly uneven across regions. The national average skill level of labor was 9.9 in 2004 and 9.6 in 2008. This puts the national average skill level at slightly above a medium skilled worker but still below a senior skilled worker. By province Hubei, Yunnan, and Guangxi ranked relatively high in average skill level in 2004. In 2008 Ningxia, Yunnan, and Xinjiang were among the top. In both 2004 and 2008, Zhejiang, Guangdong, Jiangsu, Shanghai, and other more developed provinces lagged behind in skill rankings (Figure 2). In general, average skill levels are relatively high in western regions but relatively low in eastern ones.
As for education levels, the national average was 10.5 in 2004 and 11.1 in 2008. This reflects that the average worker had come close to, but not attained, a high school diploma. By province/municipality, in both 2004 and 2008, Beijing, Jilin, Heilongjiang, Xinjiang, and Inner Mongolia had relatively high average education levels while Zhejiang, Fujian, Jiangsu, and Guangdong had relatively low averages.
For both average skill and average education levels, many developed coastal regions ranked low in average education level. This runs contrary to what one might assume about these more developed regions. However, the average level of skill or education of a province’s labor force depends not only on the quantity and proportion of high-quality labor in that area but also on the proportion of other kinds of labor.
Even though eastern regions enjoy more developed economies, these regions are mainly engaged in processing for trade. This sector absorbs large amounts of labor from central and western regions, and most laborers who move to the coastal regions to work in processing are surplus from the countryside and come with few skills and little education. This negatively impacts the average skill and education levels of eastern regions. For example, the proportions of Master’s degree holders and above in Zhejiang, Fujian, and Guangdong are 0.36 percent, 0.40 percent and 0.69 percent, respectively. These all fall below the national average of 0.71 percent and pale in comparison to Beijing with 3.66 percent, Liaoning with 1.04 percent, and Shanghai with 1.76 percent. At the same time, the proportions of workers with less than a high school diploma in Zhejiang, Fujian, and Guangdong are 58.03 percent, 49.04 percent and 46.37 percent. These statistics do not look favorable compared with the national average of 39.58 percent and the averages of Jilin (29.09 percent), Beijing (29.87 percent), and Inner Mongolia (31.50 percent).
3.2. Labor force utilization efficiency adjusted for average labor quality
To find labor force utilization efficiency adjusted for average labor quality, we first calculate the non-quality-adjusted inefficiency value of labor force utilization with Function (2). Then we apply Function (5) to the data from Figure 2 and Figure 3 to get the quality-adjusted labor input, which is then used in Function (6) to calculate the quality-adjusted inefficiency value of labor force utilization. Lastly, we compare the inefficiency value with and without labor quality adjustment in order to better study the state of labor force utilization.
3.2.1 Comparison of labor force utilization inefficiency with and without labor quality adjustment
As shown in Table 1, some provinces’ labor force utilization efficiencies are superior regardless of whether or not labor quality is taken into account. For example, Guangdong, Hubei, Hainan, and Qinghai in 2004; and Guangdong, Inner Mongolia, Liaoning, Heilongjiang, Shanghai, Hainan, Qinghai, and Ningxia, among others, in 2008. These regions top the nation in labor force utilization efficiency.
In most regions, the quality-adjusted inefficiency value of labor force utilization is higher than the inefficiency value without quality adjustment. This means labor force utilization is less efficient when labor quality is considered. Examples of this kind from 2004 include Guangxi, Liaoning, Jilin, and Yunnan; examples from 2008 include Shanxi, Sichuan, Xinjiang, and Hunan. This indicates that most regions cannot distinguish different types of labor effectively or match appropriate labor quality with appropriate jobs. In these regions, individuals of different quality are treated as the same to some degree, and differences in education and skill are at least partially overlooked. High-quality workers, for example, are not given more demanding positions but are instead doing the same low-skill jobs as ordinary workers. Since high-quality workers are not fully leveraged, their abilities are not brought to bear on the economy, which explains why labor force utilization efficiency declines when adjusted for labor quality.
A few provinces’ labor force utilization inefficiencies shrunk when labor quality was taken into account. Examples from 2004 include Shanghai, Zhejiang, Fujian and Jiangsu; examples from 2008 include the abovementioned provinces and municipalities as well as Tianjin, Jiangxi, Henan, and Shaanxi. This indicates that these regions are identifying different types of labor more effectively and allocating labor to appropriate positions. In these regions, high-quality workers are leveraged to bring about higher productivity. These regions make better use of labor of all types and consequently enjoy higher labor force utilization efficiency.
Some regions’ labor force utilization efficiency increases when adjusted for skill level but decreases when adjusted for education level. Some examples from 2004 are Tianjin, Guizhou, and Shaanxi, and from 2008 Beijing, Jilin, and Shaanxi. These regions might be effective in distinguishing labor by skill level but ineffective in doing so according to education level. The reverse was true in some provinces. Examples from 2004 include Hebei and Shanxi, and from 2008 Heibei, Guangxi, Guizhou, and Yunnan. These regions’ labor force utilization efficiencies increase when adjusted for education level but decrease when adjusted for skill level.
3.2.2. Regional difference in labor force utilization efficiency
As shown in Table 1, quality-adjusted labor force utilization efficiency varies significantly among regions. In 2004, for example, the national labor force utilization inefficiency value average was 0.056. Using that as a benchmark, we can divide all provinces, autonomous regions and municipalities into three categories.
Regions in Category 1 are on the production frontier and have the highest labor force utilization efficiency (i.e. labor force utilization inefficiency value = 0). The six provinces in Category 1 are Hainan, Jiangsu, Qinghai, Guangdong, Hubei, and Fujian. Generally speaking, in these six provinces the supply of each type of labor meets the demand. Economic development and industrial structure complement labor force structure, and all types of labor can contribute according to their abilities.
Regions in Category 2 are relatively efficient in labor force utilization (labor force utilization inefficiency value 0.056). There are 13 provinces/municipalities/autonomous regions in Category 3: Shaanxi, Anhui, Hunan, Yunnan, Xinjiang, Henan, Sichuan, Guangxi, Jilin, Jiangxi, Beijing, Gansu, and Shanxi. Most of the regions in this category are in western China, where economic development and industrial upgrading lag well behind labor force structure advancement. The economic systems in these areas might not provide enough incentives for workers; as a consequence, all types of labor are not fully leveraged.
Further study of Hubei and Guangdong can deepen the understanding of regional differences in labor force utilization efficiency. Hubei and Guangdong represent two different types of regions. In 2004, Hubei had the highest average skill level while Guangdong had the lowest. However, Guangdong’s economy is more developed than that of Hubei. In 2004, per capital GDP in Guangdong was 19,707 RMB versus 10,500 RMB in Hubei. Although both provinces are on the production frontier and have some of the highest labor force utilization efficiency ratings in China, the mechanisms in both provinces are very different, mainly because of differences in industrial structure.
Guangdong has a relatively heavy industrial concentration while Hubei has a higher share of services.4 At present, processing is still vital for Guangdong. In 2006, the export value of processing orders in Guangdong accounted for 14.3 percent of provincial GDP, and that of feeding processing was 54.7 percent; the combined share was 69 percent of Guangdong’s GDP5. But processing is at the lower end of the industrial chain, and processing work mostly demands ordinary, not skilled, workers. Guangdong is a major destination for surplus rural labor, further illustrating its dependency on unskilled labor and relatively low demand for high-quality workers.
Industries in Hubei require labor with higher skill levels. Hubei is home to China’s second largest automobile factory as well as its largest truck production factory, heavy machine tool factory, and packaging machinery factory. Information technology is one of Hubei’s major industries. Its share of the provincial industrial output was as high as 8.5 percent in 2005. All of these industries require relatively high-skilled individuals, which is why the industrial structure in Hubei has more opportunities for highly-skilled workers.
Within Category 3 areas, Beijing is worth further discussion. Beijing has the highest average education level in China and, at 8th, a very high average skill level. Despite this, its labor force utilization efficiency rank is third from the bottom. This contradiction between high labor quality and very low utilization efficiency can be attributed to three factors.
First, Beijing’s inefficiency was caused by high-quality labor being over-concentrated and excessively competitive. Senior technical title holders accounted for 4.55 percent of Beijing’s working population in 2004, the highest rate in China. This over-concentration of high-quality labor has resulted in an oversupply of this type of labor, and therefore many high-quality individuals are working below their ability which in turn leads to low efficiency.
Second, proportions of different types of labor are unbalanced in Beijing. At a given stage of development, there is an optimal proportion for each type of labor in a market. If proportions are not optimal, there will be inefficiencies. We compared the proportions of different types of labor in Beijing with those in Jiangsu, a province on the labor force utilization efficiency frontier, and with the national average. Beijing and Jiangsu are both developed regions with similar industrial structures, so a comparison is feasible. In 2004, the proportion of senior and medium professionals to junior professionals was 4:10 in Beijing, 2:10 in Jiangsu, and 2:10 nationally. Beijing had twice the number of senior professionals of Jiangsu or the nation as a whole. The proportion of senior professionals to senior technicians was 15.9:1 in Beijing, 7:1 in Jiangsu, and 8:1 nationally. Here again Beijing would seem to have too many senior professionals compared to its number of senior technicians. This excess of high-quality labor contributes to Beijing’s high inefficiency rating.
Third, Beijing’s industrial upgrading lags behind its labor structure advances, as can be seen when comparing Beijing with Shanghai. Beijing’s average skill level is far higher than Shanghai’s, but Beijing’s industrial structure has lagged behind (Wu Fuxiang, Zhu Lei, 2011). We compared the share of the value added to GDP for six major industries in Beijing and Shanghai (electronic information, finance, commerce circulation, car manufacturing, equipment manufacturing, and real estate). In Beijing in 2004 the value of all these industries, except for finance and commerce circulation, fell short of those value in Shanghai. With its industrial structure failing in these significant ways, Beijing cannot provide sufficient job opportunities for high-quality individuals. As a result, Beijing’s labor force, particularly its high-quality labor force, is not being fully utilized.
3.3 Utilization efficiency of each type of labor
3.3.1 National averages for the utilization efficiencies of different types of labor
On a national scale, in utilization efficiency on skill levels of labor, China utilizes ordinary workers the most efficiently; this is followed by junior skilled workers, technicians, senior technicians, medium professionals, and senior professionals; with the least efficiency in utilizing its medium skilled workers and senior skilled workers (Figure 4). In terms of education level (Figure 5), the utilization of college and Associate’s degree holders is the least efficient, followed by Bachelor’s degree and finally Master’s and above degree holders. Workers with high school diplomas and below are utilized relatively efficiently.
Why is the utilization of lower-quality labor more efficient? Simply put, China’s current industrial structure demands more ordinary workers. In some regions, the demand for ordinary workers even outstrips supply. Insufficient demand for senior skilled and medium skilled workers implies that China is still at the lower end of the global value chain.
3.3.2 Regional differences in each type of labor utilization efficiency
Labor force utilization efficiency varies significantly among different provinces. Looking at skill level data from 2004, all types of labor were utilized efficiently in eastern coastal regions, Heilongjiang and Liaoning in northeastern China, and Inner Mongolia, Qinghai, and Ningxia in western regions. In most central and western regions, the utilization of all types of labor is relatively less efficient.
4. Discussing the Discrepancy between Labor Resources and Economic Development
The comprehensive picture of labor quality and labor force utilization efficiency across different provinces is displayed in Figure 6. The horizontal axis represents the national average of labor quality; the vertical axis represents the national average of utilization inefficiency. Along the two axes, the diagram is split into 4 quadrants, each representing a different combination of labor quality and labor force utilization efficiency.
Regions in Category 1 are in the fourth quadrant. Category 1 regions include Hubei, Heilongjiang, Ningxia, Qinghai, and, to a lesser degree, Liaoning and Hainan. Most of these regions are in northeastern China, and some are in the west. In these regions, both labor quality and labor force utilization efficiency are rather high.
Category 2 regions are in the second quadrant. These include Shanxi, Henan, Anhui, and, to some degree, Sichuan, Jiangxi, Shaanxi, and Guizhou. Most of the regions are in central China and are described by both low labor quality and low labor utilization efficiency. These regions have two tasks before them: first, accelerate industrial upgrading in order to increase labor force utilization efficiency; second, enhance education and improve labor quality so as to develop a talent pool for long term economic development.
Regions in Category 3 are in the first quadrant. These include Gansu, Yunnan, Xinjiang and Hunan, and other provinces that might at first not seem to belong in this category, such as Beijing and Guizhou. Most of Category 3 regions are in western China with relatively high labor quality and low utilization efficiency. These regions should prioritize industrial upgrading so that their labor forces, particularly their higher-quality labor forces, can be used effectively.
Category 4 regions fall into the third quadrant. Some examples here are Shanghai, Zhejiang, Guangdong, Fujian, Shandong, Jiangsu, Hebei, and, to some extent, Tianjin. Most of these regions lie in eastern coastal areas with relatively low average labor quality but high utilization efficiency. There are two primary reasons for this situation. On one hand, these regions are more economically developed than most others, so opportunities for high-quality workers are abundant. On the other hand, many lower-quality workers, especially from central and western China, are attracted to these regions, dragging down average labor quality. These regions’ speedy development has attracted many workers from other parts of China, and as a result these regions have a responsibility to raise their labor force utilization efficiency. These regions are faced with three tasks to deal with this socioeconomic situation. First, these regions should enhance education to promote the quality of the local labor force in order to provide a talent pool that can take advantage of industrial upgrading. Second, they should formulate favorable policies to attract high-quality labor from other regions, especially from central and western China, so as to support industrial upgrading. Third, they should accelerate industrial upgrading and provide ample job opportunities for high-quality workers both native and foreign to these regions.
The discrepancy between labor resources and economic development among China’s regions is clear from the data and the diagram. Eastern regions that enjoy rapid economic growth should shoulder more of the responsibility for industrial upgrading. However, average labor quality is relatively low in eastern regions, and they lack the talent to support the required industrial upgrading. Conversely, central and western regions have high average labor quality but relatively lackluster industrial development, so labor force utilization is not efficient there either.
5. Factors Affecting Labor Force Utilization Efficiency
We used econometric models to study the factors affecting labor force utilization efficiency in order to better understand the relationship between utilization efficiency and labor quality. Our hope was to find ways to enhance labor force utilization efficiency. To make this description consistent with other studies, we refer to Wang Bing et al. (2010) and convert the inefficiency value to the efficiency value with the equation EL=1/(1+IEL). Considering that the labor force utilization efficiency value is between 0 and 1, we chose the Tobit Model for these regressions. The basic model is as follows:
(13)
In this model, t = 2004 or 2008; variables with the superscript t represent data or regression coefficients of the year 2004 or 2008. et means labor quality, which can be specified as average education level edu or average skill level tech. pk is per capita stock of fixed capital which is equivalent to each province’s stock of fixed capital divided by working population. pw is average labor remuneration, whose value is decided by dividing each province’s labor remuneration by working population and then converting to 2000 constant prices by CPI. We also incorporated the variable FDI early on to represent a region’s degree of opening up. In the regression equation for 2004, we first incorporated the ratio of rural migrants to the total working population in a city to represent labor mobility (there was no relevant data in the 2008 economic census). However, results indicate that neither of the abovementioned two variables has a significant regression coefficient, so we omitted these two variables for the sake of the following discussion.
To study how industrial structure influences labor force utilization efficiency, we added the shares of secondary and tertiary industries in GDP into Function (13), getting:
(14)
Here, ind2 and ind3 represent the shares of secondary and tertiary industries in GDP.
To study regional differences in labor force utilization efficiency, we incorporated dummy variables for different regions and interactive terms between these dummy variables and labor quality to get:
(15)
Here, d1, d2, and d3 represent western regions, central regions, and eastern regions, respectively.
To make the regression results more robust, we ran a regression by treating the inefficiency value IEL as the explained variable. In this regression, the signs of the regression coefficients are basically opposite to those in the regression with EL as the explained variable. This means that regression results are robust. In this paper, we mainly discuss the results of the regression where EL is used as the explained variable.
See Table 4 for the regression results of Function (13). In both 2004 and 2008, the regression coefficient of labor quality is negative and very significant, regardless of whether E-edu, labor force utilization efficiency adjusted for average education level, or E-tech, labor force utilization efficiency adjusted for average skill level, is used as the explained variable. This means that higher labor quality leads to lower labor force utilization efficiency. The regression coefficient of the per capita stock of fixed capital is positive. It is significant for regressions for 2004 and 2008 when labor force utilization efficiency adjusted for average skill level is used as the explained variable. This finding appears to be consistent with economic theory. The regression coefficient of average labor remuneration is positive, and it is significant when the explained variable is labor force utilization efficiency adjusted for average education level. This means raising labor remuneration is one way to enhance labor force utilization efficiency.
See Table 5 for the regression results of Function (14). Compared to the regression results of Function (13), the regression coefficient signs of labor quality, per capita stock of fixed capital, and average labor remuneration have not changed significantly. However, the coefficient of industrial structure is not significant because industrial upgrading has not yet been able to effectively raise labor force utilization efficiency.
See Table 6 for the regression results of Function (15). The regression coefficient of the per capita stock of fixed capital is significantly positive, and that of the average labor remuneration is positive but insignificant. In most regression results, the regression coefficient of labor quality is significantly negative. This indicates that eastern regions’ labor quality and utilization efficiency are inversely correlated. d1 and d2 are significantly negative, which means, regardless of the effect of explanatory variables, that the labor force utilization efficiency of central and western regions is lower than that of eastern regions. When E-edu, or labor force utilization efficiency adjusted for average skill level, is used as the explained variable, then the regression coefficient of the interactive term of d1, d2, and the average education level is significantly positive, and its absolute value is larger than that of the average education level’s regression coefficient. This means that lifting average education level can help to raise labor force utilization efficiency in central and western regions.
6. Conclusions and Suggestions
This paper presents the following findings: (1) Labor quality varies significantly among different regions. When measured by average skill level and education level, average labor quality is generally higher in western regions and lower in eastern regions. (2) In most regions, the quality-adjusted efficiency value of labor force utilization is lower than the efficiency value without quality adjustment. (3) Labor of lower quality is generally utilized more efficiently, while higher quality labor is utilized less efficiently. (4) When labor quality is taken into account, labor force utilization efficiency is lower in western regions and higher in central and eastern regions. (5) Further econometric analysis indicates that labor quality is inversely related to labor force utilization efficiency. Labor force utilization efficiency can be raised by increasing the per capita stock of fixed capital and labor remuneration. In central and western regions, enhancing average education levels can also help to raise labor force utilization efficiency.
After analyzing these findings, we conclude that the migration of low-quality labor from central and western regions to eastern regions has led to the emergence of higher labor quality in central and western regions. However, due to slow economic development and laggard industrial upgrading, high-quality labor in central and western regions is not fully leveraged. On the contrary, eastern regions, while enjoying rapid economic development and industrial upgrading, are successfully leveraging all types of labor, especially high-quality labor. However, relatively low average education levels of labor force in eastern regions may have impeded further industrial upgrading.
The discrepancy between labor resources and economic development hinders the full utilization of labor resources, especially for central and western regions. This discrepancy also constrains industrial upgrading (mainly in eastern regions) and seriously impedes coordinated development among different regions. A possible solution has two parts. The first is for the government to accelerate the promotion of industrial upgrading in central and western regions. In this way more job opportunities for high-quality labor will be provided and labor force utilization efficiency in central and western regions will rise. Companies in eastern regions should be encouraged to invest and build factories in central and western regions. The government should encourage various industries to move their operations to central and western regions; not just labor-intensive industries, but also technology-intensive industries and industries that require more skilled labor should also be urged to leave the comfort of coastal eastern regions. The government should also continue to promote labor mobility across regions and eliminate relevant institutional barriers so that high-quality labor can flow from regions where such labor will be under-utilized to regions where it can be properly applied. The government should not ask university graduates to work in central and western regions blindly; it should instead facilitate industrial upgrading in these regions to attract university graduates there. Careful and deliberate implementation of these recommendations can lead to higher efficiency and greater prosperity for China’s work force and its citizens.
References
[1] Guo, Yuqing, and Lei Jiang. 2010. “The Root Cause of Widening Regional Economic Disparity in China: Capital Deepening or Efficiency Improvement.” Journal of Quantitative & Technical Economics, 2010(7).
[2] He, Canfei, and Jinshe Liang. 2004. “The Time and Space Change in the Difference Between China’s Regional Economies: Commercialization, Globalization and Urbanization.” Journal of Management World, 2004(8).
[3] Wang, Bing, Yanrui Wu, and Pengfei Yan. 2010. “China’s Regional Environmental Performance and TFP Growth.” Economic Research Journal, 2010 (5).
[4] Wang, Xiaolu, and Gang Fan. 2004. “Analysis on the Regional Disparity in China and the Influential Factors.” Economic Research Journal, 2004(1).
[5] Wu, Fuxiang, and Lei Zhu. 2011. “Technology Insetting, Industrial Amalgamation and the Effect of Structure Transformation.” Shanghai Journal of Economics, 2011(2).
[6] Zhang, Jun, Guiying Wu, and Jipeng Zhang. 2004. “The Estimation of China’s Provincial Capital Stock: 1952-2000.” Economic Research Journal, 2004(10).