Factor Decomposition and Policy Implication of China’s North-South Regional Differences

1 Introduction

In recent years, the economic disparity between northern and southern regions of China has become one of the focal points of academic and policy-making attention. With its vast territory and large population, China exhibits significant differences in geographical location, natural environment, climate characteristics, agricultural production, residents’ living habits and traditional culture, economic development and population distribution, as well as transportation modes between the North and South. Since the beginning of the 21st century, the overall gap in regional economic development levels has been narrowing, while the economic disparity between the North and South has continued to widen, becoming a prominent feature and typical manifestation of China’s unbalanced and inadequate development (An and Li, 2023). Consequently, the issue of regional disparities has recently attracted extensive scholarly research. For instance, Sheng et al. (2018) initially analyzed the imbalance in regional development through indicators such as GDP growth rates, household income, and fiscal revenue, finding that the economic gap between the North and South showed a trend of gradual expansion. Yang et al. (2018) used total GDP as the fundamental indicator for judgment, Dai (2020) incorporated general budget revenue, Zhang et al. (2019) adopted per capita GDP as the measurement metric, and Yang et al. (2021) considered both aggregate and per capita GDP indicators—all confirming the continuous expansion of economic disparities between the North and South.

Based on the measurement and evaluation of differences in levels and trends, some studies have analyzed and identified the causes of economic disparities between northern and southern regions in China. In summary, existing research suggests that the main factors contributing to the widening economic gap between these regions include natural endowments and historical development factors (Zhang and Fan, 2021; Su, 2022; Xiao et al., 2022), institutional factors such as marketization (Dong and Chi, 2020; Cong et al., 2022; Zhang and Zheng , 2023), industrial structure factors (An and Zhou , 2021; Zheng et al., 2021; Wu , 2023), regional policy factors (Wei et al., 2020; An and Li , 2023; Wang et al., 2023), and others. These research findings have enriched and deepened our understanding of economic disparities between northern and southern China, particularly highlighting that such gaps result from the combined effects of multiple factors. However, from a methodological perspective, most studies primarily use regression analysis to examine the impact of various factors on regional disparities, without directly analyzing the direct influence of economic growth factors on regional differences.

To address the shortcomings in this methodological approach, this paper adopts a development accounting framework and utilizes the decomposition method to directly measure the specific contribution shares of growth factors such as total factor productivity, physical capital, and labor force to the output disparities between northern and southern regions. Using provincial-level regional data from China spanning 1978 to 2022, the decomposition results reveal three key findings: (1) While the overall total output disparity between northern and southern regions shows a continuous expansion trend, the gap of output per worker has remained relatively stable over the past decade, with labor force inflow and capital inflow becoming crucial factors driving faster growth in southern regions; (2) Total factor productivity disparities have consistently been a significant determinant of regional differences and will likely dominate future trends in regional disparities; (3) Physical capital disparities are notably influenced by regional policies, though policy effectiveness requires comprehensive coordination across multiple dimensions. These findings demonstrate that factor decomposition is vital for understanding the economic disparities between China’s northern and southern regions and their evolving trends, serving as a key approach to narrowing regional gaps and fostering complementary regional development patterns. This study provides a novel perspective on analyzing regional economic disparities, offering a systematic examination of the contributions of various factors to North-South disparities since China’s reform. The research holds practical reference value for comprehending economic disparities between northern and southern regions and formulating relevant policies.

The rest of the paper is organized as follows: Section 2 describes the regional disparity decomposition framework based on development accounting. Section 3 reports the indicators and data used in this study. Section 4 presents the results of the accounting decomposition. Section 5 provides the summary and policy implications.

2 Decomposition Method

Drawing on the framework of “Solow Development Accounting” developed by Klenow and Rodriguez-Clare (1997), Hall and Jones (1999), and Easterly and Levine (2001), this paper employs a production function-based decomposition method to analyze the economic disparities between China’s northern and southern regions. Building on the research of Fu and Wu (2006, 2007), we adopt the following Cobb-Douglas production function to characterize regional production:

where Y represents output, i denotes region, t indicates time, A stands for technological level (production efficiency), K represents physical capital input, H represents (effective) labor input, which is the product of labor force quantity L and human capital level h (i.e., H = Lh). The output elasticity of physical capital is defined as 1−α, while α represents the output elasticity of human capital. According to neoclassical growth theory, the output elasticity of factors generally equals their share of returns in total output.

Taking the logarithm of both sides of Equation (1), we can decompose regional output into

Using either Equation (1) or Equation (1), the output difference between northern regions (indicated by the subscript N) and southern regions (indicated by the subscript S) in China is represented as:

Alternatively, according to Equation (2), the total output differcens can be expressed in logarithmic form

The above two equations show that the output difference between the North and South can be decomposed into the contribution of physical capital difference, labor input difference and total factor productivity difference.

The decomposition above considers the total output of two regions, specifically the differences in economic aggregates. Since differences of output per worker better reflect regional development levels, the decomposition of differences in output per worker also carries significant economic implications. By dividing both sides of Equation (1) by the regional labor force size, we can express regional labor productivity as

where y=Y/L represents output per unit of labor, k=K/L represents physical capital input per unit of labor, and h=H/L represents human capital level. Taking a logarithms on both sides of Equation (5), we can get

Thus, the difference in output per worker between northern and southern regions is expressed as

Alternatively, it can be expressed in logarithmic form as

Equations (7) and (8) show that the differences in output per worker between the North and the South can be decomposed into the contribution of the difference in physical capital per worker between the two regions, difference in human capital level and the total factor productivity difference.

In addition, for multiple regions, we can use covariance decomposition to decompose the overall difference of output per worker in each region into the contribution of difference in physical capital per worker, difference in human capital level and the total factor productivity difference. That is:

3 Variables and Data

The decomposition of output differences in this paper involves variables such as regional output, physical capital, labor force, human capital, and factor returns. The sample covers 31 provincial-level regions from 1978 to 2022 (data on investment and education levels for each region in 2023 are unavailable, therefore 2023 is not included in the time series of the analysis), excluding Taiwan, Hong Kong, and Macao. According to the classification used in most current studies, the southern region in this paper includes 16 provinces: Anhui, Jiangsu, Zhejiang, Shanghai, Hubei, Hunan, Sichuan, Chongqing, Guizhou, Yunnan, Xizang, Guangxi, Jiangxi, Fujian, Guangdong, and Hainan; while the northern region comprises 15 provinces: Beijing, Tianjin, Inner Mongolia, Hebei, Gansu, Ningxia, Shanxi, Shaanxi, Qinghai, Xinjiang, Shandong, Henan, Liaoning, Jilin, and Heilongjiang.

The output is measured using the Gross Domestic Product (GDP) indicator, with units in 100 million yuan. The GDP deflation index has been adjusted to reflect the 2000 constant prices. Data are sourced from the annual volumes of the China Statistical Yearbook, while data prior to 1990 were referenced from the China Compendium of Statistics 1949–2008.

The capital investment is measured using regional capital stock indicators, with the unit being 100 million yuan. As China does not have official capital stock statistical data, we employ the perpetual inventory method for estimation. The calculation method is: 1) Adjust the total fixed capital formation over the years to 2000 prices using regional fixed asset investment price indices; 2) Obtain the initial capital stock by dividing the total fixed capital formation in 1978 by the average investment growth rate during that period; 3) Subsequent annual capital stock is calculated using the perpetual inventory method, with a depreciation rate set at 6%. The original data are sourced from the China Statistical Yearbook and the Statistical Dictionary of China’s Fixed Asset Investment. After 2016, regional fixed capital formation data were no longer published, so we utilized the total fixed asset investment figures from each region for estimation.

The labor force input is measured using the total number of employed individuals in society, with units of ten thousand. Data before 1990 are sourced from the China Compendium of Statistics 1949–2008, while data from 1991 to 2016 are adjusted based on population changes through sampling surveys and total employment figures. Data from 2020 to 2022 are derived from the China Statistical Yearbook. Most data from 2017 to 2019 are missing. We have collected data from regional statistical yearbooks on the one hand, and performed interpolation processing based on labor force growth trends in various regions, on the other hand.

Measuring human capital is relatively difficult and complex. This paper adopts a method similar to Wang and Yao (2003), using the average education level of residents to measure regional human capital levels. The indicator uses the average number of years of education for the population aged 6 and above, assuming that the average education years for residents with illiteracy/semi-literate status, primary school, junior high school, senior high school, and college or higher education are 0, 6, 9, 12, and 16 years respectively. The calculation formula is: h = prim×6 + midd×9 + high×12 + univ×16, where h represents human capital level, and prim, midd, high, and univ represent the proportion of residents with primary school, junior high school, senior high school, and college or higher education levels in the population aged 6 and above. The China Statistical Yearbook and China Population & Employment Statistical Yearbook provide sampling data for 1996–1999, 2002–2009, 2011–2019, and 2021–2022. The fourth and fifth national censuses provide data for 1990 and 2000. The third national census offered educational attainment data for the population aged 12 and above in 1982. We approximated the number of people aged 6–12 by dividing the number of students enrolled in schools that year by the school-age enrollment rate, thereby estimating the average education years for the population aged 6 and above in 1982. Data from 2001, 2010, and 2020 were calculated through arithmetic averaging of consecutive years. The trend interpolation method was used for the periods 1983–1989 and 1991–1995, with adjustments made using the average number of years of education for school enrollees as weights based on variations between the starting years. As no suitable initial year was available for the period 1978–1981, we extrapolated the results from the 1983–1989 data.

This paper uses the factor income share as the factor output elasticity in the production function (i.e., α and 1−α). The factor income share is obtained based on the income approach of national economic accounting. China’s income approach accounting decomposes the gross domestic product (or regional GDP) into four components: labor compensation, fixed asset depreciation, producer net taxes, and operating surplus. According to the indicator interpretation by the National Bureau of Statistics, labor compensation is relatively close to the total income of effective labor in national income. We take its proportion in income approach GDP as the labor compensation share, with the remaining portion representing the physical capital income share. The National Bureau of Statistics provides income approach GDP accounting data for various regions from 1992 to 2017, which we used to calculate the annual labor compensation shares across regions. The results (see Figure 1 below) show that this ratio remains relatively stable with minimal regional variations between northern and southern regions in China. We adopt the regional average values over the years as approximate results for labor output elasticity, setting α=0.5041.

Figure 1

The Proportion of Labor Compensation in GDP of Northern and Southern regions

Source: The website of National Bureau of Statistics and provincial statistical yearbooks, as well as the China Compendium of Statistics 1949–2008.

4 Decomposition Results

This section reports our decomposition results and discusses the results.

4.1 Decomposition of Total Output Differences

First, we utilized Equation (4) to calculate the total output disparity between northern and southern regions from 1978 to 2022, along with their respective factor contributions. Figure 2 presents the results. Overall, the total output gap between northern and southern regions has been on an upward trajectory, with recent fluctuations observed over the past two years. However, whether this trend will continue to narrow regional disparities remains to be seen. In terms of contribution shares, each factor exhibits distinct trends and varying degrees of influence.

Figure 2

Decomposition of Total Output Differences between the North and South

While labor input disparities have been the primary contributor to regional output differences between northern and southern China for most years, their relatively stable nature and limited fluctuation suggest they are not the key driver behind the persistent expansion of this gap. However, with the gradual relaxation of population control policies and talent attraction measures, southern regions have recently shown accelerating growth in labor input, which has also contributed to the widening overall regional output disparity.

The contribution of differences in physical capital input is volatile. From 1978 to 2000, disparities in physical capital investment were the primary factor contributing to the widening gap in total output between northern and southern regions. After entering the 21st century, with the adjustment of national regional development strategies—particularly the implementation of major initiatives such as the “Western Development”, “Northeast Revitalization”, and “Central Rise” —the central government’s investments in northern regions continued to increase, leading to a rapid narrowing of the disparity in physical capital investment between the two regions. However, around 2015, as China’s economic growth slowed down, especially with the deceleration of investment growth, the decline in investment in southern regions was relatively gradual while that in northern regions became more pronounced, resulting in a reversal of the trend of narrowing disparities in physical capital input between the North and South.

Since 1978, the most significant factor driving the widening output gap between northern and southern China has been changes in total factor productivity (TFP). Notably, during the early stages of reform and opening-up, southern regions’ TFP levels were lower than those in northern regions, with the South surpassing the North starting from the early 1990s. Over the subsequent two decades, however, southern regions experienced a much faster growth in TFP. Starting from 2006, the TFP disparity became the primary contributor to regional output differences, with its contribution share even exceeding the combined impact of physical capital and labor inputs for a period. This trend only began reversing in recent years. If sustained over time, this shift could become a driving force for narrowing the North-South divide.

Table 1 reports the changes in total output disparities and factor contribution between northern and southern regions since the reform and opening up. Overall, from 1978 to 2002, the average annual expansion rate of China’s total output disparity between northern and southern regions was 1.06%. Among these, the contribution share from changes in total factor productivity accounted for 61.34%, being the primary factor driving the widening of regional output disparities; the contribution share from changes in physical capital input accounted for 28.73%, which also had a significant impact; while the contribution share from changes in effective labor input accounted for 9.93%, showing a relatively smaller magnitude. This indicates that in the long term, production efficiency remains the key determinant of regional output growth. To narrow regional gaps and promote balanced regional development, it is still necessary to vigorously enhance the rapid improvement of production efficiency in less developed regions.

Table 1

Changes in Total Output Differences and Factor Contributions between the North and South

Annual average rate of change (%)	Total output	Total factor productivity	Physical capital	Effective labor input
1978–1990	0.86	0.39(44.98)	1.04(59.86)	–0.08(–4.83)
1990–2000	1.53	0.91(59.72)	1.48(48.09)	–0.24(–7.81)
2000–2010	1.02	1.83(178.77)	–1.60(–77.56)	–0.02(–1.21)
2010–2022	0.91	–0.28(–31.19)	1.31(71.83)	1.07(59.37)
1978–2022	1.06	0.65(61.34)	0.61(28.73)	0.21(9.93)

1. Note: The contribution share (%) is shown in parentheses in the last three columns.

To analyze the changes in regional disparities in greater detail, we divided the period into four phases (roughly corresponding to the initial stage of reform and opening-up, the comprehensive implementation phase of reform and opening-up, the strategic adjustment period of regional development, and the new era of economic development) based on factors such as China’s economic development and regional disparity trends since the reform and opening-up. We also reported the decomposition results of total output disparity changes across these phases. Comparing these four stages, we found that although the total output disparity between northern and southern regions has been expanding, its growth rate was slowest during the initial stage of reform and opening-up, fastest during the comprehensive implementation phase, and then gradually slowed down thereafter. From the perspective of factor contribution changes, several prominent features stand out: First, between 2000 and 2010, the disparity in physical capital input between northern and southern regions rapidly narrowed, fully demonstrating the role of the central government’s strategic adjustment of regional development. However, during this period, the disparity in total factor productivity between these regions expanded more rapidly. Second, since entering the new era, while total factor productivity disparities between northern and southern regions have narrowed, the disparity in physical capital input has started to widen again. Third, effective labor input disparities generally narrowed the total output disparity between northern and southern regions during most periods, but have shown significant changes since entering the new era, becoming a major factor in expanding the total output disparity. These characteristics indicate that relying solely on central government policies and material investments to narrow regional disparities is inefficient and costly. The flow of physical capital, labor, and human capital is related to policy factors, but is primarily influenced by regional production efficiency.

4.3 Regional Disparities in Output per Worker across China

To reflect the overall pattern of regional inequality in China since the reform and opening-up and to provide a comparative context for the North-South disparity analysis, we applied Equation (9) to estimate the covariance of output per worker and the contributions of various factors across 31 provinces from 1978 to 2022. Figure 4 presents the results.

Figure 4

Decomposition of Disparities in Output per Worker across China

In terms of overall trends, regional economic disparities in China narrowed slightly in the initial phase of reform and opening-up, then expanded rapidly, peaking around the year 2000. Since the beginning of the new century, these disparities have shown a pronounced downward trend, with the current level of overall disparity now lower than that at the start of the reform period. Regarding the share of contributions by different factors, physical capital per worker has been the key determinant of regional disparities in most years. Its trend closely mirrors that of output per worker, and its contribution has been the largest in magnitude. Human capital level, on the other hand, has consistently worked to reduce regional disparities, though its contribution has been relatively modest. The impact of total factor productivity on regional disparities has generally exhibited an upward trend. Particularly in recent years, its contribution has surpassed that of physical capital, making it the most significant factor driving overall regional disparities in China.

A comparison between Figure 4 and Figure 3 reveals several key insights: First, while overall national disparities have generally narrowed since 2000, the North-South disparity has continued to widen, becoming an important dimension of regional inequality. Second, the contribution of physical capital per worker to the national output-per-worker disparity is significantly larger than its contribution to the North-South output-per-worker disparity. Third, correspondingly, the contribution of total factor productivity disparity to the North-South output-per-worker gap is far greater than its contribution to the overall national disparity. These findings indicate that in the coming period, greater attention must be paid to the North-South disparity, particularly the differences in production efficiency between the two regions.

Figure 3

Decomposition of Disparity in Output per Worker between the North and South

5 Conclusion and Policy Implications

This study develops a factor decomposition framework for analyzing the economic disparity between northern and southern China based on a fundamental development accounting approach. It examines the impact of disparities in total factor productivity (TFP), physical capital, and human capital on both total output and output per worker since the reform and opening-up period. The main findings are as follows: (1) The total output gap between northern and southern regions has shown a consistent widening trend overall, although output per worker has changed little over the past decade. Inflows of labor and capital have become significant drivers of total output growth in the southern region. (2) Disparities in TFP have consistently been a critical factor influencing the North-South divide and are expected to largely shape future trends in regional inequality. (3) Disparities in physical capital are strongly influenced by the central government’s regional policies, yet the effectiveness of such policies requires comprehensive and coordinated consideration. These findings highlight the importance of factor decomposition in understanding the dynamics of economic disparities between northern and southern China. They also underscore its value as a tool for narrowing the North-South gap and fostering a regional development pattern characterized by complementary strengths. Of course, the North-South economic divide is a complex phenomenon driven by numerous factors. This study addresses the issue only from the perspective of growth factors, leaving substantial room for further research.

The report of the 20th National Congress of the Communist Party of China emphasized the need to “thoroughly implement the regional coordinated development strategy, major regional strategies, the functional zoning strategy, and the new urbanization strategy, optimize the layout of major productive forces, and establish a regional economic layout and territorial space system that features complementary strengths and high-quality development.” In our view, the conclusions of this study offer the following policy implications for promoting coordinated regional development and narrowing the economic development gap between northern and southern regions:

First, transforming regional coordinated development policies requires fostering new quality productive forces tailored to local conditions. Our analysis shows that TFP is a key determinant of regional economic growth and a primary driver of the North-South disparity. This necessitates that all regions develop new quality productive forces and rapidly enhance TFP. At the same time, we must recognize the vast differences in resource endowments and development levels across regions, which lead to varied priorities and challenges. Developing new quality productive forces cannot rely on a one-size-fits-all model; it must be adapted to local conditions. Different regions, being at different development stages, will exhibit distinct regional characteristics in this process. It is crucial to implement categorized guidance—selectively promoting new industries, models, and drivers based on local resource endowments, industrial foundations, and research capabilities. By focusing on core elements of sci-tech innovation and intensifying efforts to consolidate strengths, address weaknesses, and foster new advantages, each region can explore pathways that leverage and showcase its unique potential.

Second, achieving coordinated regional development requires deepening reform and expanding opening-up. Our research finds that advancing reform and opening-up remains crucial for achieving regional coordination. The northern region, in particular, should accelerate market-oriented reforms in property rights and factor allocation to improve the business environment. Key priorities include stimulating market vitality, supporting the development of small and medium-sized enterprises, and facilitating industrial upgrading. Simultaneously, all regions should further expand opening-up, intensify efforts to attract international and domestic capital, technology, and talent, strengthen inter-regional cooperation and exchange, and promote a positive pattern of complementary strengths and collaborative development.

Third, realizing coordinated regional development necessitates accelerating the establishment of a unified domestic and international market. Our study demonstrates that appropriate factor mobility does not lead to widened disparities in regional development levels; rather, it serves as an important means of narrowing gaps in output per worker. Therefore, it is essential to speed up the construction of a unified domestic and international market by breaking down regional administrative segmentation, removing barriers to factor mobility, and improving fiscal transfer payment mechanisms and ecological compensation mechanisms. Regions should fully acknowledge the objective laws of concentration of human, financial, and material resources in advantaged areas, further break down regional administrative divisions, comprehensively eliminate obstacles to factor mobility, and accelerate the formation of a unified, open, and competitively orderly national market for goods and factors.