Effects of Rising Wages in China on FDI Inflows into Sub-Saharan Africa
Maurice M. Amrani
University of Amsterdam, Faculty of Economics and Business Bachelor’s Thesis Business Administration
Drs. E. Dirksen August 2020
Statement of Originality
This document is written by Maurice M. Amrani who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
Sub-Saharan Africa (SSA) is one of the poorest regions globally, with many countries in the region still depending primarily on natural resources, not unlike China before 1979. An underdeveloped industrial sector has held back the region's economic development. However in recent years Foreign Direct Investment (FDI) into the region has seen a steady increase. It has been argued that as China starts running out of surplus labor, jobs may be relocated to low-income economies China trades with. In this paper it is proposed that this job shift is happening and that a positive relationship between wages in China and FDI in SSA exists. This paper looks at the relation between rising wages in China and FDI into SSA, and the possible determinants of this growth. A positive relationship between the wages in China and FDI inflow in SSA can be observed. However how this relationship is affected by income and by the natural resource endowments of specific economies remains unclear. What is clear is that Africa still has challenges to overcome before becoming a competitive alternative to China for manufacturing operations. However, the increasing costs in China should level the playing field considerably. It is concluded that this job shift can have far-reaching effects on the economies of SSA, if the right policy measures are adopted to fully take advantage of this opportunity.
Table of contents Introduction 5 Theoretical Framework 7 Methodology 17 Results 20 Discussion 25 References 30 Appendix 36
1. Introduction
Can Africa become the new China? The African continent holds some of the poorest countries in the world, and growth over the past decades has been slow. Especially in the Sub-Saharan part of Africa, many nations are still largely dependent on natural resource extraction and otherwise consist of mostly agrarian economies. While Foreign Direct Investment (FDI) into the region has seen an increase over recent years, especially from China, these investments are still largely aimed at extracting resources (UNCTAD, 2019). China, which had a similarly underdeveloped economy just 40 years ago, has seen a massive economic growth over the past few decades due to economic policy reforms and rapid industrialization. A big driver behind this growth has been the large amounts of FDI by multinational corporations (MNCs) looking to capitalize on the low labor costs in the country. China has over the years created an efficient and affordable manufacturing environment for MNCs. By utilizing its virtually unlimited rural labor pool, China started operating as the ‘factory of the world’, becoming the world’s leading manufacturer, producing anything from clothing to computers to automobile parts. However, with the country’s increases in wealth labor costs in China have been rising drastically as well (Fang & Yang, 2011; Yang, Chen & Monarch, 2010). China is slowly starting to run out of cheap labor. An aging population combined with better access to education is leading China’s unlimited labor advantage to run out. With labor costs rising, the MNCs settled in China may look to move manufacturing processes to different low-cost locations. The country may shed as much as 100 million manufacturing jobs (Lin, 2011). With wages rising steadily since the 1980’s, China’s labor advantage is slowing down. Research suggests the country may already have reached the ‘Lewis turning point’ (Lewis, 1954) where all surplus agrarian labor is
absorbed and industrial wages start rising rapidly. This presents new opportunities for developing countries like those in Sub-Saharan Africa to grow their industry.
The ‘Flying Geese’ model (Akamatsu, 1962) details how East-Asian countries have generally followed a pattern where underdeveloped countries benefit from the growth of larger economies after these reach a certain labor capacity. China may in this case be considered the ‘lead goose’ or ‘leading dragon’ (Lin, 2011) and its economic growth may be starting to spill over to the less developed countries China trades with. While not all socio-economic factors are directly comparable to China, there is cause to look into how these flying geese spillover effects are affecting the Sub-Saharan Africa (SSA) region. Historically SSA has known many socio-economic challenges restricting its development. On the other hand, FDI has been shown to have a positive effect on economic growth, China being a prominent example. If this proposed job shift is happening, the question now rises if SSA can benefit from this opportunity and if not, what challenges prevent Africa from becoming the next ‘factory of the world’?
By looking at the amount of FDI inflows into the SSA region over the years, this study tries to see if investments are boosted by the rapidly rising labor costs in China. The goal of this paper is to find if there is any correlation between rising wages in China and FDI into SSA, thus providing support for the flying geese theory of development. Furthermore, this paper aims to shed light on what factors may potentially moderate this relationship between rising wages and FDI into the SSA region. Specifically looking at manufacturing operations by MNCs and which factors may influence these firms’ decision to relocate activities abroad. Through this lens this study attempts to answer the question: how are FDI
In recent years research has been done into the potential of African countries as a manufacturing destination (Gelb, Meyer & Ramachandran, 2020; Lin, 2011; Ozawa & Bellak, 2011; Xu, Gelb, Li & Zhao, 2017). However, none of these recent studies have looked specifically at FDI inflows to SSA, or tried to correlate these with China’s rising wages. This study takes a quantitative approach, looking at the shift of FDI from China to Africa from China’s industrialization until the present. This research may increase our understanding of how developing nations could benefit from economic growth in maturing economies. Also, this research may increase support for existing economic development theory and our understanding of the factors that affect international corporations’ location decisions.
First I will discuss the existing literature on rising labor costs in China and the underlying causes. Then I will discuss the potential effects on Africa, and which factors play a role in MNCs’ investment decisions. Hereafter I will describe the quantitative methods used, and describe the data analysis done. Lastly, I will show the results of the analysis and discuss them and their implications for economic policy and future research, based on the literature.
2. Theoretical Framework
The literature review in this section is divided into three parts. First the developments of labor costs in China since the nation's economic reforms are looked at, together with the implications these have for developing economies. Then two factors are identified that may moderate these effects, namely the income level of the region, and dependency on natural resources. Low income is assumed to correlate with lower wages, and there is evidence that high dependency on natural resource exports may restrict non-resource FDI. Finally, other challenges to the development of SSA as a manufacturing
destination are discussed. While these can not all be statistically analyzed, it is important to take note of the different factors that may restrict future investments in SSA.
Rising Wages in China
Firstly looking at labor cost developments in China. Since the economic reforms in China in 1978 and onwards, Chinese wealth has steadily grown for the country to become the world’s second largest economy in terms of GDP (IMF, 2019), growing from a developing agrarian economy to an industrial powerhouse. A factor that played a big role in this growth was the large agrarian labor pool that, due to technological advancements in that sector, became available to work in low-skill industrialized sectors such as manufacturing and construction (Zhang et al., 2011). This large labor pool kept labor costs relatively low, and thus provided a favorable manufacturing location for MNCs. This ‘unlimited’ labor pool has been running out however, due to several demographic factors. Primarily an aging population and an increase in the level of education among the urban population have caused increased wage growth rapidly in recent years (Cui, Meng & Lu, 2018).
Zhang et al. (2011) argue that around 2003, China had reached the ‘Lewis turning point’. This concept, named after W.A. Lewis, indicates the point in time when a developing economy runs out of surplus labor, thus starting an accelerated increase in wage growth for low-skilled industrial work. Lewis (1954) described a dual-sector model for developing economies, with a traditional agricultural sector and a capital-intensive industrial sector. Initially the agricultural sector holds a large supply of surplus labor, partly in the form of ‘disguised’ unemployment. The industrial sector is expected to absorb this surplus labor as capital increases and technology advances. As the marginal productivity of this surplus labor is negligible or even negative, this results overall in an increase in industrial growth without an increase in wages. This process explains how China had industrialized, by utilizing its near
unlimited labor pool it could keep industrial wages low, attracting large amounts of FDI for labor-intensive industrial work. However, Lewis also stated that when all surplus agricultural labor has been absorbed, agricultural and low-skill industrial wages are expected to begin to rise (Lewis, 1954). Zhang et al. (2011) used wage survey data to find a significant increase in Chinese wages starting in 2003, indicating that China’s surplus labor has started to run out. Yang et al. (2010), who looked at the developments of wages across China, also found an increase in wages over that time. More specifically, they reported a sevenfold increase in real wages in China from 1978-2007. However up until 2007 this still remained a fraction of the cost in neighboring countries. They noted that manufacturing wage growth was still limited by the large pool of about 130-200 million laborers still available in and around cities at that time. Wei & Kwan (2018) argue that there are regional differences, and that the eastern and northeastern regions of China, that host by far the most industrial activity, had passed the Lewis point in 2010. The central and western regions still had a large surplus of agricultural laborers by then. Das & N’Diaye (2013) estimated that China will reach the Lewis turning point between 2020 and 2025, as the working age population is shrinking.
While the exact year that China reached or will reach the Lewisian turning point is debatable, it is clear that the country's labor shortages have fueled an increase in wages within the industrial sector. This is supported by data from the Chinese National Bureau of Statistics (NBS) showing average wages had more than doubled over the period 2009-2018 (NBS, 2019). Lewis (1954) mentions that when the turning point has been reached, immigration from surrounding countries and export of capital will keep the rising wages under control. This export of capital from China is expected to be the main way developing countries like those in SSA benefit from China’s increased labor costs. Additionally, the
result of these labor cost increases is that the many MNCs settled in China may look to invest in lower-cost countries to set up manufacturing operations.
Flying Geese Model
Lewis’ theory ties in to the Flying Geese theory of economic development (Akamatsu, 1962; Kojima, 2000), which describes how developing economies benefit from the industrial growth of a larger economy as rising labor costs in the latter force them to move away from labor-intensive manufacturing activities towards more capital intensive activities, a process that is currently observable in China. As a result, labor-intensive manufacturing jobs are redistributed to less developed nations the country trades with, leading to increased industrialization in those countries. The model, named after the V-formation geese commonly appear in, was first proposed by Akamatsu in the 1930’s and further developed in the 1960’s. Akamatsu noted how East-Asian countries had followed a pattern of import, domestic manufacturing and then export of commodities. Following this process, production of these goods was being passed on from the most developed to the lesser-developed countries throughout the region. This pattern had first been observed in Japan after World War II, with the country migrating its labor-intensive industries to Hong Kong, South Korea, Taiwan and Singapore. More recently Thailand, Malaysia, the Philippines and Indonesia followed, each utilizing their comparative labor advantage, followed by relocation of those sectors they had lost their advantage in (Kojima, 2000; Lin, 2011; Ozawa & Bellak, 2011). The main way transmission of industry takes place is through FDI by the larger economy, offshoring the production of goods it could no longer cost-effectively produce. This leads to an FDI-driven industrial growth in the receiving country, while at the same time promoting trade between the two countries (Kojima, 2000).
It is argued that this theory could be applied to current-day China and SSA (Lin, 2011; Ozawa & Bellak, 2011). China can be considered the ‘lead goose’, or ‘leading dragon’ here, and Sub-Saharan countries can be considered the underdeveloped ‘follower geese’ countries. Lin (2011), argues that China’s wage growth may free up as much as 100 million low-skill manufacturing jobs. It is expected then that the country, losing its labor cost advantage in labor-intensive industries such as manufacturing, will start relocating jobs to the less-developed economies in SSA.
As noted, FDI is the main way through which developed economies benefit developing countries (Kojima, 2000). FDI has played a major role in the rapid development of the Chinese economy as well. According to Chandra, Lin & Wang (2013) foreign investors helped identify growth sectors, provided advanced technology and reduced first mover risk and transaction costs when introducing new products to market. This led to the country quickly upgrading its industry. It is thus expected that FDI will be an indicator for future growth in African countries as well.
Although FDI into SSA has seen growth over the past decade, it is important to make the distinction between labor-seeking FDI and resource-seeking FDI. While both can be considered to be aimed at an export market, labor-seeking FDI in, for example, manufacturing has much more benefits for the host country (Kaplinsky & Morris, 2009; Ozawa & Bellak, 2011). The recent growth of FDI is mostly aimed at extracting the natural resources in SSA (UNCTAD, 2019), this creates far fewer jobs for the host country and, as will be discussed later, may negatively influence other types of FDI.
It should also be noted that while Chinese manufacturing firms alone have not
shown much interest yet in relocating towards Africa (Ozawa & Bellak, 2011; Xu, et al., 2017), less research has been done into MNCs from other nations investmenting in Africa.
As China’s manufacturing sector consists largely of subsidiaries of foreign MNCs, it can be expected that any increases in FDI into Africa would come not only from China, but also in large part from the MNCs settled there, with a variety of origins.
Thus, following the theory discussed above it is expected that the increase in Chinese wages will lead to increased capital investments in Sub-Saharan Africa, not only from China but from the many MNCs with manufacturing operations there as well. This brings forward the first hypothesis of this paper H1: Wages in China relate positively with FDI inflows into Sub-Saharan Africa.
Moderating Effect of Labor Cost
Now looking at what factors may influence these investment decisions. Instinct suggests that with mostly manufacturing jobs getting relocated (Lin, 2011), labor costs are a determining factor for where they are relocated. A report by Xu et al. (2017) surveyed a large number (n=640) of Chinese manufacturing firms, over half of which were foreign owned subsidiaries. Nearly half of all firms interviewed stated rising wages as their top challenge and many as their second or third, indicating that rising wages are a major concern for China-based manufacturers. The comparatively low wages and resulting labor costs in African countries are particularly attractive for labor intensive light manufacturing industries (LILM), for example textile manufacturing, shoes, electrical appliances and toys. Many of these are currently dominantly produced in China. The question that follows now is why companies then have been slow to invest into Africa as a manufacturing destination, even as wages in China, the globally leading manufacturing destination, have already started to rise.
Considering the importance of labor costs in firms’ location decisions in manufacturing, it would be expected that Sub-Saharan Africa is an ideal location for LILM
activities. However, this is not yet the case as Dinh and Clark (2012) found that the African manufacturing sector’s share of GDP was very small compared to most of the world, followed only by North Africa and the Middle East. While absolute wages in SSA may be low, several studies explain this apparent contradiction by the low productivity in African manufacturing. Gelb et al. (2020), looked at the potential of SSA as a manufacturing destination. They showed that for most SSA countries manufacturing labor cost is higher than potentially competing countries outside of SSA, when corrected for levels of GDP. In their research, only Ethiopia came out as potentially competitive in terms of manufacturing labor costs. Dinh and Clark (2012) found that a large part of firms in SSA are informal in nature. Informal firms were found to be less productive than formal firms in most countries in SSA, meaning that they account for a greater share of employment than they do of output. More specific to this research, Golub, Ceglowski, Mbaye & Prasad (2018) found that the relative unit labor costs (ULCs) for Sub-Saharan Africa were higher than China’s ULCs in 2010 even when wages were already rising. This was also explained to be mostly due to very low productivity of Sub-Saharan Africa relative to China and other Asian countries. They also point at poor infrastructure, weak institutions and political instability making African countries unattractive to foreign investment.
Still, this currently underdeveloped manufacturing sector is expected to grow increasingly as a result from the job shift from China. Considering that labor cost is a determining factor in low-skill labor manufacturing, it is expected that when labor costs are lower, investments from companies looking to relocate from China are higher. However, as actual wage data is difficult to obtain in SSA due to the highly informal nature of many firms, average GDP per capita is used as a proxy for labor costs. Thus, it is expected that any FDI increases due to the job shift from China should be greater when average earnings in
SSA, i.e. GDP per capita, are lower. Bringing forward this paper’s second hypothesis: the
relationship between wages in China and FDI is moderated by GDP per capita, such that the increase in FDI is higher when GDP per capita is lower.
Moderating Effect of Natural Resource Endowments
As mentioned earlier in this section many Sub-Saharan African countries are highly dependent on natural resource exports. Accordingly, foreign investments in the region seem to be aimed at exploiting these resources (UNCTAD, 2019). China especially has been investing heavily in extracting SSAs natural resources (Kaplinsky & Morris, 2009; Zafar, 2007). Somewhat contradictory, research by Asiedu (2013), suggests that abundant natural resource endowments may actually reduce FDI inflows. This is in line with the so-called ‘resource curse’ (Sachs & Warner, 2001), stating that countries with large natural resource endowments tend to grow slower than resource-poor countries. Most agree that a crowding-out of other economical activities, such as manufacturing due to a heavy dependence on natural resource exports plays a role in this (Asiedu, 2013; Sachs & Warner, 2001). Explanations for this phenomenon are not universally agreed upon (Mehlum, Moene & Torvik, 2006), but there is empirical evidence that a negative relationship exists between natural resource rents and non-resource FDI (Poelhekke & van der Ploeg, 2013). This indicates that when resource endowments are high, foreign investments in sectors such as manufacturing are reduced. Therefore it is expected that overall, when natural resource exports are higher, increased FDI as a result of rising wages in China will be lower. Thus the third hypothesis: The relationship between wages in China and FDI inflows in SSA is moderated by exports of natural resources, such that the increase in FDI is smaller when natural resources as part of total merchandise exports are higher.
Income and natural resource endowments are far from the only factors affecting location decisions for firms, especially when considering global value chains. Factors such as geographical location, ease of integration, infrastructure and several policy choices can be important determinants for FDI. Also, the flying geese model applies to all less-developed countries China trades with, and some of the described spillover effects could as well occur in some Southeast Asian countries like Vietnam, Cambodia and Bangladesh, which sometimes seem more preferable for companies (Ozawa & Bellak, 2011; Xu et al., 2017). These countries may then form strong competition for African destinations. Continuing on, this section discusses some of the factors potentially restricting investment by MNCs currently settled in China.
On the Chinese side, a survey by Xu et al. (2017) on a large sample of Chinese firms found that only a very small part (10%) of surveyed firms actually considered relocating, and an even smaller part towards Africa; all of those footwear firms moving to Ethiopia. Although the authors do note that the nature of the all-Chinese firm sample excludes firms that had already relocated entirely, and did not account for investment decisions made by parent companies (when over half of the companies interviewed were foreign owned subsidiaries). Still, this indicates that on the firm level many companies remain reluctant to relocate outside of China in the wake of rising wages.
Xu et al. (2017) state four factors that may keep companies from relocating to other countries: First, due to big regional differences in terms of wages, many companies may opt to move factories around within China itself. Most manufacturing operations are now located on the highly industrialized east coast, with the highest regional wages in the country (Yang et al., 2010). However, the central and western regions of China with far lower average wages are becoming more attractive to firms as labor costs rise. Domestic
relocation has the added benefits of existing infrastructure. In some cases this may even enable a firm to retain part of the existing workforce, as many laborers now active on the east coast are originally from these rural areas. Secondly, factors such as energy, water and communication infrastructure play an important role in location choice even in mostly labor-dependent industries. Over the years China has developed a very attractive business environment in terms of infrastructure, where other countries may yet have to develop these. Good infrastructure is also found to be an important determinant of FDI in SSA (Asiedu, 2006; Cleeve, 2012). Calderón & Servén (2010) found that infrastructure development in SSA positively influences long-term growth and reduces income inequality. Thirdly, as companies’ operations in China are often interdependent parts of a global value chain, the multinational ‘lead firms’ play a big role in an individual firms’ investment and location decisions. In addition, component manufacturers and assembly firms are often located together to reduce transaction costs for these firms. Fourthly, firms may choose not to relocate operations at all but instead invest in mechanization and automation technology as a response to rising wages (Xu et al., 2017).
Looking at firms that do decide to relocate abroad because of rising labor costs, there is strong competition on this front from developing Southeast-Asian countries such as Vietnam, Cambodia and Indonesia. These countries could benefit from the same spillover effects as Africa. (Xu et al, 2017; Ozawa & Bellak, 2011). Lin (2018) however, argues that although Southeast Asia has a geographical advantage, the massive population of Africa compared to these countries is needed to absorb all of China’s relocated jobs. Nonetheless, several factors on the African continent can be identified that may restrict investments. Cleeve (2012) found that traditional determinants of FDI like market size, and growth rate are important in SSA as well. However, most countries in the region score very low on these
measures globally. Other factors such as infrastructure and openness to investment are also found to be important determinants, and while not ideal in most SSA countries, these factors are more controllable by policy makers. Finally, political stability and strong institutions are found to attract more FDI. This is in line with Asiedu (2006) who finds that a good legal system and good infrastructure promote FDI, while political instability and corruption deter FDI. Furthermore, high corruption has been found to be deterring specifically to horizontal (market-seeking) FDI, possibly further skewing the amount of vertical (export-seeking) FDI into SSA (Hakkala, Norbäck & Svaleryd, 2008). In conclusion, while many determinants of FDI are outside the control of national policy makers, many endogenous factors can be influenced. This may be critically important now to compete with the other manufacturing destinations for the many jobs that China is expected to transfer.
3. Methodology Data
To test the hypotheses stated in the theoretical framework various publicly available datasets were used. For statistics on FDI inflows into Sub-Saharan Africa a dataset assembled by the United Nations Conference on Trade and Development (UNCTAD) for their annually published World Investment Report was used (UNCTAD, 2020). Wage data was retrieved from the Chinese National Bureau of Statistics (NBS) (NBS, 2019). GDP statistics were also retrieved from UNCTAD (2020). The percentage of natural resource exports as part of total merchandise exports was constructed from the World Bank database (2020) by combining the statistics for Agricultural raw materials exports, Food exports, Fuel exports and Ores and metals exports. Data availability for some variables was low for the SSA region. While annual FDI flow and GDP per capita statistics were fully available from 1970 to
2019, the natural resource exports data contained only 26 entries from 1975 to 2018. Chinese wage statistics were available from 1985 to 2019, and for the years 1978 and 1980. Variables
The dependent variable in this statistical analysis is the amount of FDI inflows in Sub-Saharan Africa, in current-day US dollars (FDI). FDI data was retrieved from a comprehensive dataset assembled by UNCTAD, consisting of global FDI flows per country and per region. UNCTAD regularly collects FDI data directly from central banks, statistical offices and national authorities, and from the International Monetary Fund (IMF) and Organisation for Economic Co-operation and Development (OECD) for its FDI database. For UNCTAD's full methodology see UNCTAD (2017).
The independent variable used in the statistical analysis is the average wage in China (WAGE), specifically the average wage of employed persons in urban units, in Chinese yuan. Urban units refer here to the work unit system in use in China (see Bjorklund, 1986). This variable was retrieved directly from the NBS. It is calculated as the total wage bill of employed persons at reference time divided by the average number of persons employed at reference time. The NBS notes that average wages from 1995 to 2008 refers to average earnings of employed persons in urban units, as a proxy for wages (NBS, 2020).
GDP per capita is used as a proxy for labor cost in SSA. This number has been retrieved from the UNCTAD database (2020). GDP statistics are presented in US dollars at 2015 prices, in millions.
The fourth variable in this model is natural resource exports, taken as a percentage of total merchandise exports (NATEXP). This variable signifies an economies’ natural resource endowment, after Asiedu (2006). Natural resource exports here are defined as
Agricultural raw materials exports, Food exports, Fuel exports and Ores and metals exports. This percentage has been retrieved from the World Bank database (2020).
Analytical Plan
The equation to be estimated is :
n(F DI) ₁(Ln(W AGE)) β₂(GDP /c) β₃(Ln(W AGE))(GDP /c)
L = α + β + +
β₄(NAT EXP ) β₅(Ln(W AGE))(NAT EXP )
+ +
To test hypothesis 1: the effect of Chinese wages on FDI flows, a linear regression will be utilized with FDI inflows into Sub-Saharan Africa as the dependent variable and Chinese average wage as the independent variable. The variables GDP per capita and NATEXP are entered as control variables. For hypothesis 2: The relationship between wages in China and FDI is moderated by GDP per capita, such that the increase in FDI is higher when GDP per capita is lower, the interaction effect between wages in China and GDP per capita in SSA is analyzed utilizing the PROCESS macro by Hayes (2020). For hypothesis 3: The relationship
between wages in China and FDI inflows in SSA is moderated by exports of natural resources, such that the increase in FDI is smaller when natural resources as part of total merchandise exports are higher, the interaction effect between wages in China and natural resource exports in SSA is analyzed, again by utilizing the PROCESS macro by Hayes (2020).
4. Results
For all statistical tests in this section a confidence interval of 95% was used, with a p-value cut-off of α = 0,05 to determine statistical significance, unless indicated otherwise. Table 1
Descriptive Statistics and Pearson’s correlations
Variable N Minimum Maximum Mean Std. Deviation 1. 2. 3. 4.
1. WAGE 37 615 90501 23114,2703 26142,8207 1 0,857** 0,846** 0,474* 2. FDI inflows (millions) 50 247,9838 45237,7325 12415,2849 14806,4331 0,857** 1 0,835** -0,002 3. GDP per capita 49 330,7703 1891,5823 981,8351 428,4371 0,846** 0,835** 1 0,203 4. NATEXP 26 68,49 82,4831 76,0971 3,7038 0,474* -0,002 0,203 1
** Correlation is significant at the 0,01 level (2-tailed). * Correlation is significant at the 0,05 level (2-tailed).
Table 1 contains the means, standard deviations of and correlations between all variables used in the analyses. Of note are the strong positive correlations between the WAGE variable and FDI inflows and between the WAGE variable and GDP per capita, both significant (p < 0,01). The former indicates some support for the first hypothesis. Additionally GDP per capita correlates strongly with both wages and FDI inflows, indicating some support for the second hypothesis. Natural resource export percentages do not correlate with FDI inflows, lending no support for hypothesis 3.
Regression Assumptions
Before performing a linear regression a number of assumptions must be met. The first assumption being that there is a linear relationship between the variables average wage in China and FDI inflows in SSA. To detect a linear relationship between the dependent
and independent variables a scatter plot is drawn. (Figure A1). A weak linear relationship between variables can be observed, but the line is not ideal.
A second assumption to confirm is that the values of residuals are independent. To test this assumption a scatterplot of standardized predicted values against residuals is drawn (Figure A2). The spread of observations in this scatterplot, in addition to a Durbin-Watson statistic of 0,387, do not indicate independence of residuals. Partly this is a result of time-series data, and this may have a large impact on the accuracy of the final analysis.
A third assumption for linear regression is the normality of residuals. To confirm normality of residuals, a P-P plot of standardized residuals is drawn (Figure A3). As residuals do not closely follow the normal plot line, normality of residuals can not be confirmed, thus this third assumption is violated.
The final assumption to test is that of homoscedasticity of residuals. To confirm this assumption the scatterplot of residuals against predicted values is consulted (Figure A2). From the spread of residuals in this plot we can not assume homoscedasticity. Heteroscedasticity is often apparent in time-series variables, however this does reduce the accuracy of the following analysis.
Due to the violation of several assumptions of linear regression the natural logarithms of the variables FDI and average wage were tested. Natural logarithmic variables were computed and tested against the same aforementioned assumptions. As shown in figure A4 these variables show a strong positive linear relationship. Figure A5 shows both greater homoscedasticity and greater independence of residuals, combined with an increased Durbin-Watson value of 0,758. While this is a considerable improvement, some autocorrelation in the residuals remains as a result of the nature of time-series data. Figure
A6 shows the normality of residuals is also improved when using logarithmic variables. While independence of residuals can still not be confirmed, there is an overall improvement over the non-logarithmic variables. Therefore it was decided to use the natural logarithms of the values of the FDI and average wage variables for this regression analysis.
Presence of multicollinearity between the focus variable and the control and moderator variables in the model was tested by looking at the Variance Inflation Factors (VIF) between them. The VIF for the variables wage and natural resource export were 3,109 and 2,425 respectively, indicating no strong collinearity. The VIF value for the variable GDP per capita reported 5,230, indicating some collinearity. While this requires some caution when interpreting the results of the analysis, the amount of collinearity is not extreme, and the nature of the samples suggests no direct link between the two variables.
Hypothesis 1
To test Hypothesis H1: Wages in China relate positively with FDI inflows in
Sub-Saharan Africa, a linear regression analysis was performed. The null-hypothesis tested
is H0: Average wages in China have no significant effect on FDI inflows in Sub-Saharan
Africa. FDI inflows into SSA are the dependent variable, average wages in China the
independent variable, and GDP per capita and natural resource exports in SSA are entered as control variables.
Table 2
Regression Model Summary
Model R R2 Adjusted R2 R2 Change F Change df1 df2 p-value
1 0,812 0,660 0,620 0,660 16,509 2 17 0,000
2 0,954 0,910 0,894 0,250 44,747 1 16 0,000
Model 1: NATEXP, GDP per capita
Table 2, model 1, shows a linear regression with only the control variables Natural Resource Exports and GDP per capita. Model 2 shows the regression with the natural logarithmic value of average wage in China added. Model 1 shows a significant R 2of 0,660, and an adjusted R 2of 0,620 compared to model 2 showing a significant R 2 of 0,910 and an adjusted R2of 0,894. This indicated respectively 66% and 91% of the variation in FDI can be explained by the models. Both models have a p-value below 0,05, and are statistically significant. Model 2 shows an R 2 change of 0,250 indicating that adding the average wage variable adds some significant explanatory power. The unstandardized B-coefficient of the wage variable is 0,744, with a t-statistic of 6,689 and a p-value below 0,05 (Table A1). This indicates that a 1% increase in wages in China, leads to a 0,74% increase in FDI inflows in SSA. In conclusion, based on these statistical tests support is found for Hypothesis 1, and the null-hypothesis can be rejected.
Hypothesis 2
To test Hypothesis 2: The relationship between the wages in China and FDI is moderated by GDP per capita, such that an increase in FDI is higher when GDP per capita is lower, the PROCESS macro by Hayes (2020) was used to find interaction effects between the variables WAGE and GDP per capita. The null-hypothesis tested is H0: GDP per capita has no
significant effect on the relationship between wages in China and FDI. The B-coefficient of the interaction effect between GDP per capita and WAGE is -0,0007, with a t-statistic of -2,1084 and p-value of 0,052 (Table 3). While the p-value is only slightly above the 0,05 cutoff value, taking into consideration the very small size of the interaction coefficient the null-hypothesis can not be rejected. The relationship between wage in China and FDI inflow in SSA is not significantly affected by GDP per capita.
Table 3 Regression Coefficients Model B-coefficient SE t p Constant 10,4122 2,4289 4,2869 0,0006 Ln(WAGE) 0,6136 0,1182 5,1934 0,0001 GDP per capita 0,0004 0,0003 1,2247 0,2396 Interaction effect -0,0007 0,0003 -2,108 0,0522 NATEXP -0,0053 0 ,0323 -0,1641 0,8719
Interaction effect: Ln(WAGE) x GDP per capita Hypothesis 3
To test the final hypothesis H3: The relationship between wages in China and FDI inflows in SSA is moderated by exports of natural resources, such that an increase in FDI is smaller when natural resources as part of total merchandise exports are higher, the
PROCESS macro by Hayes (2020) was used to find interaction effects between the variables WAGE and NATEXP. The null-hypothesis tested is H0: The amount of natural resource
exports as part of total exports have no significant effect on the relationship between wages in China and FDI. For the interaction effect between WAGE and NATEXP a B-coefficient 0f 0,0218 was found with a t-statistic of 0,6548 and a p-value of 0,523 (Table 4). The null-hypothesis is not rejected, the relationship between wages in China and FDI is not significantly affected by the export percentage of natural goods.
Table 4 Regression coefficients Model B-coefficient SE t p Constant 9,3481 0,4974 18,7931 0,0000 Ln(WAGE) 0,6588 0,1719 3,8330 0,0016 NATEXP 0,0010 0,0379 0,0267 0,9790 Interaction effect 0,0218 0,0332 0,6548 0,5225 GDP per capita 0,0003 0,0004 0,7993 0,436
Interaction effect: Ln(WAGE) x NATEXP
5. Discussion
This research aimed to find if and how FDI inflows into Sub-Saharan Africa are affected by an increase of wages in China. Following Lewis’ (1954) dual-sector theory of development and the flying geese model (Akamatsu, 1962) it was expected that capital investments into lesser developed countries China trades with, specifically those in SSA, would increase as China is starting to lose its labor cost advantage. As manufacturing jobs are getting relocated it is expected that those countries with the lowest incomes - and lowest labor costs - will benefit the most. Furthermore, as there is evidence for a negative relationship between natural resource endowments and non-resource FDI, it was expected that the effect would be lessened if natural resource exports formed a larger percentage of total exports. Through statistical regression it was attempted to confirm these expectations. The statistical results gave support only for the first hypothesis: there is a positive linear relationship between wage in China and FDI inflows in SSA . For the second hypothesis; the relationship between wage in China and FDI in SSA is affected by GDP per capita in SSA , no
GDP per capita on FDI. For the third hypothesis;the relationship between wage and FDI is negatively affected by higher natural resource exports, also no statistical support was found.
It was found that wage increases in China relate positively to FDI inflows in SSA. This gives some support to the theories of economic development discussed above for the SSA region. It also implies that for developing countries in the region it would be beneficial to prepare for increased investments. The manufacturing sector specifically should, according to the theory, see increased investments as wages in China keep rising. While anecdotal evidence suggests few China-based firms are currently willing to relocate (Ozawa & Bellak, 2011; Xu et al., 2017), pressure from rising labor costs were shown to be a concern. If wages keep rising in China, SSA countries have several improvements to make to further increase investment attractiveness. Development of infrastructure, reduced corruption and strong institutions are some areas predicted to increase FDI inflows (Asiedu, 2006; Cleeve, 2012).
No interaction effect was found between GDP per capita and wages in China on FDI flows. This may be due to the assumption that GDP per capita is correlated with labor costs being flawed. It can still be assumed that labor costs do play a role in the amount of FDI increases, especially for low-skilled manufacturing. Future research using more accurate labor cost data should give more insight into this effect. Alternatively, an explanation for the non-existent effect of GDP per capita on the increase in FDI could be that labor costs are not as big a factor for companies' location decisions as expected. There is some support for this in the fact that relatively few firms have actually relocated from China so far. Companies may prefer to lower costs in the current location by investing in technological innovation as Xu et al. (2017) found. Additionally, due to the low productivity in many African countries, as described by Golub et al. (2018), firms may consider unit labor costs and overall
production before labor costs. These factors are not captured in GDP per capita, especially considering the low employment rate in many SSA countries.
The lack of evidence for an interaction effect of natural goods exports could also be explained partially by the statistical method used. Looking at a per-country level rather than per-year may bring forward different results. A design of this scale was however beyond the scope of this research. Alternatively it could be that the crowding-out effect as seen in existing literature with investments in natural resources (Asiedu, 2013; Sachs & Warner, 2001) is reduced as wages are rising and more jobs are getting transferred to these developing countries. Companies may be more incentivized to invest in other activities than just resource extraction now that labor costs elsewhere are rising. However, little research has been done into the effect the ‘resource curse’ (Sachs & Warner, 2001) has on FDI in SSA specifically. While Poelhekke & van der Ploeg (2013) find that non-resource FDI is reduced in a resource boom, no theoretical explanations are given. Future research into this phenomenon should shed more light on the possible effects it may have on manufacturing FDI as a result of rising wages in China.
Limitations
The proposed statistical model seems to be reasonably strong, with an adjusted R 2of 89%. However, while the model is significant with a p-value below the 0,05 level, it does violate some assumptions for linear regression, which could have negatively affected the accuracy of the final results. Therefore the result of the statistical analysis can not be taken as conclusive. Furthermore, even when assumed to be valid, it is important to note that the relationship found between wages in China and FDI inflows is not a direct effect and should not be seen as such. Several factors beyond those examined in this research are likely to affect the relationship between these two variables. As discussed, labor costs are not the
single determining factor for investment decisions for most enterprises, and thus rising wages are not expected to directly affect FDI in low-wage countries. Infrastructure, political stability and several policy factors are just some other factors that have been shown to affect FDI inflows in SSA. The results do however indicate however that a certain relation exists, and this gives cause to further research to the determinants of FDI in SSA and how big the role of labor cost in other countries is comparatively.
While no interaction effect was found between GDP per capita and wages in China on FDI inflows, this does not conclude that wage rates do not play any role in the amount of FDI increases. The GDP per capita variable was chosen as a proxy for labor cost in SSA, due to lack of accurate wage statistics, assuming per capita earnings and wage correlate in SSA countries. This could however be a flawed assumption, considering the high income inequality in many SSA countries. Using more accurate wage or income data, and analyzing these on a per-country basis may improve these results.
As stated before, the statistical method used may not have been ideal for analyzing the effect that resource endowments have on FDI inflows. Data on a per country basis for a set time rather than a per year basis for the region may have given a more definitive conclusion.
Future research
Future research may look into more specific evidence that the proposed job shift from China is happening, for example by looking at the growth rate of manufacturing jobs specifically. Also a more specific look at different countries and wage levels in SSA could give more insight into what the main determinants and restricting factors are for job generation in SSA. As discussed above wage increases are not the only factor determining FDI inflows in
SSA, research into how far other endogenous factors such as the business infrastructure and policies play a role could be valuable for policy makers.
With more accurate wage data the role of labor cost in determining investment in SSA can be assessed more precisely, while also giving insight into which countries may benefit the most from the proposed job shift from China.
As investments into SSA are increasing, research into the ‘resource curse’ can be very valuable to African policy makers. How resource endowments affect economic development has implications for many currently underdeveloped economies, and the effects may become more apparent as investments in these countries are increasing.
Conclusion
In conclusion, how are FDI flows into Sub-Saharan Africa affected by rising labor
costs in China? A positive relationship between wages in China and FDI inflows in SSA was observed, which could indicate that a shift of jobs is happening. As wages in China are shown to be increasing, so are investments into SSA. While not all factors that together determine FDI increases in SSA have been identified yet, low wages alone can not be assumed to be the single determining factor. Productivity should be taken into account as well as many country-specific factors, such as political climate and infrastructure. Implications for policy makers in these economies are to invest in the right infrastructure and institutions to improve the business climate, to maximize value and make full use of the opportunities that this brings. China was able to grow from a poor country into an economic superpower in a few decades not only due to the right foreign investments, but also due to the right policy choices, now it is time for Africa.
References
Asiedu, E. (2006). Foreign Direct Investment in Africa: The Role of Natural Resources,
Market Size, Government Policy, Institutions and Political Instability. World Economy, 29(1), 63–77. https://doi.org/10.1111/j.1467-9701.2006.00758.x
Asiedu, E. (2013). Foreign direct investment, natural resources and institutions.
International Growth Centre.
Akamatsu, K. (1962). A historical pattern of economic growth in developing countries. The
developing economies, 1, 3-25.
Bjorklund, E. M. (1986). The Danwei: socio-spatial characteristics of work units in China's urban society. Economic Geography, 62(1), 19-29.
Bräutigam, D., & Tang, X. (2014). “Going Global in Groups”: Structural Transformation and China’s Special Economic Zones Overseas. World Development, 63(C), 78–91. https://doi.org/10.1016/j.worlddev.2013.10.010
Calderón, C., & Servén, L. (2010). Infrastructure and economic development in Sub-Saharan Africa. Journal of African Economies, 19(suppl_1), i13-i87.
Chandra, V., Lin, J. Y., & Wang, Y. (2013). Leading dragon phenomenon: New opportunities for catch-up in low-income countries. Asian Development Review, 30(1), 52-84.
Cui, Yuming, Jingjing Meng, and Changrong Lu. “Recent Developments in China’s Labor Market: Labor Shortage, Rising Wages and Their Implications.” Review of
Development Economics 22.3 (2018): 1217–1238. Web.
Das, M., & N'Diaye, M. P. M. (2013). Chronicle of a decline foretold: Has China reached the Lewis turning point? (No. 13-26). International Monetary Fund.
Dinh, H., & Clarke, G. (2012). Performance of Manufacturing Firms in Africa : An Empirical Analysis (pp. xxi, 211). Washington, DC: World Bank.
https://doi.org/10.1596/978-0-8213-9632-2
Eifert, B., Gelb, A., & Ramachandran, V. (2008). The cost of doing business in Africa: Evidence from enterprise survey data. World development, 36(9), 1531-1546.
Fang, C., & Yang, D. U. (2011). Wage increases, wage convergence, and the Lewis turning point in China. China economic review, 22(4), 601-610. Retrieved from
https://doi.org/10.1016/j.chieco.2011.07.004
Gelb, A., Meyer, C. & Ramachandran, V. (2016). Does Poor Mean Cheap? A Comparative Look at Africa’s Industrial Labor Costs. Revue d’économie du développement, vol. 24(2), 51-92. Retrieved from
https://www.cairn.info/journal-revue-d-economie-du-developpement-2016-2-page-51.htm.
Gelb, A., Ramachandran, V., Meyer, C. J., Wadhwa, D., & Navis, K. (2020). Can Sub-Saharan Africa Be a Manufacturing Destination? Labor Costs, Price Levels, and the Role of Industrial Policy. Journal of Industry, Competition and Trade, 1-23.
Golub, S. S., Ceglowski, J., Mbaye, A. A., & Prasad, V. (2018). Can Africa compete with China in manufacturing? The role of relative unit labour costs. The World Economy, 41(6), 1508-1528.
Hakkala, K. N., Norbäck, P. J., & Svaleryd, H. (2008). Asymmetric effects of corruption on FDI: Evidence from Swedish multinational firms. The Review of Economics and Statistics, 90(4), 627-642.
International Monetary Fund. (2019). World Economic Outlook, October 2019: Global Manufacturing Downturn, Rising Trade Barriers. Retrieved from
https://www.imf.org/en/Publications/WEO/Issues/2019/10/01/world-economic-outl ook-october-2019
Kaplinsky, R., & Morris, M. (2009). Chinese FDI in Sub-Saharan Africa: engaging with large dragons. The European Journal of Development Research, 21(4), 551-569.
Kojima, K. (2000). The “flying geese” model of Asian economic development: origin,
theoretical extensions, and regional policy implications. Journal of Asian Economics, 11(4), 375-401.
Lewis, W. A. (1954). Economic Development with Unlimited Supplies of Labour. The Manchester School, 22(2), 139-191.
Lin, J. Y. (2011). From flying geese to leading dragons: New opportunities and strategies for structural transformation in developing countries. The World Bank.
Lin, J. Y. (2018). China’s Rise and Opportunity for Structural Transformation in Africa. Journal
of African Economies, 27(suppl_1), i15-i28.
Mehlum, H., Moene, K., & Torvik, R. (2006). Institutions and the Resource Curse*. Economic
Journal, 116(508), 1–20. https://doi.org/10.1111/j.1468-0297.2006.01045.x
NBS (National Bureau of Statistics of China) (various years, 2009, 2018, 2019) China Statistical Yearbook. Beijing: China Statistics Press.
Ozawa, T., & Bellak, C. (2011). Will the World Bank's vision materialize? Relocating China's factories to sub-Saharan Africa, flying-geese style. Global Economy Journal, 11(3), 1850236.
Poelhekke, S., & van der Ploeg, F. (2013). Do Natural Resources Attract Nonresource FDI?
Review of Economics and Statistics, 95(3), 1047–1065.
https://doi.org/10.1162/REST_a_00292
Sachs, J. D., & Warner, A. M. (2001). The curse of natural resources. European economic
review, 45(4-6), 827-838.
UNCTAD. (2019). World Investment Report 2019:Special Economic Zones. https://unctad.org/en/PublicationsLibrary/wir2019_en.pdf
UNCTAD. (2020). World Investment Report 2020: International Production Beyond the Pandemic. https://unctad.org/en/PublicationsLibrary/wir2020_en.pdf
UNCTAD. (2017). Methodological note: World Investment Report 2017.
https://unctad.org/en/PublicationChapters/wir2017chMethodNote_en.pdf
UNCTAD. (2019, June). Annex table 1. FDI inflows, by region and economy, 1990-2018. Retrieved from
https://unctad.org/Sections/dite_dir/docs/WIR2019/WIR19_tab01.xlsx
Wei, Z., & Kwan, F. (2018). Revisit China’s Lewis Turning Point: An Analysis from a Regional Perspective. Asian Economic Journal, 32(4), 333–357.
https://doi.org/10.1111/asej.12162
Xu, J., Gelb, S., Li, J., & Zhao, Z. (2017). Adjusting to rising costs in Chinese light manufacturing: What opportunities for developing countries. Center for New
Structural Economics. Retrieved from
https://set.odi.org/wp-content/uploads/2017/12/SET_Survey- report_Chinese-manufacturing_Final.pdf
Yang, D. T., Chen, V. W., & Monarch, R. (2010). Rising wages: Has China lost its global labor advantage?. Pacific Economic Review, 15(4), 482-504.
Zafar, A. (2007). The growing relationship between China and Sub-Saharan Africa: Macroeconomic, trade, investment, and aid links. The World Bank Research
Zhang, X., Yang, J., & Wang, S. (2011). China has reached the Lewis turning point. China
Appendix
Figure A1
Figure A2
Scatterplot of Standardized Residuals and Predicted values
Figure A3
Figure A4
Figure A5
Scatterplot of Standardized Residuals and Predicted Values Using Natural Logarithms of Variables
Figure A6
Table A1
Regression Coefficients
Model Unstandardized B Coefficients Std. Error Standardized Coefficients Beta t p 1 Constant 16,800 4,050 4,148 0,001 GDP per capita 0,002 0,000 1,028 5,432 0,000 NATEXP -0,128 0,058 -0,419 -2,211 0,041 2 Constant 2,590 3,017 0,858 0,403 GDP per capita 0,000 0,000 0,100 0,587 0,565 NATEXP -0,006 0,036 -0,020 -0,174 0,864 Ln(WAGE) 0,744 0,111 0,882 6,689 0,000
