115urn:lsid:arphahub.com:pub:32e1b97d-7003-598d-92e7-0ceb44416cc9BRICS Journal of Economicsbrics-econ2712-77022712-7508Faculty of Economics, Lomonosov Moscow State University10.3897/brics-econ.3.e9567295672Research ArticleDoes Okun’s Law Hold for China? Some Empirical EvidenceSinghManharmanhar_singh@outlook.comhttps://orcid.org/0000-0001-5704-68251NurudeenIbrahim2Former M.PhiL. student, Department of Economics, University of Madras, Chennai (India)University of MadrasChennaiIndiaDepartment of Economics, Shehu Shagari University of Education, Sokoto (Nigeria)Shehu Shagari College of EducationSokotoNigeria
202208122022331731823E2AEE9E-07AE-5F58-81DD-42DC66C0EECE2609202227112022Manhar Singh, Ibrahim NurudeenThis is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits to copy and distribute the article for non-commercial purposes, provided that the article is not altered or modified and the original author and source are credited.
This paper seeks to estimate the applicability of Okun’s law to the situation in China between 1991 and 2020. A defining and most significant feature of this paper is that China’s unemployment rate has been proxied by youth unemployment and urban unemployment. The stochastic properties test reveals that all the three variables follow I(1) process. The paper uses this knowledge to build data generating process (DGP), which is an outstanding contribution to international research into the steady state growth. Many researchers have pointed out that the poor countries catch up faster and, consequently, their growth rate should have a trend component to it. The applied regression model has proxied the trend when estimating the operation of the Okun’s law. The inclusion of trend, strongly factual, is accounted for and reveals that Okun’s law is valid for China. Apart from the OLS estimator for testing the Okun’s law, the generalized method of moment estimator has also been used as another estimator with the first lag of both unemployment and GDP as instrumental variables. Empirical evidence supports the proposition that Okun’s law is indeed valid in the case of China contrary to the conclusion of some studies.
Singh, M., & Nurudeen, I. (2022). Does Okun’s Law Hold for China? Some Empirical Evidence. BRICS Journal of Economics, 3 (3), 173–182. https://doi.org/10.3897/brics-econ.3.e95672
Introduction
The underlying principle of the Keynesian revolution is that there always exists an excess capacity that should be used to create additional employment by pumping up aggregate demand. The government’s role is to find ways of using this excess capacity (Klein, 2016). Active involvement of government in economic activity would inevitably lead to the creation of greater employment opportunities, which in turn should spur growth.
The rate of output growth in the Chinese economy since its liberalization has been around 10 percent. In the year 2014, China overtook the United States as the largest economy in terms of Purchasing Power Parity and is expected to surpass it in nominal terms by the end of this decade (Morrison, 2019; Lin, 2020). The question of factors and policies behind such a remarkable transformation has occupied the attention of economists over the last few decades. Stripped to its essentials, the excess capacity existed in China, due to the country’s massive size, huge population and inadequate infrastructure; since 1978, the opening up of many sectors and expanding manufacturing bases were some of the key elements that provided the impetus for growth through efficient exploitation of the excess capacity organized by the Chinese policymakers. Most of the excess capacity, that has been mentioned, has been utilized. It is important to point out, however, that in 2014 growth in China started to decelerate (Lanteigne, 2015). Does it mean that the dwindling excess capacity will lead to reduction in the rate of employment generation in China?
The negative relationship between output and cyclical unemployment, first documented empirically by Arthur Melvin Okun in 1962, is known as Okun’s law (Guisinger et al., 2018; Huang et al., 2020). By his assessment, a growth rate of two percentage points would lead to one percentage point decrease in the unemployment rate (Okun, 1962). A key parameter, the nature of the economic growth in terms of productivity, is always shifting; it is expected that the underlying relationship between unemployment and growth would also change.
The answer to the above question will therefore depend on the nature and strength of the Okun’s coefficient in the case of China. Many factors might be involved in holding out the relationship between growth and employment, for instance, the role played by the labor market institutions1, but in China it is the urbanization rate that has been the major driver of growth since the beginning of liberalization process. The speedy growth of the Chinese economy is mainly due to its rapid urbanization (Wang et al., 2018). Since the country liberalized its economy, the Chinese Communist Party has given highest priority to urbanization, as they consider urbanization to be central to economic growth. The rate of urbanization in China was around 18 percent in the year 1978 and this increased to 40 percent in 2003 (Chen, 2007). More than 60 percent of the population in China now lives in urban settings (Wang et al., 2021).
This study builds on the work of the previous studies (Huang, 2003; Ball et al., 2012; Gazi & Prieto, 2016; Feng, 2018; Abubakar & Nurudeen, 2019). The underlying conclusion is that the relationship between unemployment and rate of economic growth is stable and positive. In 2021, China was the only major economy that did not witness recession, in spite of being in a lockdown in some parts of the country. The unemployment rate, globally, increased drastically. During the pandemic, the underlying relationship between growth and unemployment has become a matter of intense speculation but the question of Okun’s law needs to be explored.
From the political standpoint, the character of the relationship between growth and employment in the case of China is rather unique. Generating great employment opportunities in a democratic country is the basis on which many elections are held and often won. For the general public, having means and opportunities to get employment is of paramount concern. In a democracy, the political party that can do a better job at delivering such an outcome wins the election. As a one-party rule, the legitimacy of the Chinese Communist Party hinges on the existing positive linkages between growth and employment. In this context, the paper contributes to revealing the true relationship between output and rate of unemployment in the Chinese economy by exploiting the stationarity properties of the output that has grown exponentially. It is sometimes difficult for policy makers to control or exert the necessary degree of influence on the rate of unemployment. As they need to determine the degree of control over economy, the relationship between growth and unemployment rate is undeniably important from policy perspective.
The rest of the paper is as follows. Section 1 presents the literature review, which summarizes the previous theoretical and empirical research. The methodology is explained in Section 2. Section 3 discusses the results obtained in the course of the research. Then a conclusion is drawn and future research agenda proposed in the Conclusions section.
1. Literature review1.1. Approaches to the estimation of the Okun’s law
There are two approaches to the estimation of the Okun’s law. The first presupposes the use of level series, also called the GAP model. The second approach involves the first difference of both series (Karadzic et al., 2022; Porras & Martín-Román, 2022). It is important to point out that unemployment can also be an explanatory variable and output or income can be a dependent variable in both cases. To understand the level version of the model, consider equation (1) below,
(1)
where U and U* are the realized unemployment rate and the natural unemployment rate. Similarly, Y and Y* are the actual output that is produced and potential output or natural level of output. In the same vein, to understand the differenced version of the model, consider equation (2), shown below,
∆U_{t} = ξ(∆Y) (2)
where ∆ is the change or a differenced series. By the rule of thumb, a 1 percent increase/decrease in unemployment will lead to 2 percent increase/decrease in output. In other words, 1 percent increase in unemployment will lead to a deviation of GDP from its long run path.
1.2. Empirical literature review
The empirical evidence for the existence and validity of the Okun’s law in the case of Chinese economy is mixed. Several socio-economic reasons, unique to the Chinese economy, are perhaps the reason why the Okun’s law is not found to be holding in some empirical studies (Yin & Zhou, 2010; Wu-Liu, 2012).
In developing countries, the higher rate of growth of the country leads to an increase in the rate of urbanization, which in turn gets captured by the rate of growth of registered and formal employment. By many reckonings, China is still a developing country but has seen rapid and vast migration of people from rural to urban areas in the last two decades. Nonetheless, the level of urbanization is nowhere near the level that prevails in advanced countries. The use of the Okun’s law without factoring in the rate of urbanization is bound to lead to the conclusion that the relationship between (un)employment and growth is not holding in the case of China. The original Okun’s law is perhaps not so applicable in the case of China where the level of urbanization is still very low compared to advanced economies (Liu et al., 2018). The empirical evidence by Xuan (2003) has shown this in the case of China by constructing index of employment for three different sectors; still the revised Okun’s law does not seem to hold out in the case of China.
Yin and Zhou (2010) also show that the Okun’s law does not work for China. They argue that high employment in China could well be down to the fact that there are labor-intensive industries in the Chinese economy that anyways employs large quantities of labor. The Chinese labor market is dependent on economic growth in the countries to which it exports its merchandise. In other words, economic upturn in the United States appears to be more impactful in relation to the employment rate in China. Yin and Zhou (2010) came up with so-called Generalized Okun’s law that they count more applicable in the case of China, as compared to the standard Okun’s law more applicable to the case of developed countries.
It is essential to point out that the Chinese economy started to slow down in 2014. Since that time, no large-scale increases in employment have been observed. Basically, it is the rural to urban migration that is affected as a result of slowing growth rather than the jobs in the urban China. In the short-term, the transfers of labor from agricultural areas are correlated significantly with the economic cycles (Feng, 2018).
On the other hand, Peijiang (2014) was able to show that a 1 percentage point increase in urban registered unemployment leads to a 1.98 percent decrease in GDP, which means that the Okun’s law is valid for the Chinese economy. Yet, the paper was cautious about interpreting the Okun’s coefficient because there is a tendency that it might overestimate the impact of registered unemployment rate on GDP. One of the main fundamental reasons for the Chinese growth story was the presence of demographic dividend. The number of people of the working age started to decrease from the year 2015 (Li et al., 2017). Therefore, the central issue is whether growth is translating into commensurate increase in the rate of employment in China or not.
The empirical literature on the inclusion of youth unemployment in the Okun’s law in BRICS countries is available for India. The study that empirically examined the relationship between unemployment and output in India confirmed that although Indian output was growing, the growth was jobless because it was not up to the threshold level that ensures a decline in unemployment (Abubakar & Nurudeen, 2019). There is some evidence that the youth employment rate is the most affected during downturns since the early entrants to the job markets are the first to lose their jobs when there is a slump in economic activity.
However, before the results that has been estimated is put forth, it is useful to reveal the relationship between growth and employment in China.
A novel attempt, from the methodological standpoint, was made by including the trend in the output growth. The growth theory according to Solow (1956) provides a realistic theoretical explanation for the catching up process. This has been taken into account by the inclusion of trend component. That needs to be seriously investigated in the case of China since it demonstrates a classical and prototype example of the catching up process.
2. Methodology
The growth rate of the nominal GDP has been used as a proxy for the output. Both the youth unemployment and the urban unemployment rate were used for estimating the Okun’s law. The data was extracted from the economic research by the Federal Reserve Bank of St. Louis and all the series spanned from 1991 to 2020. The Augmented Dickey Fuller test (Dickey & Fuller, 1979) henceforth, (ADF) and Dickey Fuller generalized least square test (Elliott et al., 1996; Leybourne et al., 1998), henceforth DF-GLS, have been used in order to test the stochastic properties of the series. The specification of the ADF (1979) has been shown in (3).
(3)
The series will be considered to be stationary if the probability value is negative and less than 5 percent. Elliott et al., (1996), introduced DF-GLS test in order to address some of the disadvantages found in KPSS (1992). Therefore, the test is used in order to validate the finding of the ADF (1979). This test assumes two components in the model given in equation (4) below,
Y_{t} = ad_{t} + U_{t} (4)
where U_{t} = U_{t}_{– 1} + e_{t} and e_{t} follow a normal distribution and ad_{t} is thus the deterministic component of the model. The null hypothesis will be accepted in favor of unit-root if a = 1 and if a < 1 then the alternative hypothesis is accepted. Using both these tests would help to determine the stochastic properties of the series.
The estimation of the relationship between output and unemployment was conducted using the OLS estimator. The specification of the regression used to estimate the parameters of the constant, represented by a_{1}, trend which is represented by a_{2} has been shown in (5).
Y_{t} = a_{1} + a_{2}t + a_{3}X_{t} + e_{t} (5)
Where Y_{t} is the dependent variable and X_{t} is the independent variable, a_{1} and a_{3} measures the average and slope of the relation, and a_{2} is the deterministic component of the model while e_{t} is a white noise assumed to be independent and identically distributed with zero mean and a constant variance. The Generalized Method of Moment estimator (GMM) was also used. The first lag of both the unemployment rate and GDP were used as instrumental variables.
3. Results and discussion
Table 1 captures the estimates of the stochastic properties of each series. The estimated t-statistic has been compared with the tabulated values at a given level of significance. The unemployment rate and youth Unemployment are stationary at first difference. Similarly, the GDP series is also first difference stationary and it is significant at 5 percent level of significance. For the DF-GLS test, the results shows that all the variables are first difference stationary. Essentially, the results of the DF-GLS seem to support the evidence of the ADF. The main takeaway from here is that all the variables become stationary at first difference.
Stochastic Properties of the Variables
Variables
ADF Statistics
DF-GLS
Decision
Level
First Difference
Level
First Difference
Unemployment
-2.437
-3.295***
-2.309
-3.359**
I(1)
GDP
-1.871
-4.122***
-2.048
-3.453**
I(1)
Youth Unemployment
-1.339
-3.950***
-2.131
-3.918*
I(1)
Source: Computed by the Authors.
The estimated linear regression model has been presented in table 2, shown below. The underlying assumption of the data generating process (DGP) favors intercept and trend for all the series in the unit-root test. The regression model included intercept as well as the trend in the estimation. The regression model is therefore specified and estimated using equation (3) above. The estimation testifies the significance of both constant and trend. Keeping in line with the objective of the paper, GDP was taken as the endogenous series to find out if the GDP’s targeted growth of 8.5 percent can be achieved given the World Bank’s position. This is firstly assessed by a linear regression model using an ordinary least square (OLS) estimator. The results have been presented in table 2, shown below. Additionally, the generalized method of moment (GMM) estimator was employed in order to check and assess the robustness of the regression result, reported in table 3. The regression result shows that 1 percent increase in unemployment leads to 3.96 percent decrease in Chinese GDP. This is further assessed with a GMM estimator, and it shows quantitatively, that 1 percent increase in unemployment will decrease Chinese GDP by 4.68 percent; both OLS and GMM estimates are significant at 1 percent level of significance. Given the targeted growth of GDP by 8.5 percent pronounced by the World Bank, our regression model showed that the rate of unemployment would decline by 0.05 percent; this implies that unemployment would turn out to be 4.95 percent by the end of 2021 considering 5 percent unemployment rate existed in 2020. While our GMM model showed that 8.5 percent growth rate would instead increase unemployment by 1.591 and the rate of unemployment would turn out to be 6.59 by the end of 2021 fiscal year. Youth unemployment for the ages of 16-24 is higher in China than for the age group beyond 24 years.
Regression Model Estimates
Variables
Coefficient
Standard Error
T-Statistics
Probability Value
Constant
8.72
1.26
6.920
0.0001
Trend
8.35
3.84
21.72
0.0001
Unemployment Rate
-3.96
4.31
-9.20
0.0001
Source: Computed by the Authors.
GMM Model Estimates for the Overall Unemployment
Variables
Coefficient
Standard Error
T-Statistics
Probability Value
Constant
1.05
1.83
5.74
0.0001
Trend
9.05
5.05
17.92
0.0001
Unemployment Rate
-4.68
5.97
-7.84
0.0001
Source: Computed by the Authors.
Conclusion
The estimate of the regression points to Okun’s law holding in the case of China. The inclusion of trend, for which there is a strong theoretical justification due to Solow (1956), does indeed seem to make a difference in the case of China when it comes to analyzing the relationship between employment and output. The evidence seems to point to the existence of firm relationship between employment and output. Apart from the OLS estimator for testing the Okun’s law, the generalized method of moment estimator was also used as the other estimator where the first lag of both the unemployment and the GDP were used as the instrumental variables. Empirical evidence supports the conclusion that Okun’s law is holding in the case of China contrary to the results of some studies. This paper incorporates another version of the generalized version of the Okun’s law, with the trend of the output growth; it is more applicable to emerging or developing market economies, which will catch up faster with developed countries. One can propose two promising lines of future research. Firstly, the behavior of the Okun’s coefficient over the course of the business cycle can be explored for the BRICS countries. The other line of research assumes that the Okun’s coefficient BRICS countries changed after the pandemic.
Regression Model Estimates for Youth Unemployment
Variables
Coefficient
Standard Error
T-Statistics
Probability Value
Constant
8.22
1.25
6.55
0.0001
Trend
9.78
5.71
17.12
0.0001
Unemployment Rate
-2.01
2.30
-8.79
0.0001
Source: Computed by the Authors.
GMM Model Estimates for Youth Unemployment
Variables
Coefficient
Standard Error
T Statistics
Probability Value
Constant
1.05
1.83
5.745
0.0001
Trend
9.05
5.05
17.92
0.0001
Unemployment Rate
-4.68
5.97
-7.842
0.0001
Source: Computed by the Authors.
ReferencesAbubakarJ.NurudeenI. (2019) . Economic growth in India, is it a jobless growth? An empirical examination using Okun’s law. The Indian Journal of Labour Economics, 62 (2), 307-317. https://doi.org/10.1007/s41027-019-00165-wAnZ.GhaziT.PrietoN.G. (2016) . Okun’s law: unfit for low and lower middle-income countries? International Monetary Fund. https://www.imf.org/external/np/seminars/eng/2016/GlobalLaborMarkets/pdf/Ghazi_Session1_paper.pdfBallL.LeighD.LounganiP. (2017) . Okun’s law: Fit at 50? Journal of Money, Credit and Banking, 49 (7), 1413-1441. https://doi.org/10.1111/jmcb.12420ChenJ. (2007) . Rapid urbanization in China: A real challenge to soil protection and food security. Catena, 69 (1), 1-15. https://doi.org/10.1016/j.catena.2006.04.019DickeyD. A.FullerW. A. (1979) . Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74 (366a), 427-431. https://doi.org/10.1016/j.catena.2006.04.019ElliottG.RothenbergT. J.StockJ. H. (1996) . Efficient Tests for an Autoregressive Unit Root. Econometrica, 64 (4), 813-836.FengL. (2018) . The Labor Market and China’s Economic Cycle: A Discussion of Okun’s Law in Relation to China. Social Sciences in China, 39 (4), 53-78. https://doi.org/10.1080/02529203.2018. 1519223GuisingerA. Y.Hernandez-MurilloR.OwyangM. T.SinclairT. M. (2018) . A state-level analysis of Okun’s law. Regional Science and Urban Economics, 68, 239-248. https://doi.org/10.1016/j.regsciurbeco.2017.11.005HolmlundB. (2014) . What do labor market institutions do? Labour Economics, 30, 62-69. https://doi.org/10.1016/j.labeco.2014.06.007HuangG.HuangH. C.LiuX.ZhangJ. (2020) . Endogeneity in Okun’s law. Applied Economics Letters, 27 (11), 910-914.KaradzicV.BackovicT.StreimikisJ. (2021) . Relevance of Okun’s Law in Montenegro. Montenegrin Journal of Economics, 17 (1), 59-69. https://doi.org/10.14254/1800-5845/2021.17-1.4KleinL. R. (2016) . The Keynesian revolution. Springer.LanteigneM. (2015) . Chinese foreign policy: an introduction. Routledge.LeybourneS. J.MillsT. C.NewboldP. (1998) . Spurious rejections by Dickey–Fuller tests in the presence of a break under the null. Journal of Econometrics, 87 (1), 191-203.LiH.LoyalkaP.RozelleS.WuB. (2017) . Human capital and China’s future growth. Journal of Economic Perspectives, 31 (1), 25-48. https://doi.org/10.1257/jep.31.1.25LinJ. Y. (2020) . China’s economic outlook against the backdrop of the COVID-19 pandemic and US-China tensions. Journal of Chinese Economic and Business Studies, 18 (4), 327-331. https://doi.org/10.1080/14765284.2020.1854024LiuX.LamR.SchipkeA.ShenG. (2018) . A generalized Okun’s Law: Uncovering the myth of China’s labor market resilience. Review of Development Economics, 22 (3), 1195-1216. https://doi.org/10.1111/rode.12379MorrisonW. M. (2019) . China’s economic rise: History, trends, challenges, and implications for the United States. Current Politics and Economics of Northern and Western Asia, 28 (2/3), 189-242. http://dehai.org/dehai/assets/dehai/2021/370555-20150911_RL33534_7d506d97ac04bd17c079bd8791c8dfee1da3eb65.pdfOkunA. M. (1962) . Potential GNP: its measurement and significance. In: Proceedings of the business and economic statistics section of the American Statistical Association, Alexandria, VA: American Statistical Association (pp. 89–104).PeijiangZ. (2014) . Has Okun’s Law Become Invalid in China? Institute of Economics and Management, Southwestern University of Finance and Economics, Chengdu, China. https://m.sohu.com/a/207325304_673573/?pvid=000115_3w_PorrasM. S.Martín-RománÁ. L. (2022) . The heterogeneity of Okun’s law: A metaregression analysis. GLO Discussion Paper, No. 1069, Global Labor Organization (GLO), Essen.SolowR. M. (1956) . A contribution to the theory of economic growth. The Quarterly Journal of Economics, 70 (1), 65-94.WangJ.CaoS. J.YuC. W. (2021) . Development trend and challenges of sustainable urban design in the digital age. Indoor and Built Environment, 30 (1), 3-6. https://doi.org/10.1177/1420326X20976058WangQ.SuM.LiR. (2018) . Toward to economic growth without emission growth: The role of urbanization and industrialization in China and India. Journal of Cleaner Production, 205, 499-511. https://doi.org/10.1016/j.jclepro.2018.09.034Wu-LiuZ. H. A. N. G. (2012) . Dynamic Okun’s Law: Test and Comparison with Chinese and US Data. Journal of Guizhou University of Finance and Economics, 30.XuanZ. W. H. (2003) . Okun’s Law and China’s Unemployment. World Economy, 6. http://en.cnki.com.cn/Article_en/CJFDTotal-SJJJ200306005.htmYinB.ZhouJ. (2010) . China’s High Economic Growth and Low Employment: Empirical Study of Okun’s Law in China. Finance and Economics, 262 (1), 56-61.ZhangJ.WuJ. (2018)The Chinese labor market, 2000–2016. IZA World of Labor, 437. https://doi.org/10.15185/izawol.437
Bertil Holmlund (2014) provides an excellent review of how the labor market institutions affects the rate of unemployment.