Economic growth and height of youth: a panel data analysis on 27 provinces in China, 1985– 2014

Background: This study aimed to empirically examine the influence of China’s macroeconomic development on the height of Chinese youth in the past 30 years, using provincial panel data collected from more than one million children. Methods: Panel data from seven longitudinal surveys (1985, 1991, 1995, 2000, 2005, 2010, and 2014) of the Chinese National Surveys on Students’ Constitution and Health, including students aged 7–22 years from 27 provinces, were utilized for data analysis. Fixed-effects models were used to estimate the association between economic growth and height. Results: For every 1% increase in per capita disposable income (PCDI), the average height of students significantly increased by approximately 0.009%. Stated another way, this implies that a doubling of income is associated with 0.9% increase in height, which is significant for height. The coefficient of PCDI in the last decade is higher than that of in the early two decades. The average height of boys was approximately 3.9% higher than that of girls. The average height difference between high and low ages in the sample was 1.9%. The impact of policies on students' height was extremely small and may have little practical significance. Conclusion: China’s economic growth has a significant positive effect on the height of urban Chinese students without stagnation.

the association between macroeconomic factors and height are cross-sectional studies [4,7,12,14], whereas longitudinal studies are limited. Therefore, the present study aims to empirically evaluate how powerful the impact of China's macroeconomic development on the height of Chinese youth is, using provincial panel data collected from more than one million children in the past 30 years.

Data
Height data Data were derived from the report of 1985, 1991, 1995, 2000, 2005, 2010, and 2014 Chinese National Surveys on Students' Constitution and Health [17][18][19][20][21][22][23]. The CNSSCH is a series of surveys conducted by the Ministries of Education, Health, Science and Technology, the State Ethnic Affairs Commission, and the State Sports General Administration of the People's Republic of China. This study included only subjects of Han ethnicity, who account for 92% of the total Chinese population. The respondents came from 27 out of 31 provinces of China, excluding Hainan and Chongqing, both of which were founded after 1985. Qinghai and the Tibet autonomous region were also excluded because the former missed the 1995 survey, whereas the latter did not participate in nearly all surveys. Each province had equal sample sizes from three socioeconomic classes (i.e., "upper," "moderate," and "low") at the regional level. The participants were students aged 7-22 years from primary to college levels, who were selected from the same areas in each province from 1985-2014. Respondents were selected by stratified cluster sampling from certain classes, and the clusters were randomly selected from each grade in selected schools. Table 1 presents the sample sizes at each examination period.
Metal column height measuring stands (200 cm with a 0.1-cm precision) were used to measure each respondent's stature without shoes on.

Socioeconomic data
The data of per capita disposable income (PCDI) were derived from the respective statistical yearbooks of China's provinces (1985, 1991, 1995, 2000, 2005, 2010, and 2014) and the statistical database of the Chinese economic network [24].

Variables
To overcome heteroscedasticity and enhance the stationarity of data, we took the logarithm of all variables [25]. Estimation approach Panel econometric model was employed to estimate the association between students' height and PCDI and other factors, using height as the dependent variable and PCDI as a key independent variable. To reduce endogeneity, several control variables were introduced. To overcome heteroscedasticity and enhance the stationarity of data, we took the logarithm of all variables. Stata12.0 software was used to conduct the static model empirical analysis. lny it = α i + β 1 lnPCDI it + β 2 sex + β 3 age + β 4 P90 + β 5 P97 + β 6 P07 + ε it where lnyit is log of the height status in province i at time t, lnPCDI is the log of the ecnomic level of province i, sex, age,P90,P97and P07are control variables that are expected to relate height, sex is dummy variable,age is a numerical variable

Trends in height
The average height of male and female students increased steadily from 1985-2014 without any tendency to plateau (Figures 1 and 2, respectively). The research focuses on 7-, 13-, 16-, and19-year-old students as first-grade ages in the primary, junior high school, senior high school, and university levels, respectively. Among all ages, the increment in the height of males was largely pronounced at puberty(at 13years old), with an increase of 9.3 cm in three decades. In contrast, among all ages, the increment of the height in females was largely pronounced at childhood(at 7 years old), with an increase of 5.87 cm in three decades. The second largest increment of height in males was 6.46 cm at 7years old relative to 4.66 cm in 7 females at 13 years old. Figure 2 presents the socioeconomic changes in China between 1985 and 2014. As can be seen, urban PCDI increased from 739 to 28,843yuan during this period.

Empirical analysis
On the basis of the natural logarithm of each variable, the robust command was used to correct the standard error with white heteroscedasticity, so as to make the result more robust. All following regressions eliminated the outliers.
Table2 present the estimated results of the fixed effects models which is the test of the association between China's economic growth and the height of urban students after controlling for other variables, such as gender, age, and policy. A significant positive coefficient of 0.009 was found in this model. Given the log-log specification, this coefficient represented the elasticity of height with respect to PCDI: an increase of 1% in PCDI occurred with an increase in average height by approximately 0.009%. This indicated that, for every 1% increase in PCDI, the average height of youth in China increased by approximately 0.009% given that other factors remain unchanged. Stated another way, this implies that a doubling of income is associated with 0.9% increase in height, which is significant for height. This finding reveals that economic growth may exert a promoting effect on students' height. The coefficient of sex is 0.039, which is highly significant at the statistical level of 1%. The coefficient can be explained as the average height of the male sample, which is approximately 3.9% higher than the average height of the female sample. The coefficient of age is 0.019, which passes the significance level of 1%. This coefficient can be interpreted as the average height difference between adjacent ages of the sample students at 1.9%. The impact of policies in 1990 and 1997 on students' height was negative and highly significant. However, the coefficient value is extremely small and may have little practical significance.
As shown in Figure 1, 2005 is the turning point not only of China's macro-economy, but also of a new round of healthy economic growth. Therefore, to further explore trends of association between PCDI and students' height, the model was estimated As shown in Column (4), the elasticity of students' height is 0.011, which is highly significant. The coefficient is higher than that of in the early two decades. This finding may indicate that the Chinese economy has had additional positive effects on the height growth of children and adolescents from 2005 to 2014. In summary, the results show that the positive effect of the Chinese economy on students' height may gradually increase without any tendency to plateau.

Discussion
Using provincial-level 5-year interval panel data of 27provincesfrom 1985-2014, the current study reveals that economic growth has a significant positive effect on the height growth of youth. Data also show a linear relationship between economic growth and height of youth, suggesting an average height increase with continued increase in PCDI without any stagnation. Moreover, the study also indicates that the effect of economic growth on students' height in the past ten years is greater than that in the previous 20 years. In fact, numerous studies reported a positive association between economic status and healthy output in both developed and developing countries [4,6,7]. Increase in height growth entered a period of rapid growth in the past 30 years. [26][27][28].  [29]. In 1993, the rationing system was abolished, open markets became dominant, and a modern food system began to take shape [30], which improved crop and poultry productivity [31] and dramatically enriched the diet of the Chinese people. Entering the 21st century, China's economy achieved a new round of development. With the improvement of living standard and of health consciousness, people gradually prioritized healthy foods, such as milk, fish, and shrimp [32]. These are rich in high-grade protein and calcium essential for bone development [33]. The increase in high-quality protein and calcium intake, which is   Trends in Height and PCDI in Females