1. Introduction
The growing aging population poses a huge challenge to China’s industrialization and modernization. For rural areas, where the population is aging even more than in urban areas, off-farm employment of young workers and their prolonged out-migration to urban areas has further deepened the actual degree of aging and exacerbated its adverse effects on agricultural production and rural economic and social development. Of all of the challenges that the aging population poses for rural China, the aging of the agricultural labour force and the consequent scarcity of high-quality agricultural labour and the shortage of agricultural labour reserve has become one of the most talked about issues in current research. It has triggered a concern throughout Chinese society regarding the question of ‘who will farm in China in the future?’ [
1]. The shortage of effective supply of agricultural labour is not only strongly related to the development of agriculture and the rural economy, but also, if this situation persists for a long time, it may weaken the foundation of China’s grain production and food supply, and subsequently, Chinese food security and social stability [
2].
However, from the perspective of allocative efficiency, off-farm employment of young rural labour is not only in line with China’s current economic development strategy, but it also supports household income maximization through the division of labour in the family. This means that the labour input in agriculture depends on the long-term effective labour supply and allocation across agricultural and non-agricultural sectors for rural elderly people [
3]. Therefore, from this perspective, the realistic needs of China’s social and economic development require us to continue to study the labour supply of rural elderly people and identify the important determinants of their supply decisions through rigorous scientific analysis.
Although age, sex, education, employment opportunities, and wages are important factors of the decision, for the elderly, health is the key determinant [
4]. Due to the decline in physical function, the health of elderly people is more vulnerable to the shock of fatigue and illness and their recovery capability after health shocks is weaker. In addition, the depreciation of health accelerates with age [
5]. This indicates that health issues are more common and obvious among the elderly. The deterioration of health directly decreases productivity and work capacity, causing them to decide to either shorten their working time and reduce labour intensity or exit the labour market completely. Thus, in theory, the labour supply decision of older people is more susceptible to health variations [
6]. And this is even for the rural elderly people in China. However, the relationship is not limited to this. Because of the sharp differences in labour intensity and work environment, there are great differences in the health requirements between agricultural and non-agricultural employment [
7]. The effects of health on labour supply for these people are also on the allocation between agricultural and non-agricultural sectors.
Almost all Chinese studies on the effects of health on labour supply appeared after 2000 [
8,
9,
10,
11], and most of them explored this effect using rural samples. From the studies using a rural full-sample and out-migration sample, health improvements have a significant positive effect on labour supply [
12,
13,
14,
15]. However, for rural elderly people, conclusions are inconsistent. In here, some studies found that there is a ‘ceaseless toil’ phenomenon among elderly people in rural China; that is, they do not significantly reduce labour supply due to health deterioration as they grow older [
16,
17]. Therefore, it was concluded that health is an insignificant factor on their labour supply. But, some others provided evidences for the significant positive effect of health [
1,
18,
19].
Reviewing the existing literature, we have identified three potential challenges. First, there are limitations to the measures of health used, such as the subjective Self-Assessed Health Status (SAH) or objective medical indicators like Body Mass Index (BMI), Activities of Daily Living (ADLs), Instrumental Activities of Daily Living (IADLs) and chronic diseases, and this may be the reason for the inconsistent findings. For example, in the study of Tan and Zhou, the proxy variable of health, measured by SAH, has no impact on the labour supply of the elderly, while the variable indexed by BMI is strongly significant [
17]; Second, the endogeneity that arises from the simultaneous causality of health and labour supply decisions and the sample selection in labour supply is not fully considered in the estimation, and this could also be a source of biased regression results. Third, age and gender variation in the relationship between health and the labour supply decision and the difference of these relations between agricultural and non-agricultural employment are not discussed in-depth within an identical framework.
In light of the above three challenges, this paper attempts to complete two innovations. On the methodology, after analysing the advantages and disadvantages of the existing measurements of health and endogeneity in the econometric estimation of the health and labour supply, we construct a latent health stock index (LHSI), as presented by Bound et al. [
20] and Disney et al. [
21], and then comprehensively use the one-period lagged health index, the Heckman two-stage method (Heckman method), and the Bourguignon-Fournier-Gurand two-stage method (BFG method) in the estimation to ensure the effectiveness and reliability of the results. On the content, we examine the overall impact of health on the rural elderly labour supply. Furthermore, we explore the role of health in agricultural and non-agricultural sectors and the gender and age differences of health effects.
The remainder of this paper is structured as follows: the regression model, the definition of variables, and the data description are explained in
Section 2. Then, we present the empirical results in
Section 3, and we conclude with the findings and consider the implications of our work in
Section 4.
4. Conclusions
We use CHARLS data to examine the effect of health on the labour supply decision of the rural elderly in China and its gender and age differences. Given the measurement bias of the global health index, the simultaneous causality of health and labour supply decisions, and the sample selection in the labour supply, we construct an LHSI to eliminate measurement bias and then comprehensively use the one-period lagged LHSI and the Heckman and BFG methods to deal with the simultaneous causality and sample selectivity in model estimation. Our results show that:
- (1)
Health has a significantly positive impact on the overall probability of LFP. With other conditions constant, the LFP probability would increase by 8.72 per cent, on average, marginally with improvements in health.
- (2)
In the sub-divided employment types, health has a significantly positive influence on the probability of LFP in agricultural employment, off-farm employment, and off-farm self-employment. The marginal effect from improvement in health across these three types employments are 2.10 per cent, 15.59 per cent and 10.74 per cent, respectively, reflecting a stronger impact on LFP in off-farm employment than it in agricultural employment.
- (3)
With regards to working time, health improvements have a significant increasing effect. Holding other factors constant, the marginal effect of health is, on average, 149.2 h per year (about 18.65 days).
- (4)
In the sub-divided employment types, health is insignificant for working time in all three types of employment.
- (5)
From the gender comparison on LFP, we believe that the health condition changes can significantly affect the LFP of rural male and female elders. As far as the degree of impact is concerned, health effects on the LFP for non-agricultural employment are significantly greater in males than in females, while the impacts on males in agricultural employment and non-agricultural self-employment are slightly less than those on females.
- (6)
From the gender comparison on labour time, health has a strong positive effect on overall working time, and the influence on males is significantly larger than females. However, in the sub-divided employment types, this effect of health becomes insignificant in both male and female groups.
- (7)
From the age comparison on LFP, improvements in health have a positive impact on overall LFP which increases with age. However, for agricultural employment, even though good health leads to a higher possibility of LFP, the marginal impact is negative when individuals are 65 years old or younger and positive when they are older than 65 years old. For off-farm employment and off-farm self-employment, good health corresponds with a higher probability of LFP in these two types of employment and the marginal effect from health improvements all experience an inverse U-type process that rises at first and then decreases with increases in age.
- (8)
From the age comparison on labour time, health has a significant impact on overall working time when rural individuals are less than 66 years old and its impact increases with age. However, in the sub-divided employment types, this impact becomes insignificant in all age groups.
Based on these findings, we believe that the assertion of the ‘ceaseless toil’ for rural elders in China is hardly sustained empirically. For older people in rural areas, although the phenomenon that they are still working after the age of 60 is more common than among their urban counterparts, the results show that the increasing marginal effect of health on the labour supply with age growing have oppose the concept of the ‘ceaseless toil’. Meanwhile, given that agricultural labour input is increasingly dependent on the supply of rural elderly labour due to the integrated effect of population aging and continuous out-migration from agriculture for rural young labour, effective measures to improve the health of rural elderly people are key to alleviating the labour input shortage and ensuring production stability China’s agricultural sector. However, in policy design, the conflict between labour supply and health welfare of elders requires the governments to coordinate their relationship and clarify the key population by policies guiding. Our research proposes a focus on rural elders under the age of 66 as the focus of policy guidance, and their labour supply capacity guaranteed by health needs to be scientifically evaluated. And this would be future work in our researches.