City Size, Wage of Peasant workers and Expansion of. Mega-cities

Transcription

1 City Size, Wage of Peasant workers and Expansion of Mega-cities Based on Survey Data of Peasant workers in Six Provinces of Central and Western China Hongzhong Fan a a School of Economics, Huazhong University of Science and Technology(HUST), Wuhan , China Corresponding author: Hongzhong Fan hongzhong@mail.hust.edu.cn 1 / 22

2 City Size, Wage of Peasant workers and Expansion of Mega-cities Based on Survey Data of Peasant workers in Six Provinces of Central and Western China Abstract: Through the regression analysis of survey data of peasant workers in six provinces of central and western China, it's showed that there was no significant difference of the nominal wages among the peasant workers who work in the mega-cities, large cities, big cities and the middle and small cities. This finding proves that in China which is a developing country with great employment pressures, the wage negotiation ability of average quality workers in mega-cities are weak, they could not pass the higher living cost in mega-cities to the firms there and increase the investment cost of the firms by means of asking for higher nominal wages. As a result, firms in mega-cities enjoy the profits brought by the agglomeration economies but do not bear or completely bear the cost brought by the agglomeration diseconomies, however, average qualify workers in mega-cities bear the cost brought by the agglomeration diseconomies but do not enjoy the profits brought by the agglomeration economies, the allocations of the agglomeration economies and agglomeration diseconomies between firms and workers are mismatch. Under the circumstances, the market mechanism of house prices and other agglomeration diseconomies to prevent over-expansion of mega-cities fails to work. Because average quality workers' nominal wages in the mega-cities are nearly the same with that in the middle and small cities, and agglomeration economies bring higher production efficiency in mega cities, firms will choose continuously to invest in mega cities and create new job opportunities there, which constantly attract peasant workers and other average quality workers in small cities who urgently need job opportunities. So it is unavoidable to cause over-expansion of mega-cities in China as Beijing, Shanghai, etc., though house price there has outdistanced the purchasing power of average quality workers. Key words: Agglomeration Economies Agglomeration Diseconomies Peasant workers 1 Peasant workers refer to the peasants from the rural areas migrate into cities to work, they owned land in the rural areas but worked in the cities, they are mainly engaged in middle or low-income physical jobs and they do not have decent house or apartment to dwell in cities, so they usually dwell in the construction sites, dormitories with rough conditions and slums, but most of them have house and families in the rural areas, they treat their hometown which is in the rural area as the "home", they are always go back to their "home" and meet their families in important festivals. The number of the Chinese peasant workers is huge, totally close to 250,000,000 which is half of the total labour force in Chinese cities. According to the hukou system in China, peasant workers has no urban Hukou of the cities they work, but has rural hukou in their rural hometown. 2 / 22

3 1. Introduction 1.1. The Hetero-allocation of Economies of Agglomeration and Diseconomies of Agglomeration According to the theory of urban economics, the formation and development of city depends on agglomeration economies and diseconomies formed by the production and population concentration (Fujita, 1989). Agglomeration economies make the marginal income of firms and workers rise, and they are the market factors promoting firms and workers to concentrate in cities. Agglomeration diseconomies make the marginal cost of firms and workers rise, and they are the market factors to prevent firms and population from excessive agglomeration. The concurrent effect of agglomeration economies and diseconomies is that, against city size, the net marginal benefit of firms and workers change in an inverted U shape, with the U curve vertex representing the effective city size (Au and Henderson,2006a,2006b) 1. Here's an implicit assumption that the firms and workers in large cities enjoy similarly the benefits of the agglomeration economies, and also similarly bear the cost brought about by agglomeration diseconomies. However, although this assumption is true basically in developed countries such as the U.S., it is false in developing countries like China. From the perspective of micro-foundations, agglomeration economies have the following several kinds of forms: the concentration of production and population has generated the information spillovers, reducing the information cost relevant to technology, suppliers, purchasers and market conditions(fujita and Ogawa, 1982), and increasing consumer selectivity; the expansion of the market size makes possible the marketization production of intermediate goods which originally is not worth of trade, thus reducing the production and transaction cost of intermediate inputs (Becker and Henderson, 2000); the concentration of production and population and the expansion of the market size help realize the diversity of intermediate inputs which are provided for regional export firms, thus increasing the efficiency (Dixit and Stiglitz, 1977; Abdel-Rahua and Fujita, 1990); the concentration of production and population has reduced the information cost of the labor market, thus improving efficiency of matching between firms and workers (Helsley and Strange, 1990); Because of the geographical adjacency between buyers and sellers, the transportation cost (Krugman, 1991) and the search cost for retail (Caplin and Leahy, 1998, see Henderson (2002) for a review) have been reduced. Although large cities provide higher efficiency in production, large cities residents bear higher living costs, such as higher house prices and much longer commute time, noise and pollution and so on (Muth, 1969; Fujita and Ogawa, 1982). For example, in the United States and Latin American countries, the cost of living in large cities is more than two times of that in small cities (Thomas, 1980; Henderson, 1988), and the cost of living in Paris is 89-94% higher than that in other regions of France(Rousseau,1995). The marginal investment cost (including public facilities, public service and housing) of absorbing an extra family in average urban areas of Bangladesh, Egypt, Indonesia, and Pakistan is three times that in rural areass higher still for the primate city of a country (see Henderson (2002) for a review). Zheng (2001) found that agglomeration diseconomies in Tokyo metropolitan are mainly represented by too high house 1 Economic models with an endogenous number of cities postulate an inverted U shape of real income per worker against city size (Henderson, 1974; Helsley and Strange, 1990; Black and Henderson, 1999; Fujita, Krugman and Venables, 1999; Duranton and Puga, 2001, see Au and Henderson (2006a) for a review). 3 / 22

4 prices, too long commute time and poor environmental quality, and that agglomeration economies mainly stem from the concentration of company headquarters, government organizations and financial industries. It is known from the above literature review that agglomeration economies mainly bring higher efficiency for the firms, while agglomeration diseconomies mainly increase workers' living costs through house price, traffic, environment and other channels. Through wage negotiations in labor markets, the rising living costs with the agglomeration of production and population in a city will firstly affect workers nominal wage in the city, and then affect the investment cost of the firms in the city. Therefore, the effective city sizes are realized as a trade-off between the marginal benefits or higher production efficiency of firms from agglomeration economies and the increased living costs of workers from agglomeration diseconomies (Davis, Henderson, 2003). From the conclusion of the above documents, we can get an important corollary: firms are the primary direct beneficiaries of agglomeration economies, while workers are the primary direct victims of agglomeration diseconomies, but under the condition of perfect market mechanism and full employment (or quasi-full employment), workers in large cities can claim for higher nominal wage as compensation from firms. Otherwise, they will choose to migrate to other small and middle-sized cities. In this way, the higher living costs of workers in large cities will be passed on to firms, and transformed into investment costs of the firms. This process of cost conduction and transfer will effectively prevent over-concentration of production and population in large cities. It is worth noting that, there are two important preconditions for realizing conduction and transfer of the higher living cost of workers in large cities into the investment costs of firms, one being full employment or quasi-full employment, in which condition workers have enough power to bargain with firms on wages, the other one being perfect market mechanism, mainly referring to that workers can migrate and choose jobs among different cities with no or very little cost. For the developed countries such as the U.S., these two preconditions basically exist. Even though the United States also has certain unemployment, it is relatively much close to full employment compared with developing countries such as China and India. According to the American census of 2000, the nominal wage of workers in large metropolitan area with population size larger than 1.5 million is 32% higher than that in rural areas and small cities with population less than 250,000. If nominal wages are adjusted according to the living cost index and compared in the form of real wages, then the real wage of workers in large metropolitan areas with population larger than 1.5 million has no significant difference from that in rural areas and small cities with population smaller than (Baum-Snow, Pavan, 2011). This suggests that workers in large metropolitan areas pass on the higher living costs to firms by requiring for higher nominal wages, which means that the higher living costs in large metropolitan areas increase the investment cost of firms, but do not influence real incomes of workers there. However, in most developing countries 1, the two important preconditions for workers in large cities to pass on the higher living costs to firms are often difficult to meet. For most developing countries like China, there are a large number of unemployed average quality workers in cities or in rural areas. Except for a small proportion of high quality workers, ordinary workers in large cities in developing countries often bear heavy employment pressures and face high migration costs, and have very weak ability to negotiate for wages with firms. In order to get or 1 Even the same is with the process of urbanization for countries such as Japan, South Korea and Brazil, which entered rapid urbanization after the Second World War. 4 / 22

5 keep employment opportunities, they have to endure high house prices, crowded traffic and bad environment, live in slums, and accept survival wage which can only maintain continuation of life (Lewis, 1954), and which has no difference from the nominal wage of workers in middle and small cities, both equal to the survival wage for maintaining life continuation of workers. Therefore, in developing countries of heavy employment pressures, ordinary workers in large cities cannot require for higher nominal wages from firms, and then they can't pass on smoothly the higher living costs brought about by agglomeration diseconomies to firms, which should increase the firm investment cost. Consequently, firms enjoy high production efficiency brought about by agglomeration economies, but do not compensate for higher living costs of workers. In other words, for labor markets in large cities in developing countries, firms enjoy benefits brought about by agglomeration economies, but do not bear (or bear completely) the costs brought about by agglomeration diseconomies; however, workers bear the costs brought about by agglomeration diseconomies, but do not enjoy benefits brought about by agglomeration economies. Thus, there occur hetero-allocation about agglomeration economies and agglomeration diseconomies between firms and workers, which is significantly different from labor markets in large cities in developed countries, where the nominal wages of workers in large, middle and small cities have significant differences, and firms and workers take in the same way agglomeration economies and agglomeration diseconomies. All in all, the implicit assumption in the urban economic theory that agglomeration economies and diseconomies likewise have effects on firms and workers, which although accords with situations of developed countries such as the U.S., does not necessarily accord with situations of developing countries like China. We can define the phenomenon occurring in labor markets in developing countries as the hetero-allocation of agglomeration economies and agglomeration diseconomies, "hetero-allocation" for short. The existence of hetero-allocation between agglomeration economies and diseconomies leads to an important market failure, that is, the important market function of house prices and other agglomeration diseconomies to adjust the flow directions of capital and population is serious failure. Firms in large cities enjoy benefits of agglomeration economies, but do not bear the due agglomeration cost, and need not provide due compensation for housing and transportation costs of workers; they will continue to choose to invest in large cities, and create new jobs there. Workers in urgent need of jobs who cannot take so much consideration of house prices of large cities constantly pour into big cities to seek employment opportunities, which will inevitably cause fast expansion of large cities. Compared with workers in middle and small cities, workers in large cities of the United States have higher nominal wages, but real wages adjusted according to the living cost index have no significant differences (Baum-Snow, Pavan, 2011), which has for the reasonable spatial distribution of production and population the following significances: on one hand, workers in large cities have higher nominal wages, thus increasing the investment cost of firms there; on the other hand, they have higher living costs, thus reducing the attraction of large cities for population, which forms the important market mechanisms of preventing capital and population from excessively gathering in large cities, and can effectively prevent the outbreak of large city diseases, such as high house prices, heavy traffic, environmental degradation and etc. However, there may not exist such kinds of market mechanisms in China and other developing countries Wage decision mechanism of Chinese peasant workers and the market function failure 5 / 22

6 of agglomeration diseconomies Peasant workers are the major sources for the expansion of Chinese urban population, and the main force of Chinese urbanization. For instance, in 2010 peasant workers account for 35.9% of the resident population in Beijing, the new resident population of Beijing is mostly peasant workers, and peasant workers account for 77% of the resident population in Shenzhen city in Guangdong province. In national view, approximately 250 million peasant workers account for nearly half of all the employment population of Chinese second and third industry, and thus form an important component of Chinese urban labor market. Therefore, the study of peasant workers' wage decision mechanism is of great value for the study of the decision mechanism of Chinese labor wages. The average cultivated land for peasant household in China rural areas is about half a ha., and the agricultural marginal productivity in the countryside for most of the peasant workers is very low and even equal to zero, so if not working part-time in the city, they have no marginal contribution to the household income. In other words, most of the peasant workers are surplus rural labor, and their opportunity cost to work in the city is almost equal to zero. According to the labor supply theory by Lewis, firms mainly decide wages of peasant workers on the basis of the opportunity cost of peasant workers working in cities (Lewis, 1954). The opportunity cost for peasant workers working in cities is the marginal contribution of farming in the countryside for household incomes, and it has no direct relation with sizes and house prices of cities where they work. Firms can hire enough peasant workers only by providing wages slightly above the survival wage for maintaining their life continuation, so as to appropriately compensate for some emotional costs of being far away from home, costs of physical strength and travel expenses. Thus, for a long time, the nominal wages of Chinese peasant workers in large cities are very low, far from sufficient to rent or purchase decent houses in cities where they work 1. However, although the wages of most peasant workers can neither afford purchasing houses nor renting decent houses in cities, after several years of accumulation, their wages can afford building decent "western-style building" in the rural home. Due to restrictions of the wage decision mechanism, the household registration system, the land system and the social security system, peasant workers gradually form a life pattern of "earning money in cities, and building houses in rural hometown". And in the acquaintance society of rural hometown, the psychology of comparison consumption among peasant workers on their houses and durable goods has strengthened this life pattern. Therefore, the purpose for them working in cities is often "earning money, building houses at rural hometown and getting married ". It is because of this kind of life style that, according to the second agricultural census data of China, among all employees of peasant workers, those under age of 30 years account for 52.6%, those between 30 to 40 years account for 29.5%, and those above 40 years old account for 17.9% (research group, 2010). It is suggested that, generally peasant workers above the age of 35 years begin to retreat from the labor market, and most of them have dropped out of the labor market after 40 years old. This is because 1 Large, middle and small cities of eastern, central and western part of China have in different degrees the problem of "peasant worker shortage", but it does not mean that peasant workers are not surplus labor forces. According to some researches, there are still a large number of surplus labor forces in the rural area of China. The main reason for the shortage of peasant workers is that it is difficult for them to realize citizenization, which on the one hand, leads to that a large number of peasant workers can not leave home, and on the other hand, leads to that peasant workers retire from the urban labor market too early between 35 to 40 years old. (Fan Hongzhong, Lian Yujun(2010), Wang Jin, Zhong Xiaohan (2011). 6 / 22

7 when above 40 years old, most peasant workers have solved the problem of housing in their hometown and married. At the same time, their physical strengths can't compete with young peasant workers, and their emotional costs of being far away from home also increase greatly. Despite that a few peasant workers settle in cities finally, they do not represent the mainstream. This behavior pattern of peasant workers can be represented by the following formula: Max: W C a C (a) b In the formula (a), W represents the nominal wage of peasant workers, Ca the clothing and foods and other daily life costs of peasant workers in cities, and C the dwelling cost of peasant workers in cities. Their traveling expenses for working places have been reduced greatly due to the improvement of provincial transportation facilities and thus can be neglected. The severe employment situation and the opportunity cost of working out determine the wage negotiation status of peasant workers, and they can only acceptw, which is a little higher than the survival wage. In order to maximize formula (a), they can only minimize b C andc. Therefore, they would like to live in construction sites, rough collective dormitories, and even city slums, so as to save more money for building houses at rural hometowns, and even for some consumption of comparison in the hometown. The value acquired by subtracting a b Ca and Cb from W constitutes incentive factors for peasant workers to work in cities by overcoming emotional costs and physical force costs. The behavior pattern of peasant workers leads to basically the same dwelling pattern of peasant workers in large, middle and small cities. Peasant workers of manufacturing industry generally dwell in humble collective dormitories built by firms, which are often free of charge or charge for lower rents. The bigger the city, the larger the number of peasant workers lives in one dormitory. Many self-employed peasant workers dwell at night in the store where they work in the day, and many peasant workers of the construction industry dwell in free and temporary sheds near construction sites. And many peasant workers choose to dwell in remote villages of rural-urban continuum or crowded urban villages. These places lack necessary public service facilities, and have bad hygienic environment, and may also take a long time at rush hours to go to work for living there, but the rent there is very cheap, and these places can be called the Chinese type of "slum" 1. In order to save the cost of dwelling, some of peasant workers in large cities even choose to dwell in the basement of buildings. To sum up the above free or cheap dwelling patterns, peasant workers may have the same dwelling cost both in large cities and small cities. Even if there is difference, it is far smaller than the difference of the costs of dwelling in a decent house or apartment between large cities and small cities 2. The equilibrium condition for peasant workers to flow between large and small cities can be represented by the following formula (b): 1 According to the theory of urban economics, housing rents of villages in urban-rural continuum of both big cities and small cities are the same. It might take longer commuter hours in large cities, but according to the behavior pattern of peasant workers, they do not pay much attention to time compared with money. 2 In 2009 Dang Guoying asked about the living cost of one peasant worker selling water along Changan Street near Jianguomen of Beijing. The peasant worker said he lived in a nearby hutong, and the living cost of every month is 200 Yuan (Dang Guoying, 2011). Quoted from the speech craft by Dang Guoying in December 2011 on the first Central Plains Rural Development Forum held in Xuchang, Henan. 7 / 22

8 W W C C 0 C C 1 A 1B Ab Bb Ah Bh (b) In formula (b), W 1 A and W 1 B respectively represent nominal wages of peasant workers working in the large city A and small city B, C Ab 和 C Bb respectively represent real dwelling costs of peasant workers in the large city A and small city B, and C Ah andc Bh respectively represent the costs of dwelling in a decent house or apartment in the large city A and small city B. Since food, clothing and other commodities can be tradable through spaces; formula (b) assumes thatc, daily life costs like clothing and food, are the same in large and small cities. a Formula (b) has three important meanings: (1) As for as the difference of decent housing costs between large and small cities is concerned, peasant workers have basically the same dwelling cost and nominal wage in both large and small cities, which are irrelevant with city size, house prices and decent housing costs. (2) The market function of house prices and agglomeration diseconomies preventing peasant workers and capital from excessively gathering in mega-cities seriously fails to work. To carry out empirical research of this theoretical conclusion is one of the important purposes for this paper. 1.3 The market failure and the expansion of the mega-city The empirical research on the relationship between Chinese city size and wages of peasant workers has the following important significances: Firstly, Compared with urban permanent-resident labor forces, peasant workers have very high regional flowability, and they can level wages of laborers with the same quality through the labor market competition, thus exerting important influences on labor wages of cities with different size in China. According to the research of national 1% population sampling survey data in 2005, Xing Chunbing (2008) has found that ordinary peasant workers' wages are very close to that of ordinary urban permanent-resident workers. Therefore, the study of the relationship between city size and peasant workers wages help to reveal the relationship between Chinese city size and wages of ordinary workers. Secondly, in recent years, with the rapid development of Chinese economic growth and urbanization, there emerges a phenomenon difficult to explain theoretically in the development processes of Chinese mega-cities. On the one hand, house price of the mega-city outdistances purchasing power of ordinary urban residents, and on the other hand, population size of mega-cities expand continually and rapidly. For instance, although in recent years house prices of Beijing and Shanghai outdistance purchase powers of ordinary residents, according to the data of Beijing statistic bureau, in the year 2005, 2006, 2007, 2008, 2009, the resident population in Beijing increases respectively 453,000, 430,000, 520,000, 620,000, 600,000, equivalently increasing by the population size of a middle city per year. Most of the new population in Beijing is peasant workers. And according to the 2010 census data, the increase of population size in Shanghai even surpasses that in Beijing. The reason for this phenomenon may be that, nominal wages of peasant workers in Beijing, Shanghai and other mega-cities are higher, although the housing and transportation cost and other living costs are also very high. According to the above analysis, the more possible reason is that, nominal wages of peasant workers in Beijing, Shanghai and other mega-cities have no significant difference with those in other middle and small cities. It 8 / 22

9 is only because mega-cities have more job opportunities, and the market mechanism of housing prices, transportation costs and other agglomeration diseconomies preventing the flow of peasant workers fails to work. If the latter reason is found to be true according to empirical researches, what is implied is that it s difficult to realize space rational allocation of Chinese capital and population only relying on the market mechanism. This conclusion helps the government to take measures for guiding the rational flow of capital, preventing urban diseases in large cities, and promoting the healthy development of Chinese urbanization. There may be some other reasons for the assumption that nominal wages of metropolitan peasant workers are not necessarily higher than that in middle and small cities: (1) the level of public service has important influence on the flow of population(jennifer, 1982). Large cities of China often have much higher public service levels than small and middle cities, so young peasant workers may prefer to accept lower nominal wages and are thus willing to work for a few years in large cities before going back home "to build houses and get married". (2) According to the study by Lucas (2004), it is found that the rural population expects faster increase of the human capital and wages for working in large cities in the long term. So in view of the whole life cycle incomes, the rural population would rather receive lower real wages at the moment and thus are willing to work in large cities. But, whatever the reason is, the consequence is that, the peasant workers ability of wage negotiation declines, and the market mechanism of house prices and traffic costs regulating the flow of peasant workers fails to work severely. There are many theories and empirical researches about excessive concentration. For instance, Todaro (1969) and Lucas (2004) also explain to some extent the overexpansion phenomenon of the mega-city through growth prospects of the human capital or income. A large number of literatures argue that national political institutions and government policies, such as the preferential policies adapted by the central government of a state for individual cities (usually the capital) leads to absence of fair competition stage among cities, and thus lead to excessive concentration 1. The contribution of this paper lies in that it finds new mechanisms leading to the overexpansion of mega-cities in developing countries, that is in developing countries like China, due to the low wage negotiation ability of ordinary labors, their nominal wages have no significant difference between big cities and small cities (high quality labor forces are excepted, because their wages are different between big and small cities due to their scarcity), thus leading to that the market mechanism of house prices and agglomeration diseconomies to prevent capital and population from excessive flowing into large cities fails to work. The direct significance of this conclusion is that, the preferential policies for large cities will not only lead to the expansion of the mega-city, the decline of metropolitan ordinary labors life levels and the appearance of slums as well, this is because the wage of ordinary labors in large cities (such as Chinese peasant workers) does not contain decent housing costs in large cities, and the wages of ordinary labors between large and small cities have no significant differences. Although the study of relations between nominal wages of Chinese peasant workers and the city sizes helps to explain the expansion of Chinese mega-cities, empirical studies about relations between Chinese city size and peasant worker wages are almost blank. The main reason for this phenomenon may be that, China is lack of accurate data about city size and peasant worker wages. Although Chinese statistic department provides different population sizes of urban household 1 See in particular ades and Glaeser, 1995; Renaud, 1981; Henderson, 1988; Moomaw and Shatter, 1996; Henderson and Kuncoro, 1996; Henderson and Becker, 2000; also see Davis and Henderson(2003) for a review. 9 / 22

10 registration and the average wage of formal workers in enterprises and institutions above designated size, neither does it provide sizes of informal and peasant workers nor the wage of them. In addition, the economic development levels of cities in eastern, central and western China are greatly different and the costs of living among cities with the same population size also have great differences. This requires us that in the research of relations between city size and wages of peasant workers, urban economic development levels of different areas should be taken into account. During the Spring Festival 2 of 2011, a stochastic questionnaire survey was specially organized among peasant workers of 78 natural villages in six Midwest provinces of China including Hunan, Hubei, Henan, Anhui, Sichuan, and Jiangxi, and the relevant data of peasant workers in different size of cities was collected. Through quantitative analysis of the data, this paper has found that city size has no significant influence on nominal wages of Chinese peasant workers, and there are no significant differences on nominal wages of peasant workers among mega-cities, large cities, big cities, middle and small cities. It is indicated from this conclusion that, high house prices and high costs of living in Beijing, Shanghai and other Chinese mega-cities can't stop influx of peasant workers, the market mechanism of high house prices, high costs of living and other agglomeration diseconomies preventing Chinese giant cities from overexpansion fails to work severely, and thus other means should be considered to prevent Chinese mega-cities from over-expansion. Next, the second section is the definition of the peasant worker occupations and cities of different size; the third section introduces the questionnaire survey data of this paper; the fourth section conducts regression analysis of peasant worker wages and city size; the fifth section summarizes the whole paper. 2. The definition of peasant workers' occupation and sizes of cities 2.1. Definition of peasant workers' occupations Peasant workers' occupations may have an important impact on their nominal wages. For example, some jobs require in relative high technique and some jobs will harm to the bodies. The ones who engage in such occupations may get higher nominal wages, such as painters. Therefore, the correct analysis of relationship between wages of peasant workers and city size should consider impact of peasant workers' occupations. According to the characteristics of occupations peasant workers mainly engage in and National Bureau of Statistics Census of Agriculture questionnaire on peasant workers in occupational classification, we put the occupations of peasant workers into the following 11 categories this paper. (1)Small business owners or self-employed workers. They have a certain investment, such as the bun shop, long-distance transport (car owners) and small nursery owners. (2)Professional and technical workers: engineers, electrical designers, software designers, interior designers, quality supervisors and doctors. Generally these occupations require formal education. In this survey, farmers often consider that their children who graduated from universities and are working without settling down to buy houses as peasant workers. These kind 2 The Spring Festival is the most important holiday in China, and peasant workers usually choose this holiday to go back home for family reunion, so the Spring Festival is the best time for the investigation of the peasant worker wage and its influencing factors. 10 / 22

11 of labour's wages are worthy of analysis. They have good education background and the figures help to study the relationship between different degrees and wages. (3)Over middle-level managers of Enterprises: division manager and senior white-collar workers (including mid-level). (4) Ordinary staff: refers to the ordinary clerk in enterprises and institutions, under mid-level (not including middle-level ). (5) Sales personnel: Sales personnel refers to professional sales personnel without fixed place (such as real estate sales personnel, pharmaceutical sales personnel (excluding pharmacy staff, pharmacy staff are included in (7) - Business Workers), electrical or other equipment sales personnel such as elevator sales personnel (excluding store sales staff, which in (7) ). (6) The manufacturing industry workers. (7) Business enterprise workers, including sales staff or tally clerks in supermarkets, department stores, drug stores, computer stores and so on. (8)Catering industry workers and household service workers( excluding chef), including waiters, pantry workers and household cleaners. (9) Building and fitment workers: masons, carpenters, painters, plumbers, aluminum workers, excluding non-technical workers which including in (11) category: other occupation. (10) Other skilled workers: tailors, cooks, drivers, machine workers, turners, fitters, nurses, kindergarten teachers, mechanics, technicians and so on. (11) Other occupation: unlicensed vendors, waste collection, unskilled building workers and other unskilled workers Definition of city size Regional economy development levels of China are various. We not only lack of accurate data of urban population, but also hard to judge the living cost according to urban population due to differences in levels of urban development. In order to reveal the relationship among the city size, urban living cost and nominal wages of peasant workers, this paper uses two different methods to define the different city sizes and does regression analysis respectively in order to get reliable conclusions According to house price Moulton (1995) has considered that house price and cost are the most important element lead to difference of regional living cost. The experimental living cost index of U.S. Census Bureau, FMR index, assumes the differences of living cost between regions is only caused by the housing cost (Curran et al, 2006), because the non-tradable goods prices and housing costs are very relevant, while the prices of tradable goods are almost the same. So under the condition that Chinese statistic department does not publish accurate data about urban population, to some extent, the house price reflects differences of living cost better than population size. Because of that, we define cities with different size and living cost according to house price. Chinese cities( towns ) living and price Yearbook 2010 gives dwelling area and sales figures of prefecture level cities and over in According to these data, we calculate and give first definition about mega-city, large city and other city. 11 / 22

12 Mega-cities include Beijing, Shanghai, Guangzhou and Shenzhen, whose house prices in 2009 were over ten thousand Yuan per Square meter. Large cites refer to cities whose house price were over 5000 Yuan per Square meter in 2009, including Tianjin, Dalian, Wuhan etc 1. The cities whose house price in 2009 under 5000 Yuan per Square meter are other cities According to administrative level and developing level China's economy development model is government-led. That means, the cities with high administrative level can get more preferential policies and development opportunities and the result is that the cities with high administrative level generally have larger population size and better public services. The eastern cities in China are major gathering places of peasant workers. They are also more developed regions in China. The housing and living costs there are generally higher than those of the same administrative level central and western cities. Considering with China s economy development and administrative level, we defined the mega-cities, large cities, big cities and other cities. Among them, the mega-cities are Beijing, Shanghai, Guangzhou, and Shenzhen. Large cities include 13 sub-provincial cities excluding Chongqing, Suzhou, Foshan, Guangzhou, and Shenzhen. Although Chongqing is a provincial city, its house price and the cost of living are closer to Wuhan. Although Suzhou and Foshan are prefecture-level cities, according to the sixth census data, Suzhou has more than 10 million resident population and its house price and size of the population are in excess of Wuhan; Foshan, which is near Guangzhou, is closer to downtown of Guangzhou than Guangzhou Huadu District. Therefore, Foshan can actually be seen as a sub-center of Guangzhou metropolitan circle. On March 19, 2009, Guangzhou and Foshan signed the "Agreement of Co-operation between Guangzhou And Foshan ", and agreements about the urban planning, transport infrastructure, industrial cooperation and environmental protection, which mark the official integration of Guangzhou and Foshan. Big cities include other capital cities in midwestern province and some big cities in eastern provinces. They are Wuxi, Changzhou, Zhenjiang, Yangzhou and Nantong in Jiangsu, Shaoxing, Tanzhou, Jinhua and Huzhou in Zhejiang, Dongguan, Zhuhai and Shantou in Guangdong, Quanzhou and Fuzhou in Fujian, Yantai in Shandong and Shijiazhuang in Hebei. Other cities are the cities excluding in the list above. 3. Data Explanation and Description This survey chooses Jiangyou City in Sichuan, Wanjiang County and Lixin County in Anhui, Zhijiang City in Hubei, Duchang County in Jiangxi. Nan County in Hunan and Puyang City in Henan. These regions not only output large number of peasant workers but also locate at flat area. The reason that we choose flat area is to control cost of information about peasant workers and cost of transport. If we do survey in mountain area, although farmers are from the same county, the cost of information about peasant workers and cost of transport are different significantly. This survey use stratified random interview and questionnaire survey method. Investigators selected the 2-3 township which were adjacent to, and then randomly selected villages, 1 Large cities includes Tianjin, Dalian, Nanjing, Wuxi, Suzhou, Yangzhou, Wenzhou, Ningbo, Jiaxing, Huzhou, Shaoxing, Zhoushan, Taizhou and Lishui, Fuzhou, Xiamen, Quanzhou, Jinan, Qingdao, Wuhan, Zhuhai, Dongguan, Foshan, Haikou, Sanya, Chengdu and Simao. 12 / 22

13 and each villages randomly selected 10 to 20 farmers for investigating. When we did survey, investigators read the questionnaires and peasant workers or heads of households answered, and then investigators filled out the questionnaires. The survey includes nominal wages and personal characteristics of 1156 peasant workers, but a small number of peasant workers did not accurately report the workplace. So the effective sample is This paper studies that peasant workers' wages in different -sized cities. As indicated below, if the size of city is divided into 3-4 groups, each group of samples has nearly 200 people or more. More importantly, the individual characteristics of peasant workers are very similar. For example, peasant workers have the same level of education. Most of them graduated from middle or primary schools. At the same time, peasant workers are also broadly similar occupations. Most are self-employed workers, the construction or manufacturing workers. For these two reasons, when the purpose of this paper is concerned, the author believes that over 1,000 peasant workers in the sample should have a strong representation. (1) workplace of sample peasant workers In the 1150 samples of peasant workers, there are 338 workers work in the home province and 812 workers work in other provinces. It shows that more than 70 % of peasant workers work in other provinces, among which there are 629 people work in eastern cities. Table 1 shows distribution about sample of peasant workers in different sizes of cities according to two definition methods mentioned above. From Table 1, we can know that the majority of peasant workers work in the middle and small towns. Table 1:The number of sample peasant workers in different sizes of cities Unit: People According to 1st city size Mega-city Large City Other City definition method According to 2nd city size Mega-city Large City Big City Other City definition method (2) Wage of sample peasant workers in different sizes of cities There are 1079 peasant workers reporting monthly average nominal wages explicitly. The average value is 2639 Yuan and the standard deviation is 1836, which include 13 peasant workers whose monthly nominal wages have surpassed 8000 Yuan. We consider these peasant workers whose monthly average wages surpassed 8000 Yuan as the strange samples. Table 2 gives description statistical result about peasant workers' monthly wages (the ones' wages are less than 8000 Yuan) and dwelling cost in the different sizes of cities. From Table 2, regardless of according to first or second kind of city size definition method, there is no significant difference of peasant workers' monthly average wages and housing cost between mega-cities, large cities, big cities and middle and small cities. 13 / 22

14 Table 2: description statistical result about peasant workers' monthly wages and dwelling cost in the different sizes of cities Unit: Yuan Wage Dwelling Cost Mean Value Median Standard Deviation Mean Value Median Standard Deviatio n Accordin g to 1st city size definition method Accordin g to 2nd city size definition method Mega-cit y Large City Other City Mega-cit y Large City Big City Other City (3) Other characters of sample peasant workers In the survey, there were 2 peasant workers who were illiteracy. And there were 195 people finishing primary school education, 790 people finishing junior high school education, 72 people finishing high school education, 45 people finishing technical secondary school education, 25 people finishing junior college education and 29 people finishing higher education. It means illiterate peasant workers are almost non-existent and the majority of peasant workers have junior middle school education and primary school education. The number of these two kinds of peasant workers accounted for 16.9 % and 68.3% respectively in total. Age Work Duration Table 3: ages and work duration of sample peasant workers Mean Value Median Maximum Value Minimum Value Standard Deviation 30.9 ( years old) 4.1(years) Among these investigated peasant workers, there are 231 small business owners or self-employed workers, 35 professional and technical workers, 15 over middle-level managers of Enterprises and institutions, 51 general staff, 25 sales personnel, 215 manufacturing industry workers, 36 business enterprise workers, 57 catering industry workers and household service workers, 110 building and fitment workers, 212 other skilled workers and 152 workers who work 14 / 22

15 in other occupation. Among these investigated peasant workers, male workers are 753 and Communists are 26 (There are 14 communists work outside their home provinces.) Table 3 gives description statistical result about ages and work duration of these investigated peasant workers. 4. Econometric Analysis Result Model (1) and Model (2) are used in this paper. log( wage) 0 1age 2age2 2time 3gend 4 party 5 prov EDU VOC SIZE (1) log( wage provin) 0 1age 2age2 3time 4gend 5 party EDU VOC SIZE (2) log( hcos t) 0 1age 2age2 3gend 4 party 5 prov EDU VOC SIZE (3) Above models, wage refers to peasant workers' monthly average wage. In some documents about peasant workers' wage, some are in monthly wage measurement and some are in hourly wage measurement. Also others use these two measurement methods. In our samples, there are many self-employed workers. It is difficult to measure their working time hourly. So it is more reasonable to use monthly average wage in this paper. age refers to peasant workers' ages, whose unit is year. age2 is square of age. According to some studies, age and age2 affect wage significantly (Yao Jun, 2010). The interval of peasant workers' age is between 15 and 58 years old. Growing with age, peasant workers' strength and energy may be inverted "U" curve. However, experience will increase with age. To sum, coefficient of age should be positive and coefficient of age2 should be negative. Some documents have not considered square of age, but it is considered in this paper. time refers to how long peasant workers have done their reported work, whose unit is year. To some extent, time reflects peasant workers' skills. gend is a dummy variable referring to gender of peasant workers. Male is 1. party is another dummy variable referring to communist party member. If a peasant worker is a communist member, value of party will be 1. Generally, only a capable person can be a communist member in countryside. To some degree, we hope party will reflect peasant workers' ability. prov is a dummy variable referring to the province where peasant workers work. If they work in home province, value of prov will be 0, otherwise it being 1. Working far away from home means less chance to get together with family. It means more cost of feeling and visiting so that value of prov should be positive. Moreover, prov may also reflect peasant workers' individual ability, because the ones working far away from home need better communication ability, adventure spirit and gumption. 15 / 22

16 EDU is a group of dummy variable referring to education degree. They are junior middle school level, senior middle school level, technical secondary school level, junior college level and high education level. VOC is also a group of dummy variable including 10 dummy variables referring to 10 occupations ( excluding manufacturing industry workers). SIZE is group of dummy variable referring to size of different cities where peasant workers work. When they work in mega-city like Beijing, Shanghai, Guangzhou and Shenzhen, SIZE 0 is 1. When they work in a city whose house price is over 5000 Yuan/ square meter, SIZE 1 is 1. In a word, SIZE 0 and SIZE 1 represent for mega-city and large city according to first method to define city size. If we use second method to define city size, when peasant workers work in large cities, SIZE 2 is 1. When they work in big cities, SIZE 3 is 1. So SIZE 0, SIZE 2 and SIZE 3 are used to control mega-city, large city and big city according to second method. The focus of this paper is to investigate whether the coefficients of SIZE 0, SIZE 1, SIZE 2 and SIZE 3 are significant when we do the regression analysis. In other words,we want to analyze whether nominal wages of peasant workers working in different sizes of cities change significantly. In Model 2, provin is a variable which multiplies by wage. It results from transferring 0 of prov to na in Model 1. That is to say, Model 2 will do regression analysis about peasant workers working outside home provinces. Compared with Model 1, Model 2 not only controls peasant workers' ability and cost of feeling better but also removes correlation between prov and SIZE 0, SIZE 1, SIZE 2 and SIZE 3. However, samples are less than those in Model 1. In Model 3, hcost refers to housing cost (or dwelling cost) of peasant workers. The other variables in Model 3 are the same with Model 1 and Model 2. In Model 3, prov is used to control cost elements. When peasant workers working outside home provinces, it is more common that they live with their wives and children and housing cost may be increase. Moreover, Model 3 still use variables like age, EDU, VOC and gend. In the end of this paper, we give regression results of housing cost of peasant workers according to Model 3 ( see Appendix Table 1). We find that relationship between the housing cost of peasant workers and city size is not statistically significant. Table 4 gives results according to OLS analysis, among which regression 1, regression 3 and regression 4 are the results using first definition method and regression 2, regression 5 and regression 6 are the results using second definition method. From regression 1 and 2 which only contain dummy variables like SIZE and prov, although most coefficients are not significant, we find all coefficients of SIZE are negative. It shows that wages of peasant workers working in mega-cities, large cities and big cities are not higher than those in middle and small cities when we only consider whether peasant workers work outside home provinces. Regressions 3-6 in Table 4 increase other control variables. However, we do not report EDU and VOC in order to narrow the space of table. Compared regression 3, 4 and regression 5, 6, when we add age2 into the model, it increases both significance of age and adjusted R2. In regression 2-6, it presents that effect of duration is positive and significant, which means duration increases human capital. Meanwhile, to be communist member also has positive effect on wages. To certain degree, to be communist member reflects peasant workers ability. Also, working outside home provinces influences wage positively and significantly. It means these peasant workers have to need higher cost of feeling and visiting or better personal capability. 16 / 22