Abstract

The report of 18th CPC National Congress has pointed out that in response to changes in both domestic and international economic developments we should speed up the creation of a new growth model and ensure that development is based on improved quality and performance.Under the new international situation，it is quite important for developing countries to expand trade openness，improve productivity and promote long-term economic growth.This paper analyzed the evolution process of productivity and trade openness theory，used macro data，industry data and micro data respectively，and discussed the effect of trade openness to productivity from the perspective of tariff barrier，non-tariff barrier，trade penetration and firm activity.The methods we used in this paper include normative research，empirical research，simulation analysis and regression analysis，and we preceded our discussion by country level，industry level and firm level.

In the country level，according to the research of Restuccia and Rogerson and Hsieh and Klenow，factor misallocation is one of the most important sources of total factor productivity differences between countries.We used this theory and the model of international trade and resource misallocation developed by Ferreira and Trejos to study developing countries.We introduced two indicators named tariff cost and non-tariff comprehensive cost in order to measure trade cost and to analyse how trade openness and trade cost impact the productivity differences between countries.Based on the trade cost data supplied by Economic and Social Commission for Asia and the Pacific Comprehensive Trade Cost Database，this paper took the method of simulation analysis to investigate 80 developing countries by using data about the year 2005.The results showed that trade can improve the level of output given input，which enabled countries benefit from trade.The existence of trade barriers caused the distortion of resources allocation among sectors and weakened gains from trade.Most developing countries can still realize their output increase from trade in the case of considering tariff barriers only.However，after taking non-tariff trade barriers into account，many countries wasted quite a number of trade benefits.Trade openness could reduce total factor productivity gap between developing countries and developed countries but this effect would be weakened due to the existence of trade barriers.

In the industry level，we introduced the theoretical model developed by Melitz and Ottaviano and empirical model developed by Chen et al.to the study of developing countries.Based on the industry level data supplied by CEPII database，we examined of the effect of trade openness to labor productivity of developing countries in industry level by using the data of 26 manufacturing industry from the year 1981 to the year 2006.We used the methods of OLS estimation，fixed effect model and first-order difference estimation，respectively，and found that domestic trade openness increased domestic productivity，while foreign openness acted to diminish it.The domestic relative trade openness to foreign had a positive impact on domestic relative productivity.The results are consistent with theoretical model and passed the robustness test.The results suggested that the improvement of trade openness in developing countries introduced more foreign firms to domestic market，which increased the product variety in domestic market and promoted domestic competition.The competition increased average productivity.Hence，trade openness in developing countries presented competition effect to domestic market and had a positive impact on the improvement of domestic industrial relative productivity，the reduction of productivity gap with the developed countries and the enhancement of domestic industrial competitiveness.

In the firm level，we used Levinsohn-Petrin method to estimate the total factor productivity of enterprises.Levinsohn-Petrin method is useful to solve the simultaneity problems caused by the correlation of inputs and productivity shocks in micro data and to recover the disadvantage of Olley-Pakes method because Olley-Pakes method could estimate total factor productivity of enterprises only when they have continuous investment.We also used two other indicators to measure productivity.One is total factor productivity calculated by fixed effect model and the other one is labor productivity calculated by added value per worker.We introduced two indicators to measure trade openness named port cost，which means the cost of transporting a 20-foot container cargo from the city that enterprise located to the port designated by the Institute of Comprehensive Transportation，and port distance，which means the great circle distance from the city that enterprise located to the nearest major port.We classified the whole sample of enterprises to four categories according to their trade status.They are enterprises that only export，enterprises that only import，enterprises that engaged in both export and import，and enterprises that do not engaged in international trade.We controlled the other determinants of enterprise productivity，such as management ability，staff's quality，salary incentive，utilization of information technology，research and development，employee training and decision-making structure.We examined the effect of trade openness and trade status to enterprise productivity based on the data of nearly 12000 manufacturing enterprises in China's 120 large and medium-sized cities in the year 2004.

The results showed that trade openness in the city that enterprise located had a positive impact on enterprise's total factor productivity and labor productivity.The reduction of comprehensive transport costs and the decrease of port distance would improve enterprise's total factor productivity and labor productivity.The results maintained significant after controlling for the enterprise trade status，enterprise characteristics and other determinants of enterprise productivity.Meanwhile，comprehensive transport costs are an important part of the non-tariff trade costs.When we controlled for all of the other determinants of enterprise productivity，the effect of comprehensive transport costs to enterprise productivity was more than twice of the effect of port distance.This conclusion maintained robust when we used Levinsohn-Petrin method to calculate total factor productivity，fixed effect model to calculate total factor productivity and added value per worker to calculate labor productivity.The conclusion suggests that non-tariff trade costs are a significant barrier to productivity growth and the reduction of non-tariff trade costs，especially the reduction of comprehensive transport costs，are quite important for enterprise productivity improvement.In addition，among the determinants that could positively affect enterprise productity，the ratio of bonus to employee wages，the education level of employees and the percentage of workers who often use computer to work have the most significant roles.The percentage of workers who had participated in information technology training and the degree of general manager's annual income relating to the company's performance also has important effect.This conclusion suggests that emphasizing the performance motivation of managers and employees，improving employees' education quality and information technical ability，and enhancing the career training inside firms are quite important to enterprise productivity.

This thesis mainly contributes to the existing literature as follows：

（1）The research of Restuccia and Rogerson and Hsieh and Klenow pointed out that factor misallocation is one of the most important sources of total factor productivity differences between countries.This point of view has been widely applied to the studies about productivity differences between countries.However，most researches analyzed factor misallocation and productivity differences by using closed economy model，and the research that applied this theory to trade openness is very limited.This paper followed the theory of factor misallocation and productivity differences，and introduced the open economy model developed by Ferreira and Trejos to study developing countries.We have not found the relevant domestic literature that had discussed on this issue.

（2）Melitz and Ottaviano constructed the trade model with heterogeneous firms from the perspective of trade openness inducing product market competition.Chen et al.transferred the theoretical model to empirical analysis and interpreted the international differences in industrial productivity by the international differences in trade openness.We introduced the theoretical model developed by Melitz and Ottaviano and empirical model developed by Chen et al.to the study of developing countries in the industry level research，which extended the application of theory.We have not found similar domestic literature on this issue according to our investigation.

（3）The existing literature about trade openness is generally focusing on the effect of tariff reduction to productivity.However，as tariff rates have been significantly reduced in worldwide，non-tariff barriers become the major impediment of international trade.In recent years，researchers pay great attention to the topic of non-tariff trade barriers，but the investigation of the relationship between non-tariff barriers and productivity is scarce.In this thesis，we analyzed the effect of non-tariff trade cost to productivity in developing countries systematically.To be specific，we explored non-tariff cost data provided by ESCAP comprehensive trade cost database in macro level and explored comprehensive transportation cost in China Investment Climate Survey provided by World Bank in micro level.

（4）The estimation of exports' impact on productivity will be biased if one do not consider the import activity of enterprises and the same thing happens when one estimates the imports' impact on productivity.Hence，it is necessary to provide a complete view of trade by grouping firms according to their exposure to foreign markets.Such detailed analysis is scarce in the trade literature and is mainly about developed countries.This thesis classified the whole sample of enterprises into four categories according to their trade status.They are enterprises that only export，enterprises that only import，enterprises that engaged in both export and import，and enterprises that do not engaged in international trade.We controlled the other determinants of enterprise productivity，such as management ability，staff's quality，salary incentive，utilization of information technology，research and development，employee training and decision-making structure to examine the effect of trade openness and trade status on China's manufacturing enterprise productivity.

Key words：trade openness；total factor productivity；labor productivity；non-tariff barriers；developing countries