,

1.School of Business,Hunan University of Commerce,Changsha 410205,China; 2.Hunan University of Commerce,Changsha 410205 China

Abstract Under the analysis framework of output distance function,this paper makes overall measurement of rural Agriculture Logistics Efficiency (ALE) by constructing non-parametric production frontier function model using related panel date of 28 provinces over the period 1994-2014 in China.The accession to the WTO is the most important milestone in the process of China’s globalization.It employs the bi-directional fixed effects model and the difference in differences (DID) method to study the impacts of globalization (joining WTO) on ALE and its mechanism of action in the context of China.The results show that the WTO makes ALE of China generally increased by about 34.2%,promoting ALE with a long-term dynamic gradual process; long-term impact of globalization on ALE is about 83.3%,which is higher than short-term effect; the promotion of ALE is obtained mainly by the improvement of rural logistics infrastructure,rural informatization level,regional division of agriculture and the quality of rural labor force brought by globalization; agriculture logistics professional level has not yet acquired obvious enhancement during the process of globalization in China so far.

Key words Globalization,WTO,Agriculture Logistics Efficiency (ALE),Distance function,Difference in differences (DID)

1 Introduction

In the process of globalization during recent decades,China has realized significant economic growth.At the same time,rapid economic growth has led to remarkable agrarian development.The average annual growth rate of China’s agriculture output value reaches 4.8% from 1978 to 2014.Especially after the WTO accession on 2001,China’s agricultural production capacity has prominently enhanced,grain output increases for 7 years consecutively and cotton,oil plants,sugar,meat,milk,eggs,aquatic products,vegetables product yield keep stable development as well.It has allowed the penetration of market and other forces of capitalism into the agricultural production system,leading to advances in production techniques,along with huge distribution of agricultural product output[1].The improvement in agricultural comprehensive production capacity should not only rely on agricultural technology progress,also depend on the construction of efficient agricultural product logistics system[2].Low efficiency in rural agriculture logistics will lead to more losses and higher cost of agriculture products (especially the fresh and perishable agriculture products) and weaken agricultural comprehensive competition ability in the world market.

China is still characterized by the urban and rural binary structure.Agriculture logistics become vital to smooth of supply channel of agriculture products for the whole country.By reducing delivery times,maintaining product quality and reducing shipping costs,improvement in logistics efficiency have greatly facilitated trade of agriculture products.Agricultural Logistics Efficiency (ALE) runs as an objective comprehensive evaluation index to measure the quality of agricultural products logistics system[3].It reflects the ability of using limited resources to achieve maximum output in agriculture logistics,and represents a comprehensive consideration that involves the agricultural product market transaction efficiency and division of labor in agriculture.

Accession to WTO is a landmark in the economic globalization of China.In this paper,we investigated the globalization impacts on the efficiency of agriculture logistics in China,calculated and analyzedALE,aiming at the important time point of China’s accession to the WTO.

2 Literature review

The research on the empirical measure of commodity logistics efficiency is derived from the study of the technical efficiency.As early as 1950s,scholars began to study the measurement methods of technical efficiency in economic activities.The concept and measure methods of the technical efficiency were proposed[4-5]respectively based on input and output angle,marking the generation of the technical efficiency theory.With further development of the study,two kinds of technical efficiency measurement methods are gradually formed: one is frontier production function method within the category of econometrics; the other includes Data Envelopment Analysis (DEA) and C2GS2 technical efficiency evaluation model in the field of operational research.Charnes,Seiford and Cooper are the representative scholars.

The degree of market integration and access convenience[6]are also utilized to evaluate the efficiency of logistics.Lele[7]used empirical economic analysis method for the first time to measure the degree of market integration.So far,several main methods for measurement has been already generated after continuous improvement by scholars,namely Ravallion Model method (Ravallion Model),correlation coefficient method,the Co-integration test and price limits Model method (Parity Bounds Model),etc.These methods have been widely applied in agricultural market integration and the empirical study of agricultural products logistics efficiency.Kumar[8]investigated 55 middlemen of agricultural products and 60 farmers in Uttar Pradesh Hamirpur area in India to analyze chickpeas circulation channels of spreads and logistics efficiency,showing that the logistics cost occupied a large proportion in agricultural prices,and the cause of high logistics costs included poor transportation equipment,lack of storage facilities,excessive market regulation.Chahal[9],using a similar approach,carried out investigation and analysis of 10 large scale,15 middle scale and 26 small scale fishermen’s organizations in Punjab (Indian) region.The research results indicated that the fishermen,implementing the contract marketing and cooperation marketing ways by taking part in rural cooperative organizations,can improve logistics efficiency of aquatic products and avoid the middlemen excessive exploitation.

Many other scholars also studied the influence factors of the agricultural logistics efficiency and promotion countermeasures.It is believed that long agricultural product circulation channels[10-11]low degree of marketization of agricultural products[2,12],low degree of agricultural product circulation organization[13],poor agricultural product logistics infrastructure[14-15]have restricted the agricultural product logistics speed and efficiency.Therefore,some scholars suggest shortening the circulation channel to reduce the loss of agricultural products in circulation[16-18].Besides,there are many ways of increasing the agricultural logistics efficiency,such as agricultural vertical coordination,chain’s upstream and downstream cooperation of agricultural industry,utilization of logistics technology[19-21].

The main disadvantage of previous studies that analyses existing agriculture logistics is that they more or less neglected the agriculture products logistics on a national scale from the perspective of the rural areas as well as its intertwined connection with globalization effects.In view of this,we tried to make overall measurement ofALEunder the framework of output distance function with related national macro data and explore the impacts of globalization onALEin the context of China.

3 Measurement of rural agriculture logistics efficiency (ALE) in China

3.1MethodandmodelAccording to the Production Economics theory,the traditional Production Function can visually express the technology relationship between input and output of a certain industry or enterprise.However,there are inherent defects and limitations.First of all,the transfer function is probably not smooth and continuous,so that it can’t be solved by using calculus techniques; Second,dealing with multi-objective optimization problem is also quite difficult,even the calculus method is introduced; the most important thing is that the traditional production function approach assumes that production has been conducted on the production frontier all the time,namely,all enterprises’ technologies are efficient.But the truth is that most production in real economy deviates from the optimal production boundary because of various reasons,which means the non-efficiency of production extensively exits.Therefore,we will introduce the method of distance function to measure efficiency of agriculture logistics in the context of China.

3.1.1Research method—Distance Function.Malmquist[22]and Shephard[23-24]respectively put forward the concept of distance function which is not widely exploited until recent decades.We neither indicated the specific behavior object and the specific production function expression,nor made assumption of efficiency distribution,but applied the distance function to measure the relationship between input and output in this study.The agriculture logistics production frontier can only be estimated based on volume samples from various areas of China related to rural agriculture logistics.Therefore,this method requires a lot of sample data and complicated calculation.

According to Shephard[24],the output distance function is utilized to measure the distance between production point of producers and production possibilities frontier.Regional agriculture logistics output distance function is defined by the agriculture logistics output setP(xk).

Agriculture logistics output distance function can be defined by the set ofP(xk):

DO(xk,yk)=inf{μ∶μ∈P(xk)}=sup{μ∶μyk∈P(xk)}-1

Then,ALEis determined by the following function as per Fare[25]:

ALE=D0(xk,yk)

whereALEis an index valued between 0 and 1.According to this definition,the value of region’s agriculture logistics output distance function becomes 1 when an observation point is just right on the production possibilities frontier.At this time,theALEequals 1 (or 100%).

3.1.2Model specification.Since concrete product function is dispensable,we adopted the method of linear programming,building a nonparametric piecewise linear production frontier with whichALEfor each region can be calculated respectively.

Suppose time seriest=1,2,…,T,regionk=1,2,…,K,and assuming constant returns to scale (CRS).RegionalALEcan be obtained by the model with non-parametric linear programming as follows:

3.1.3Sample data.To measureALEprecisely,data from 28 provinces and regions in China with 21 years’ time span is gathered to comprise a total of 588 sample points.The averages were employed for missing value of certain variables in observation samples to develop a balanced panel data.The volume of region rural agriculture products circulation was collected as output index of rural agriculture logistics,while setting capital stock and employee quantity related to rural agriculture logistics as the indicators for agriculture logistics inputs.The volume of region rural agriculture products circulation is represented by the product of rural residents family average main agriculture products (grain,oil,fruit,vegetables) sold per person and the amount of regional rural population; the productive original value of fixed assets of rural residents family (including transportation facility value) is taken as the capital stock.The above data were derived from the China statistical yearbooks and statistical yearbooks throughout the years in various provinces.Meanwhile,the data of rural logistics employees came from China rural statistical yearbooks and China logistics yearbooks.As to variables affected by price fluctuations,the deflator price index of each year for each province was applied to translate the nominal value into real one.

Table 1 ALE rankings of regions in China

As can be seen from the above distribution,the regions with high and relatively highALEin China concentrate in the Central East of China,and the worst six areas are all located on western China with the lowestALEvalues which are all less than 0.3.Shandong,Guangdong,Jiangsu,and Zhejiang stay in a position of relatively high efficiency or above during 1994-2014.Shanghai’sALEbegan to decline after 2004; in contrast,Henan upgraded its ranking to the second level in 2004.

From the point of static comparative analysis,China’s rural areas with high and relatively highALE,a total of five,account for a little below 18% of the whole country; the number of regions withALEvalue below 0.6 is 12,accounted for 43%,or nearly half of the country’s regions have failed inALE.Among them the lowest six regions withALEvalue below 0.3 take up more than 21% of the total; there are 11 provinces and regions where theirALEare between 0.6 and 0.7,accounting for 39% of China.It is concluded that the overall level of China’s ruralALEis not high,even tend to be lower than common degree (fair condition).

However,there is no denying the fact that a general trend of the averageALEof the entire nation has been gradually increasing.The national averageALEvalue and coefficient of variation are shown in Table 2 and Fig.1.

With China’s accession to the WTO in 2001,foreign agricultural products began to enter China market while a large number of domestic agricultural products were sold abroad due to substantially dropped agricultural tariffs.In addition,China’s agricultural output steady rose after 2000,which multiplied quantity of agricultural products to sell as well.So,continuousALEimprovement was required to meet rapid growth in volume of agricultural products in China.

At end of the transition period of China’s accession to WTO by 2004,China’s agricultural product market experienced a comprehensive opening to outside world and foreign agricultural products entered into China in volume quantities,which caused a great impact to farmers in China.Farmers had to halt some production and adjust product structure.This is the reason that China’s ruralALEvalue declined in 2004 and 2005.But in 2006,theALEhas a substantial increase by 9.5% year-on-year,and began to enter orbit to rise since then.Through years of reform,market economy has been penetrating to China rural areas where people spontaneously organize some leading enterprises,industry association and the wholesale market to improve the degree of systematization themselves,which in turn promoted the regions’ALE.

From the perspective of the difference between regionalALE,a development trend is shown that the coefficient of variation experienced a rise before slowly decline as per Fig.1.Since 2005,regionalALEdifferences began to shrink.With more profound understanding of market economic orientation widely obtained,regional government started to encourage,rather than obstacle,the opening of the agricultural market; in the meantime,the linkage between domestic and foreign market has been getting closer,accompanied by unceasing enhancement of regional division of labor and specialization level in rural areas.All above change theALEdifference between regions with a tendency to gradually become smaller,but the speed of changing remains slow.

Table 2 National average ALE value and coefficient of variation in China during 1994-2014

Fig.1NationalaverageALEvalueandcoefficientofvariationinChinaduring1994-2014

4 Impacts of globalization on ALE of China

4.1MethodandmodelThere are factors such as agricultural development cycle,the regional characteristics of heterogeneity exerting influence on theALE.The accession to the WTO is seen as a milestone in the process of the economic globalization of China.Wrong conclusions could be drawn if the WTO impact on China’sALEwere evaluated simply according to the changes ofALEvalue before and after the WTO accession.For that reason,we adopted two measurement methods,two-way fixed effects panel data method and difference in differences method,to obtain as ideal as possible estimates of the WTO impact on China’sALE.Both methods can effectively control estimation bias triggered by factors changing with the individual or time.In addition,we can test the reliability of the article’s conclusion by comparing estimate results the two methods bring in.

4.1.1Two-way fixed effects model.Two-way fixed effects model uses virtual variables which reflect the individual characteristics and time characteristics to describe some certainty information that is contained in non-observable factors affecting the dependent variable.The impact of these factors,therefore,can be controlled and separated by employing fixed effects model.A two-way fixed effects panel data model containing various virtual variables of area and time is established as follows:

(1)

where timet=1,2,3,…,T,regioni=1,2,3,…,I,the number of explanatory variablesm=1,2,3,…,M,yitstands for the ACE value of the regioniat the timet;xmis a set of observable interpretation variables;Witis a dummy variable representing the WTO accession (before the WTO accessionWit=0,after the WTO accessionWit=1);λishows fixed effects of the regioniwhileγtis fixed effects of the timet.λicontrols the individual characteristics,such as the regional heterogeneity,that do not change with time andγtcontrols the time characteristics that don’t vary with individual,such as agricultural development stage,macro economic conditions and government policy.μitis random error with assumption that there is no correlation betweenμitandALE.The most important coefficient isαthat measures the average change of ChineseALEbefore and after the WTO accession.

In model (1),αmeasures the direct impact of the WTO onALE,because the indirect impact caused by explanatory variables has been separated.If we remove the explanatory variables from model (1) ,would be estimated as total impact onALE(the sum of direct effect and indirect effect).The model takes the form:

yit=αWit+λi+γi+μit

(2)

(3)

Likewise,if the explanatory variables are removed from model (3),we can obtain the WTO’s general dynamic effect onALE.The model is written as:

(4)

4.1.2Difference in differences (DID) method.The DID method has been widely applied to solve the problem of policy effect evaluation since its first use by Ashenfelter and Card[26].This method becomes a useful tool for observing the change of a treatment group and control group before and after an issued policy to determine whether the policy has an effect,because introducing the control group into model can effectively separate the effects of other variables except policy,that makes policy effect judgment more accurate.The accession to the WTO is an important milestone in the integration of the Chinese economy into the global market.It also can be regarded as an important policy decision in China.

The DID estimation equation is established as follows:

(5)

wherejis group category (treatment groupj=1,control groupj=0);djis region virtual variable (dj=1 whenj=1 anddj=0 whenj=0);dtis time virtual variable (before WTO accessiondt=0,after WTO accessiondt=1);xmare other explanatory variables; parameterα1represents change of the two groups over time whileα2embodies the change without time.α3is the estimated difference in difference results reflecting the effect of the policy change.

The choice of control group is crucial before estimation process.This problem can be converted into estimation of theALElevel in China assuming no WTO case.A feasible solution is to simulate prediction ofALEvalues and other variables of each year after the WTO (2002-2014) through the time series of exponential smoothing based on data ofALEvalues and other variables in China before the WTO accession (2001).Series of predicted values obtained can be considered to be simulation ofALEand variables in condition of assuming no WTO case in China.

4.2VariablesanddataprocessingSince 1994,China’s agricultural products trading system has been making transition from planned economy to market economy.So the time starting point for samples collected is the year of 1994.We set theALEvalues of each province during the calendar years as explained variable.Other explanatory variables include: (i) rural logistics infrastructure,expressed as villages traffic smooth degree; (ii) quality of rural human resource,expressed as average educated time of rural labor force; (iii) rural informatization level,expressed as telephone penetration and Internet penetration on villages and township in countryside; (iv) rural agriculture logistics professional level,expressed as the percentage of rural agriculture logistics personnel to total rural employment; (v) level of agriculture labor division,expressed as the ratio of rural agricultural labor force to rural non-labor force.

Based on above,provincial level panel data with total of 21 years from1994 to 2014 are collected and derived fromChinaStatisticalYearbook,ChinaRuralStatisticalYearbook,ChinaLogisticsYearbook,China’sAgriculturalDevelopmentReport,andChinaAgriculturalStatisticsforFiftyYears.Statistical descriptions of the variables are listed in Table 3.

Table 3 Statistical descriptions of variables

4.3Resultsoftwo-wayfixedeffectsmodel

4.3.1Estimation results.Estimation results of the models (1),(2),(3),and (4) are shown in Table 4.The estimated results of models (1) and (2) embody average impact caused by the WTO accession on China’sALE,while the results estimated from models (3) and (4) reflect the dynamic influence of WTO on China’sALE.

4.3.2Average effects analysis.The first two columns of Table 4 report the WTO’s average impact on China’sALE,of which column (2) contains the total impact and the second one (1) shows the direct impact with controlling indirect impact caused by explanatory variables.From the estimated results,the WTO makes the overall value ofALEincreased by about 0.146 units,with an average increase of 34.2% over the previous WTO.It is a relatively significant influence,which shows that the WTO has indeed promoted the upgrading of China’sALE.For other explanatory variables,estimation results indicates that the rural logistics infrastructure,quality of rural HR,rural information level and district agricultural division have significant positive influence on theALE.Division of labor can promote agriculture growth by improving the level of the specialization of agricultural production and business efficiency,and then help to expediteALE.It is also concluded that the influence of the rural agricultural logistics specialization level is weak so far in China.

Table 4 Estimation results of two-way fixed effects panel model

Note: figures in parentheses are fortstatistic;*,**,***respectively stand for 10%,5%,and 1% significance level.

4.3.3Dynamic effect analysis.Columns (3) and (4) in Table 4 separately indicate the WTO’s dynamic total impact and dynamic direct effects onALE,so we can easily understand the progressive influence of WTO toALEand its implementation speed.The estimated symbols for all explanatory variables are consistent with the static regression.Logistics infrastructure in rural areas,the rural labor force quality,rural information level and agricultural division perform significantly,while the professional level of the rural agricultural logistics is still not significant.The influence of WTO in the year of 2001 was not obvious,yet it increased for the next year (2002),and expanded year by year.By the fifth year of China’s accession to WTO (2006),its effects bringALEsignificant improvement with 0.104 units variable coefficient,and then 0.112 units in the sixth year.It reached 0.131 units in the seventh year,although the influence declined during the next two years,0.127 and 0.119 units respectively.WTO’s influence on theALEbegan to significantly appear when China fully opened its agricultural products market to outside after 2004.Estimation results and the above analysis shows that the effects of WTO on China’sALEown obvious progressive dynamics,and it will continue to work for a long time.

4.4ResultsofDIDmodelThe relevant data before the WTO entry is exploited to simulate the value ofALEand other explanatory variables of 2002-2014 (after WTO accession) by exponential smoothing of time series.The predicted values with the previous data before 2001 are composed of a new simulation panel data which can be considered as an alternative to control group in the absence of real data.

4.4.1Empirical results and description.The model (5) is estimated by using the simulation results.In order to test the robustness ofALEin time,we adopt the way of gradually increasing the variable to express the influence of the explanatory variables on the dependent variables.In addition,considering the time lag effect,the model is divided into long-term model and short-term one.The long-term period is 1994-2001 and 2002-2014 and the short term refers to 1999-2001 and 2002-2004.

According to the estimation results in Table 5,the explanatory variables are not significantly changed by the change of the model variables,and the conclusion of this paper is proved to be robust to a certain extent.

The most important parameter estimation in Table 5 is the DID estimation results that reflect the influence of WTO toALEin China.The impact of the policy changes is embodied by the coefficient (that is ) in front of the DID variables (i.e.the time dummy variables and regional dummy variables cross terms).The estimated value of reflects the change of theALEgeneration before and after the WTO accession.

According to Table 5,estimation results of the coefficient for both long-term and short-term model are significant positive,indicating that WTO accession has indeed promoted the growth ofALEin China.But Long-term model estimation (0.066) is significantly higher than that of short-term model.In the sense of average,the long-term effects of WTO accession onALEis about 0.030 units more than short-term effect,increased by 83.3%.This shows clearly that the impact of the WTO accession to theALEin China is not temporary,and its influence has a long-term evolutionary nature,especially after 2004.The results keep consistent with that of dynamic model of the fixed effects.Therefore,they fully validate the reliability of our conclusions.

4.4.2Analysis of mechanism based on DID.In order to test mechanisms through which the WTO affectsALE,we further used the DID method to estimate the impact of WTO on other explanatory variables.These variables include the rural logistics infrastructure,the quality of rural human resources,the level of rural informatization,the division of agriculture in region and the specialized level of agriculture logistics.We employed the long-term model to estimate two stages (before and after the WTO accession) respectively for a better reflection of the trend and variation.The results are shown in Table 6.

Table 5 DID model estimation results

Note: the figures in parentheses are fortstatistic;*,**,***respectively stand for 10%,5%,1% significance level.

Table 6 The DID estimation results of explanatory variables

The last column of Table 6 reports the DID estimation results of these variables.They reflect the degree of effect caused by the WTO accession on these variables.From Table 6,it can be seen that the entry into WTO has a significant impact on the rural logistics infrastructure,the level of rural informatization,the quality of rural labor and the level of agricultural division in China,and the improvement of agriculture logistics professional level is not obvious.

Entering WTO is a milestone in the opening up of China’s agriculture,and it is important to improve the degree of agricultural marketization in addition to promoting the reform of agriculture and the adjustment of agricultural policy.This has forced us to increase the efficiency of circulation in the course of the agricultural products transaction,and the improvement of the level of logistics infrastructure,expanding the supply of agricultural products in rural areas and enhancing the radiation ability of market,are the most direct ways to boost the efficiency of agriculture logistics.DID value of rural logistics infrastructure is 0.025,which is located in the first of all variables,showing that China’s rural logistics infrastructure has relatively most significant improvement after the WTO accession.The rapid development of agricultural informatization brought about by the opening up of agriculture,with improvement of the rural information exchange between agricultural enterprises and farmers,has greatly raised the efficiency of agriculture transaction and logistics.In addition,in recent years,China has accelerated the optimization of agricultural structure adjustment and regional layout.The regional division of agriculture has continued deepening to speed up the circulation of agricultural products.Meanwhile,the WTO accession also contributes to the improvement of the rural labor force quality,accordingly enhance the quality of agriculture logistics service and reduce the related management cost.The results in Table 6 show that the WTO makes quality of rural human resource in China increased by about 46.7%.

Small-scale production known as one of significant characteristic of China’s agriculture,for the moment,remains unchanged,which leads to a tiny difference of agriculture logistics specialization level before and after the WTO accession,say only 0.001.The low level of agriculture logistics specialization has caused the majority of small-scale producers not be able to have a good grasp of market opportunities,which restricts further promotion ofALE.

According to the results of Table 6,we can draw the conclusion that the WTO accession promotes growth of China’sALEand the motive mainly comes from China’s investment and construction of rural logistics infrastructure,rural informatization level,regional agricultural division of labor and rural labor quality,but China’s agriculture logistics professional level is still low without significant improvement after the WTO accession.

5 Conclusions

(i) TheALEin China has been on the rise in the overall level since 2000; regional variations in theALEdifferences show a trend that is from expanding to lessening.The regionalALEdifference becomes gradually narrowed with economic and social development of China,but the narrow speed is getting slow.

(ii) Globalization promotes China’sALEto rise.The WTO accession has made China’sALEvalue higher,on average,than that of pre-accession by about 34.2% in overall.This promotion role in China is not merely of the immediate,but of long-term with an obvious dynamic progressive characteristic.Long-term’s influence of globalization onALEin China is 83.3% more than that of the short-term.

(iii) The entry into WTO has not markedly boosted the China’s agriculture logistics professional level,but there is a significant improvement,after the WTO accession,in the rural logistics infrastructure,the rural informatization level,the division of labor and the quality of rural labor,which indeed contribute to promote the improvement ofALEin China.

During the globalization process,China has been participating in agriculture cooperation and labor division by increasing alignment with the world agriculture.In the meantime,China’s agriculture is also facing fierce external competition.Efforts to improve the comprehensive agricultural production capacity and competitiveness are the important measures to meet the challenges,and high efficiency of agriculture logistics system is one of the most important support and guarantee to enhance the agricultural competitiveness in China.The models and the results could be available and useful for those who just has interests in knowledge about situation of agriculture logistics in China and who need to deal with related policy-making.