泰國水稻生產地區效率差異之研究-共同邊界包絡分析法之應用

Measuring Efficiency Differences of Thai Paddy Rice Production Regions: An Application of Metafrontier DEA Approach

王傑(ANUPONG WONGCHAI)

指導教授 ： 彭克仲

國立屏東科技大學/國際學院/熱帶農業暨國際合作系/博士(2013年)

https://doi.org/10.6346/NPUST.2013.00102

摘要

泰國是世界稻米生產輸出國之一，水稻生產在泰國經濟上與勞動力上扮演主要角色。自1959年起，泰國經歷了提升水稻產值的發展政策，而這些政策已在全國各地遍布實施，因而逐漸形成泰國的主要四個不同的稻作生產地區，且各種不同的水稻生產技術。因此，研究目的旨在衡量: (i) 技術效率和共同邊界技術差距比率，(ii) 泰國水稻生產力的變化趨勢，和 (iii) 影響泰國水稻生產超效率值的因素。本研究範圍為將泰國的76個省，劃分為四個主要區域群組。時間序列資料，蒐集自泰國農業經濟辦公室和泰國氣象部 2002/03年到 2011/12年之間收穫年度的數據。
首先，共同邊界資料包絡分析方法與投入主導型的概念被用來揭示共同邊界、 TE 和 TGRs 的值。在DEA模型中，水稻的產值視為產出，而耕地、勞動力、耕耘機具、種子、化肥和農藥則視為投入。
第二，藉由資料包絡法與Malmquist生產力指數來揭示水稻生產力變化的趨勢。該指數可以分解為兩個組成元件，即技術變動和效率 變動。
最後，經由兩階段的超效率 DEA 和 Tobit 回歸模型，來探討哪些因素影響技術效率。經Tobit迴歸模型分析發現有7個因素和泰國的水稻生產技術效率有著密不可分的關係。這些因素包括:家務勞力與家庭的大小的比值、教育水準、水稻耕作訓練、水稻耕作經驗、灌溉土 地、土地所有權及降雨量。
此分析可分為三個主要部分:宏觀層面分析、微觀層面分析以及宏觀和微觀層面的分析。宏觀層面分析乃根據共同邊界，衡量區域效率 值。而微觀層面分析則根據區域邊界，評估省級效率值。宏觀和微觀層面分析則是比較以上兩者的值。
在宏觀層面分析顯示，北方執行了最高的平均共同邊界值為 0.956。然南部則是最低的平均共同邊界值為0.806。東北部和中部分別顯示最高和最低平均 TE值分別為 0.979 和 0.944。而在 TGRs方面，北方為最佳值，南方則是最差值。 另Malmquist生產力指數分析顯示，僅有中部在生產力變動有所進展。與此同時，其他區域則生產力退化。超效率資料包絡分析結果發現，所有區域都有一個或多個高平均效率 值，顯示多數的決策單位具效率。此外，經Tobit 迴歸分析結果發現，增加灌溉土地和降雨量，對增加技術效率值具顯著影響力。
在微觀層面的分析發現，具執行效率的決策單位，北部9個、 東北部8個、 中部17 個、 南部7個。 Malmquist生產力指數指出，進展最佳的狀態要屬：北部的Nakhon Sawan、東北部的Udon Thani、 中部的Ratchaburi 以及南部的Narathiwat。超效率資料包絡分析發現，具超效率最高值的有:位於北部的Phichit、 東北部的Surin、中部的Bangkok和南部的Narathiwat。此外，Tobit 迴歸分析結果發現，在統計學層級的微觀層面有不同的影響力。
在宏觀和微觀的層級分析總結認為，有一半以上的決策單位顯示 TE 值最高。然而，只有7個省的TGRs值較高。另生產率指數的總體因素方面顯示，由於的技術和效率的變動，超過10年的收穫年度生產力變動有退化的趨勢。在此層級的平均超高效率值是大於1。因此，對76 省生產邊界超高效率值而言，7個因素顯示出不同的正面和負面影響。

關鍵字

Malmquist生產力指數 ； 共同邊界資料包絡分析法 ； 泰國水稻生產 ； 技術效率 ； 技術差距比率

並列摘要

Thailand has a strong of paddy rice production that represents a significant portion of the Thai economy and intensive labor force. Since 1959, Thailand has experienced a development of policies oriented to improving the paddy rice production. These policies have been implemented throughout the country, particularly at a regional level. Consequently, technology production of paddy rice is variously different across the four regions of Thailand.
Therefore, the purposes aim to measure (i) technical efficiency (TE) and technology gap ratios (TGRs) with reference to a metafrontier, (ii) trend of Thai paddy rice productivity’s change, and (iii) factors affecting super-efficiency scores in Thai paddy rice production. The research areas consisted of 76 provinces grouped into four main regions. Time-series data were collected from the Office of Agricultural Economics in Thailand and the Meteorological Department in Thailand during the crop year 2002/03 to 2011/12. A total of 76 provinces were conducted as decision making units (DMUs) in this study.
Firstly, a metafrontier data envelopment analysis (DEA) approach with input-orientated concept was utilized to reveal the scores of metafrontier, TE, and TGRs. The paddy rice quantity was considered as an output, together with arable land, labor force, tractor, seed, fertilizer, and pesticide were used as the inputs in the DEA model.
Secondly, the trend in paddy rice productivity’s change over a ten-year period was revealed by DEA Malmquist productivity index. Its index can be decomposed into two components, technical change and efficiency change. The Malmquist productivity index’s score with greater than 1.000, equal to 1.000, and less than 1.000 indicates the productivity progression, productivity stagnation, and productivity regression, respectively.
Finally, the two-step procedures of super-efficiency DEA and Tobit regression model were adopted to investigate what factors affect the super-efficiency scores. Seven factors that mostly associated with the technical efficiency of paddy rice production in Thailand were considered as the independent variables in the Tobit regression model. These factors included the ratio of household labor/household size, education level, rice-farming training, rice-farming experience, irrigated land, land tenure, and rainfall.
The empirical analyses compose three main parts, comprising macro-level analysis, micro-level analysis, and macro and micro-level analyses. The macro-level analysis measures a regional production function with reference to the metafrontier. The micro-level analysis assesses a provincial production function with reference to its regional frontier. The macro and micro-level analyses compare a provincial function with the metafrontier.
The macro-level analysis showed that the North performed the highest average metafrontier score at 0.956. The South was the lowest average metafrontier score at 0.806. The Northeast and the Central revealed the highest and lowest average TE score at 0.979 and 0.944, respectively. For the TGRs, the North was the best score, while the South was the worst score. The Malmquist productivity index showed that only the Central had a productivity progression. Meanwhile, the other regions depicted a productivity regression. The super-efficiency DEA showed that all regions had a high average efficiency score greater than 1.000. The majority DMUs performed efficiently. In addition, the Tobit regression results found that higher amount of irrigated land and rainfall could increase the super-efficiency scores at one percent significant level. Whereas, the lower number of rice-farming training and land tenure could increase the super-efficiency scores at one percent significant level as well.
The micro-level analysis found that a total of nine DMUs in the North, eight DMUs in the Northeast, 17 DMUs in the Central, and seven in the South that performed efficiently. The Malmquist productivity index noted that Nakhon Sawan was the best progression in the North, Udon Thani was the best in the Northeast, Ratchaburi was the best in the Central, and Narathiwat was the best in the South. The super-efficiency DEA model implied that Phichit in the North, Surin in the Northeast, Bangkok in the Central, and Narathiwat in the South were the highest super-efficiency scores that were presented in the different four regions. Moreover, the Tobit regression statistically displayed the coefficient estimates with significant levels.
The macro and micro-level analyses concluded that there were more than a half of DMUs showing the highest TE scores at 1.000. However, only seven DMUs performed the highest TGRs’ scores of 1.000. The total factor productivity index presented a regressed trend during the ten crop years because of a decrease in technical change and efficiency change. An average super-efficiency score at this level was greater than 1.000. Consequently, all seven factors showed the differently positive and negative impacts on the super-efficiency scores in the Tobit regression model derived from the 76 provincial production frontiers.
Although, the paddy rice production in Thailand had a high technical efficiency score, it still revealed the high values of technology gap ratios. In bridging the gaps, all relevant parties including the Thai government, the farmers, the rice research institutes, and development partners are supposed to integrate some practical ways to increase the efficiency of rice production.

並列關鍵字

Malmquist productivity index ； metafrontier data envelopment analysis approach ； paddy rice production ； technical efficiency ； technology gap ratios

參考文獻

Abu, O. 2011. Fertilizer Usage and Technical Efficiency of Rice Farms under Tropical Conditions: A Data Envelopment Analysis (DEA). Journal of Agricultural Science 2(2):83-87.

Aigner, D., C. A. K. Lovell, and P. Schmidt. 1977. Formulation and Estimation of Stochastic Frontier Production Function Models. Journal of Econometrics 6(1):21-37.

Amparo, M. B., G. M. Constantino-Jose, and S. G. Ramon. 2012. Efficiency Analysis of Small Franchise Enterprises through a DEA Metafrontier Model. The Service Industries Journal 32(15):2421-2434.

Andersen, P., and N. C. Petersen. 1993. A Procedure for Ranking Efficient Units in Data Envelopment Analysis. Management Science 39(10):1261-1264.

Bridhikitti, A., and T. J. Overcamp. 2012. Estimation of Southeast Asian Rice Paddy Areas with Different Ecosystems from Moderate-Resolution Satellite Imagery. Agriculture, Ecosystems and Environment 146(1):113-120.