31 7 Vol.31 No. 7 2011 7 ECONOMIC GEOGRAPHY Jul. 2011 1000-8462(2011)07-1178 - 07 DEA 1 2 1 1 2 1 2 1. 100101 2. 100049 DEA 2009 DEA F304.7 A 1980 DEA 2010 7 8 1 DEA 8 [1] DEA [6] DEA [7] 1 [2] DEA Decision [3] Making Unit, DMU 2 SFA Data Envelopment Analysis DEA 3 DEA DEA DEA C 2 R BCC [4-5] DEA 1978 A. Charnes W. SFA W. Cooper E. Rhodes CRS Constant Fried Lovell Schmidt Yaisawarng Return Scale C 2 R DEA DMU DEA 2011-01 - 26; 2011-04 - 22 2007BAD80B03
7 1179 DEA Front4.2 β n μ n 2 2 δ v n δu n [8] Banker Charles Cooper VRS 2 2 Variable Return Scale C 2 R β 赞 n μ 赞 n δ 赞 v δ 赞 [12] n u n Jondrow BCC [9] u n i E 赞赞 u n i v n i +u n i 赞 v n i C 2 R BCC v 赞 n i [13] TE CRS VRS DEA TE VRS E 赞赞 v n i v n i +u n i 赞 =S n i - z i β 赞 n -E 赞赞 u n i v n i +u n i 赞 3 CRS DEA SE A t t+1 x n i = x n i + 赞 Maxi {z i β 赞 n }-z i β 赞 n 赞 + 赞 Max i {v n i }-v n i 赞 n=1 2 N i=1 2 I 4 TE CRS(t) = TE VRS SE [10] A x TE VRS = D v t+1 (xt+1 y t+1 ) SE= D v v n i xn i t (xt y t ) D t+1 (xt+1 y t+1 ) v c c D t (xt y t 赞 ) D t (xt y t ) D t+1 (xt+1 y t+1 赞 z i β 赞 n ) v n i Max i {z i β 赞 n } D c (x t y t ) DMU DEA 赞 Max i {z i β 赞 n }-z i β 赞 n 赞 DMU C 2 R BCC 赞 Max i {v n i }-v n i 赞 DEA- Solver Pro 5 DMU DMU SFA Fried DEA A DEA / x n i x n i C 2 R BCC DEA DMU SFA [11] DMU I 2 S n i =x n i - Σλ i x n i 0 n=1 2 N i=1 2 I 1 i=1 2.1 S n i DEA i n 2010 7 8 x n i 3 SFA 1 S n i =f n (z i β n )+v n i +u n i n=1 2 N i=1 2 I 2 18 z i =(z 1i z 2i z ni ) i=1 2 I N f n (z i β n ) S n i v n i +u n i 9 v n i u n i v n i 12 2 u n i v n i ~N(0 δ v n ) u n i ~N(μ n 2 δ u n) 163 1 2 2 2 γ=δ u n δ v n+δu n 1 2 2 2 γ=δ u n δ v n+δu n 0 2009 Battese Coelli
1180 31 Fig.1 2.2 1 The sketch map of investigation route 1 2 3 8 4 2.2.1 DEA 1 1 DEA Tab.1 Raw data of input and output indicators of DEA model / / / / / /kg / LA 7.3 0.6 24 57.1 519.9 4 418.9 1 425.3 XC 11.4 0.0 47 62.2 624.6 4 759.8 2 183.1 JX 11.8 1.6 38 61.4 566.5 4 215.2 2 936.9 NC 40.2 3.1 96 286.8 1 718.3 9 620.6 4 350.2 QR 13.6 1.1 49 109.3 788.7 9 320.7 1 533.3 QJ 22.7 1.8 48 156.7 1 485.2 5 633.8 5 150.6 LD 27.4 0.7 81 178.2 1 624.3 6 150.5 984.5 DJ 57.5 13.1 68 407.3 4 917.9 15 980.4 4 325.4 GR 24.2 1.7 86 182.3 1 554.6 6 392.8 3 521.6 ND 6.6 0.9 41 394.6 538.9 3 751.2 4 646.4 KL 12.8 1.3 80 82.5 373.6 4 450.1 4 198.7 XS 12.9 1.3 33 73.8 315.2 1 786.4 1 382.9 20.7 2.3 58 171.1 1 252.3 6 373.4 3 053.2 DEA [14] MATLAB Kendall s tau- b 7 2 5% 2.2.2 2 4 3 7% [15-16] 1 2 7 Kendall(tab- u) Tab.2 Kendall(tab- u) statistical value of 7 variables tab- u - 0.036 0.165-0.014 0.092-0.096 0.228 0.315 P 0.029 0.013 0.042 0.007 0.037 0.006 0.019 tau- b Kendall
7 1181 4 163 TE CRS TE VRS SE Tab.4 Total technical efficiency, technical efficiency and scale efficiency of 163 samples in the first stage [17-18] TE CRS TE VRS SE LA 0.419 0.583 0.719 Irs XC 0.468 0.692 0.677 Irs 3 JX 1.000 1.000 1.000 - NC 0.454 0.617 0.736 Drs QR 0.358 0.558 0.642 Irs QJ 0.326 0.534 0.604 Irs [6 19] LD 0.414 0.569 0.728 Drs DJ 1.000 1.000 1.000 Irs GR 0.407 0.586 0.695 Drs [20] ND 0.486 0.774 0.628 Irs KL 1.000 1.000 1.000 - XS 1.000 1.000 1.000-3 irs drs - Tab.3 The statistical attribute of environmental variables 461.0 189.0 217.3 38.3 [0.786/ 0.743+0.786 ] 16.0 0.0 7.7 9.1 83 500 870 3 925.8 415.6 3 3.2 SFA 3.1 DEA DEA 163 C 2 R- I BCC- I 3 DEA- Solver FRONTIER SFA 5 4 5 2 2 2 DEA γ δ u n(δv n δ u n ) 12 0 1% 0.4 0.7 LR 5% SFA 0.743 0.786 163 0.584 0.743 0.786 41.6% 1-0.584 SFA 51.4% Tab.5 5 SFA The results of stochastic frontier analysis 7.314*** 1.296** 13.058** 17.227*** 36.825** 0.058** 0.003* 0.013 0.002 2.741*** - 0.218** 0.094-0.106-0.121*** 12.843*** - 2.573** 0.039-0.005* - 18.996*** 13.372*** δ 2 45.336*** 7.761** 4.309*** 118.275*** 53.668*** γ 0.306 0.009* 0.001*** 0.000*** 0.000*** Log- likelihood function - 198.337** - 206.512-223.916-213.798** - 217.056* LR test of the one- sided error 9.493 11.782 15.954 11.423 16.848 * ** *** 10% 5% 1%
1182 31 0.743 0.875 [21] 58.2%[0.875/ 0.875+0.628 ] DEA DEA SFA DMU 20.5% 3.3 DEA SFA DEA- Solver C 2 R- I BCC- I 4 DEA 6 6 163 Tab.6 The true efficiency of 163 samples in the third stage TE CRS TE VRS SE LA 0.419 0.583 0.719 Irs 2010 LA 0.278 0.837 0.332 Irs DEA XC 0.384 0.881 0.436 Irs JX 1.000 1.000 1.000 - NC 1.000 1.000 1.000 - QR 0.222 0.709 0.313 Drs QJ 0.347 0.835 0.416 Irs C 2 R I 12 LD 0.275 0.852 0.323 Drs 7 8 DJ 1.000 1.000 1.000 - GR 0.226 0.586 0.385 Irs ND 1.000 1.000 1.000 - KL 0.261 0.799 0.327 Irs 8 XS 1.000 1.000 1.000 - Tab.8 The slack of input indicators of non- DEA efficiency samples DEA Wilcoxon H 0 H 1 7 218.82 487.51 Z - 0.314 p 0.029 α 0.05 p 0.786 0.628 4 / / / / LA 2.1 1.6 13.3 29.2 235.8 XC 4.6 2.9 11.4 32.3 208.1 QR 3.7 2.4 15.6 18.2 193.5 QJ 5.3 4.4 18.7 21.9 133.6 LD 6.2 5.8 12.4 30.1 174.8 GR 3.6 3.1 9.5 17.4 168.7 α KL 2.9 2.4 16.2 24.1 79.3 4.1 4 000m 4 6 163 0.584 0.582 0.875 0.628 8 12 7 Wilcoxon Tab.7 The results of Wilcoxon sign- rank test N Z 14 a 14.86 218.82-0.314 0.029 22 b 22.17 487.51 36 a < b > Z
7 1183 163 25.9 37.2% 25.7% 314.7% 23.1-13.5% 10 Tab.10 The optimization proposal on labor input of non- DEA efficiency samples 26 hm 2 / 25.22 hm 2 /% 378.25 4.2 KL 82.5 106.6 24.1 29.2 9 12 163 20 40 71.4% 9 Tab.9 The input and utilization of main material cost of non- DEA efficiency samples / / /% <10 LA 519.9 235.8 45.4 0.33 10 20 XC 624.6 208.1 33.3 0.44 QR 788.7 193.5 24.5 0.31 KL 373.6 179.3 47.9 0.33 >20 QJ 1 485.2 133.6 9.0 0.41 LD 1 624.3 174.8 10.8 0.32 GR 1 554.6 168.7 10.6 0.39 LA 57.1 86.3 29.2 51.1 XC 62.2 94.5 32.3 51.9 QR 109.3 127.5 18.2 16.7 QJ 156.7 134.8-21.9-14.0 LD 178.2 148.1-30.1-16.9 GR 182.3 164.9-17.4-9.5 5 9 10 45% DEA 263 0.33 20 2009 10.2% 163 0.35 U [22-23] DEA 4.3 80% 10 [1].
1184 31 [J]. 2005(7) 72-81. sis[j]. Journal of Productivity Analysis 2002(17) 157-174. [2] And row Fisher. State growth and Social Exclusion Recent Economic Growth in Tibet[J].The 10th International Conference of the International Association of Tibetan Studies Oxford UK 2003. [12] Battese and Coelli. Frontier Production Functions Technical Efficiency and Panel Data with Application to Paddy Farmers in India[J]. Journal of Productivity Analysis 1992(9) 38-75. [3] Simar L Wilson P. Estimation and Inference in Two-stage Semiparametric Models of Production Processes[J]. Journal of Econometrics 2007(136) 31-64. [13] James Jondrow. On the Estimation of Technical Inefficiency in the Stochastic Frontier Production Function Model[J]. Journal of Economics 1982(19) 233-238. [4] Pekka J Korhonen Pyry-Antti Siitari. A dimensional decomposition approach to identifying efficient units in large-scale DEA models[j]. Computer&Operation Research 2009(36) 234-244. [14] [15]. [J]. 1997(6) 35-38.. [5] Preeti Tyagi Shiv Prasad Yadav S P Singh. Relative Performance of Academic Departments using DEA with Sensitivity Analysis[J]. Evaluation and Program Planning 2009(32) 168-177. [16] [J]. 2005 23(4) 620-624.. [6]. [J]. 2003(1) 66-68. DEA [J]. [17]. 2009 35(9) 92-102. [J]. ( ) 2005 35(6) 105-112. [7] Mccarty T Yaisawarng S. The Measurement of Productivity Effi- [18]. ciency Techniques and Applications[M]. New York Oxford Unversity Press 1993. [8] Charnes A Cooper W W Rhodes E. Measuring the Efficiency of Decision Making Units [J]. European Journal of Operational Research 1978(2) 429-444. [9] Banker R D Charnes A Cooper W W. Some Models for Estimating Technical and Scale Inefficiences in Data Envelopment Analysis[J]. Management Science 1984(8) 533-536. [10] Toshiyuki Sueyoshi. DEA-Discriminant Analysis Methodological Comparison among Eight Discriminant Analysis Approaches [J]. European Journal of Operational Research 2006(169) 247-272. [11] Fried H O Lovell C A K Schmidt S S. Accounting for Environmental Effects and Statistical Noise in Data Envelopment Analy- [J]. 2004 (5) 10-17. [19]. [J]. 2004 23(2) 219-227. [20]. [J]. 2002 57(4) 459-468. [21] Pastor J T. How to Account for Environmental Effects in DEA An Application to Bank Branches [M]. Spain Universidad de Alicante 1995. [22] Tone K. A Slack-Based Measure of Efficiency in Data Envelopment Analysis [J]. European Journal of Operational Research 2001(130) 498-509. [23]. [J]. 2002 12(4) 425-429. THE ANALYSIS AND IMPROVEMENT FOR AGRICULTURAL PRODUCTION EFFICIENCY OF HOUSEHOLDS IN THE YLN REGION OF TIBET BASED ON A THREE- STAGE DEA MODEL AND THE MICRO- DATA OF RURAL HOUSEHOLDS ZHU Fan 1,2,YU Cheng - qun 1,ZENG Rong 1,2,XU Shao - yun 1,2 (1. Institute of Geographical Sciences and Natural Resources Research,Chinese Academy of Sciences,Beijing 100101,China; 2. Graduate University of Chinese Academy of Sciences,Beijing 100049,China) Abstr act: In order to eliminate the adverse effects of environmental and random factors on the production efficiency, the paper makes an empirical study on the agricultural production efficiency of households in the YLN region of Tibet in 2009 through a three-stage DEA model and the micro-data of rural households. The results shown as follows: environmental and random factors, such as the amount of subsidy, the average education years, and nonfarm income of rural households, has a remarkable influence on production efficiency of rural households. Under the homogeneous background, the decision-making and management efficiency of rural households is relatively high, and the regional differences of efficiency are small statistically, the scale efficiency is proved to be the bottleneck of the improvement of the agriculture production efficiency. Then, based on the projection analysis, this study proposes policy recommendations, such as structural readjustment in agriculture, material input optimization and reasonable allocation of labor resources. Key words: three-stage DEA model;agricultural productivity;environmental effect;tibet;ynl region 1985 E-mail:zhuf.igsnrr@tom.com