Handbook on Data Envelopment Analysis (eBook)

Handbook on Data 
3 
Preface 7
About the Authors 11
Contents 21
Contributors 23
Chapter 1: Data Envelopment Analysis: History, Models, and Interpretations* 
27 
1.1 Introduction 27
1.2 Background and History 29
1.3 CCR Model 33
1.4 Extensions to the CCR Model 44
1.4.1 Nondiscretionary Inputs and Outputs 44
1.4.2 Categorical Inputs and Outputs 46
1.4.3 Incorporating Judgment or A Priori Knowledge 47
1.4.4 Window Analysis 49
1.5 Allocative and Overall Efficiency 52
1.6 Profit Efficiency 55
1.7 Recent Developments 60
1.8 Conclusions 62
References 62
Chapter 2: 
66 
2.1 Introduction 66
2.2 RTS Approaches with BCC Models 68
2.3 RTS Approaches with CCR Models 73
2.4 Most Productive Scale Size 79
2.5 Additive Models 82
2.6 Multiplicative Models 86
2.7 Summary and Conclusion 91
Appendix 93
References 94
Chapter 3: 
96 
3.1 Introduction 96
3.2 Sensitivity Analysis Approaches 97
3.2.1 Algorithmic Approaches 98
3.2.2 Metric Approaches 98
3.2.3 Multiplier Model Approaches 102
3.2.4 A Two-Stage Alternative 106
3.2.5 Envelopment Approach 109
3.3 Summary and Conclusion 115
References 115
Chapter 4: 
117 
4.1 Introduction 117
4.2 Using Price Information 120
4.3 Reflecting Meaningful Trade-Offs 122
4.4 Incorporating Value Information and Managerial Goals 125
4.5 Choosing From Alternate Optima 129
4.6 Looking for Non-zero Weights 133
4.7 Avoiding Large Differences in the Values of Multipliers 136
4.8 Improving Discrimination and Ranking Units 139
4.9 Conclusions 144
References 146
Chapter 5: 
151 
5.1 Introduction 151
5.2 DEA Technologies 152
5.3 Projecting onto the Frontier 156
5.4 Productivity Indexes 162
5.5 A Dynamic Malmquist Productivity Index 170
References 172
Chapter 6: 
174 
6.1 Introduction 174
6.2 Problem Settings Involving Ordinal Data 175
6.2.1 Ordinal Data in RandD Project Selection 175
6.2.2 Efficiency Performance of Korean Telephone Offices 177
6.3 Modeling Ordinal Data 179
6.3.1 Permissible Worth Vectors 182
6.3.2 Criteria Importance 186
6.4 Solutions to Applications 187
6.4.1 RandD Project Efficiency Evaluation 187
6.4.2 Evaluation of Telephone Office Efficiency 188
6.5 Problem Settings and Issues Involving Qualitative Data 189
6.5.1 Implementation of Robotics: Identifying Efficient Implementers 190
6.5.2 A Fair Model for Aggregating Preferential Votes 190
6.5.3 Multiple Criteria Decision Modeling: Ordinal Data, Criteria Importance, and Criteria Clearness 191
6.5.3.1 Evaluating Vendors for Complex Systems 192
6.5.3.2 Country Risk Evaluation 192
6.5.3.3 Mutual Fund Selection 193
6.5.3.4 Ordinal Data in Multicriteria Modeling: Evaluation in Terms of Subsets of Criteria 193
6.6 Discussion 194
References 194
Chapter 7: 
196 
7.1 Congestion 196
7.2 Comparison of Two Literatures on Congestion 201
7.3 Färe, Grosskopf, and Lovell (FGL) Approach 201
7.4 Cooper, Thompson, and Thrall (CTT) Approach 205
7.4.1 A Numerical Example 207
7.5 A Unified Additive Model 209
7.6 Estimating the Output Effects of Congestion 211
7.7 Extensions 214
References 215
Chapter 8: 
217 
8.1 Introduction 217
8.2 The SBM Model 218
8.2.1 Production Possibility Set 218
8.2.2 Input-Oriented SBM 219
8.2.3 Output-Oriented SBM 220
8.2.4 Nonoriented SBM 221
8.2.5 An Illustrative Example of SBM Models 222
8.2.6 The Dual Program of the SBM Model 223
8.3 Extensions of the SBM Model 224
8.3.1 Variable Returns-to-Scale Model 224
8.3.2 Weighted-SBM Model 225
8.3.3 Super-SBM Model 226
8.3.4 An Illustrative Example of Super-SBM Models 227
8.4 Further Extensions 227
8.4.1 Dealing with Nonpositive Data in the SBM Models 227
8.4.2 Variations of the SBM Models 229
8.4.3 A Compromise of Radial and Nonradial Measures of Efficiency 230
8.5 Concluding Remarks 230
References 231
Chapter 9: 
232 
9.1 Introduction 232
9.2 Efficiency and Efficiency Dominance 233
9.3 Stochastic Dominance and Joint Chance Constrained Efficiency 236
9.3.1 Potential Uses 239
9.4 Stochastic Efficiency in Marginal Chance Constrained Models 244
9.5 Satisficing DEA Models 252
9.6 Concluding Remarks 258
References 259
Chapter 10: 
262 
10.1 Introduction 262
10.2 Efficiency and the Theory of the Firm 263
10.3 Estimation 265
10.4 A Statistical Model 268
10.5 Some Asymptotic Results 269
10.6 Bootstrapping in DEA/FDH Models 271
10.7 Implementing the Bootstrap 274
10.8 Monte Carlo Evidence 279
10.9 Enhancing the Performance of the Bootstrap 287
10.10 Conclusions 289
References 290
Chapter 11: 
293 
11.1 Introduction 293
11.2 Hypothesis Tests When Inefficiency is the Only Stochastic Variable 295
11.2.1 Statistical Foundation for DEA 295
11.2.2 Efficiency Comparison of Two Groups of DMUs 296
11.2.3 Tests of Returns to Scale 297
11.2.4 Tests of Allocative Efficiency 299
11.2.5 Tests of Input Separability 301
11.3 Hypothesis Tests for Situations Characterized by Shifts in Frontier 302
11.4 Hypothesis Tests for Composed Error Situations 306
11.4.1 Tests for Efficiency Comparison 307
11.4.2 Tests for Evaluating the Impact of Contextual Variables on Efficiency 308
11.4.3 Tests for Evaluating the Adequacy of Parametric Functional Forms 311
11.5 Concluding Remarks 313
References 314
Chapter 12: 
316 
12.1 Introduction 316
12.2 Two-Stage Processes 318
12.3 Centralized Model 319
12.4 Stackelberg Game 322
12.5 DEA Model for General Multistage Serial Processes Via Additive Efficiency Decomposition 325
12.6 General Multistage Processes 328
12.6.1 Parallel Processes 328
12.6.2 Nonimmediate Successor Flows 329
12.7 Conclusions 330
References 331
Chapter 13: 
333 
13.1 Introduction 333
13.2 Performance Measurement Approaches in Banking 334
13.2.1 Ratio Analysis 334
13.2.2 Frontier Efficiency Methodologies 335
13.2.3 Other Performance Evaluation Methods 336
13.3 Data Envelopment Analysis in Banking 336
13.3.1 Banking Corporations 337
13.3.1.1 In-Country 337
13.3.1.2 Cross-Country Studies 338
13.3.2 Bank Branches 339
13.3.2.1 Small Number of Branches 339
13.3.2.2 Large Number of Branches 340
13.3.2.3 Branch Studies Incorporating Service Quality 341
13.3.2.4 Unusual Banking Applications of DEA 341
13.4 Model Building Considerations 342
13.4.1 Approaching the Problem 342
13.4.2 Input or Output? 342
13.4.3 Too Few DMUs/Too Many Variables 343
13.4.4 Relationships and Proxies 344
13.4.5 Outliers 344
13.4.6 Zero or Blank? 345
13.4.7 Size Does Matter 345
13.4.8 Too Many DMUs on the Frontier 346
13.4.9 Environmental Factors 346
13.4.10 Service Quality 347
13.4.11 Validating Results 347
13.5 Banks as DMUS 347
13.5.1 Cross-Country/Region Comparisons 348
13.5.2 Bank Mergers 349
13.5.2.1 Selecting Pairs of Branch Units for Merger Evaluation 351
13.5.2.2 Defining a Strategy for Hypothetically Merging Two Bank Branches 351
13.5.2.3 Developing Models for Evaluating the Overall Performance of Merged Units Through the Selection of Appropriate Input and Output Variables 
352 
13.5.2.4 Calculating Potential Efficiency Gains 352
13.5.2.5 Identifying Differences in Cultural Environments Between the Merging Banks 352
13.5.2.6 Calculating Potential Synergies 353
13.5.3 Temporal Studies 354
13.5.3.1 The Models 354
13.5.3.2 Window Analysis 355
13.5.3.3 Malmquist Productivity Index 357
13.6 Bank Branches as DMUS 361
13.6.1 The Production Model 362
13.6.2 Profitability Model 363
13.6.3 Intermediation Model 364
13.6.4 Model Results 365
13.6.5 Senior Management Concerns 366
13.6.6 A Two-Stage Process 367
13.6.7 Targeted Analysis 368
13.6.8 New Role of Bank Branch Analysis 369
13.6.9 Environmental Effects 370
13.7 Validation 372
13.7.1 Validating a Method with Monte Carlo Simulation 372
13.8 Conclusions 373
References 374
Chapter 14: 
380 
14.1 Background and Context 381
14.2 Research Issues and Opportunities 384
14.2.1 Evaluating Design Alternatives 384
14.2.2 Disaggregated Process Evaluation and Improvement: Opening the ``Input/Output Transformation Box´´ 385
14.2.3 Hierarchical Manufacturing System Performance 386
14.2.4 Data Measurement Imprecision in Production Systems 389
14.2.5 Dynamical Production Systems 391
14.2.6 Visualization of the DEA Results: Influential Data Identification 395
14.3 A DEA-Based Approach Used for the Design of an Integrated Performance Measurement System 396
14.4 Selected DEA Engineering Applications 399
14.4.1 Four Applications 399
14.4.1.1 Evaluating Efficiency of Turbofan Jet Engines (Bulla et al. 2000) 399
14.4.1.2 Measurement and Monitoring of Relative Efficiency of Highway Maintenance Patrols (Cook et al. 1990, 1994) and of Highway Maintenance Operations (Ozbek et al. 2010a, b)6 
400 
14.4.1.3 Data Envelopment Analysis of Space and Terrestrially Based Large Scale Commercial Power Systems for Earth (Criswell and Thompson 1996) 
401 
14.4.1.4 The Relationship of DEA and Control Charts (Hoopes and Triantis 2001) 401
14.4.2 The Effect of Environmental Controls on Productive Efficiency 402
14.4.3 The Performance of Transit Systems 403
14.4.4 Other Engineering Applications of DEA 404
14.5 Systems Thinking Concepts and Future DEA Research in Engineering 405
14.5.1 The Need for Operational Thinking 405
14.5.2 Contribution to Performance Measurement Science 406
14.5.3 Relationship of the DEA Model with the Real World 407
References 408
Chapter 15: 
420 
15.1 Introduction 420
15.2 Universities1 
423 
15.2.1 Introduction 423
15.2.2 Conceptual Framework 424
15.2.3 Research Design 426
15.2.4 Results and Analysis 428
15.2.5 Concluding Remarks 429
15.3 Hotels2 
430 
15.3.1 Introduction 430
15.3.2 Conceptual Framework 431
15.3.3 A Quick Guide to Selecting Inputs and Outputs 433
15.3.4 A Numerical Example 433
15.3.5 Concluding Remarks 435
15.4 Real Estate Agents3 
436 
15.4.1 Introduction 436
15.4.2 Research Design 437
15.4.3 Analysis of Results 439
15.4.4 Concluding Remarks 441
15.5 Commercial Banks5 
442 
15.5.1 Introduction 442
15.5.2 Conceptual Framework 443
15.5.2.1 Modeling Profit Efficiency 443
15.5.2.2 Key Profit Centers and Estimating Corresponding Data 444
15.5.3 Methodology 446
15.5.3.1 Overview of Network DEA 446
15.5.3.2 Network Slacks-Based Measure of Efficiency 449
15.5.3.3 Data and Simulation 450
15.5.4 Results and Analysis 451
15.5.4.1 Profit Efficiency Using Traditional DEA (SBM) 451
15.5.4.2 Profit Efficiency Using Network SBM and Simulated Profit Center Data 452
15.5.5 Concluding Remarks 454
15.6 Epilogue 455
References 457
Chapter 16: 
461 
16.1 Introduction 461
16.2 Brief Background and History 463
16.2.1 Acute General Hospitals and Academic Medical Centers 464
16.2.2 Nursing Homes 465
16.2.3 Department Level, Team-Level, and General Health-Care Studies 466
16.2.4 Physician-Level Studies 467
16.2.5 Data Envelopment Analysis Versus Stochastic Frontier Analysis 468
16.2.6 Reviewer Comments on the Usefulness of DEA 469
16.2.7 Summary 470
16.3 Health-Care Models 471
16.3.1 Clinical Efficiency Definitions 471
16.3.2 How to Model Health-Care Providers: Hospitals, Nursing Homes, Physicians 472
16.3.3 Managerial and Clinical Efficiency Models 473
16.3.3.1 Medical Center and Acute Hospital Models: Examples of Managerial Efficiency 475
16.3.3.2 Nursing Homes: Another Example of Managerial Efficiency 476
16.3.3.3 Primary Care Physician Models: An Example of Clinical Efficiency 477
16.3.4 Hospital Physician Models: Another Example of Clinical Efficiency 478
16.3.5 Profitability Models: A Nursing Home Example 479
16.4 Special Issues for Health Applications 481
16.4.1 Defining Models from Stakeholder Views 481
16.4.2 Selecting Appropriate Health-Care Outputs and Inputs: The Greatest Challenge for DEA 483
16.4.2.1 Take Two Aspirin and Call Me in the Morning 483
16.4.2.2 Using DEA to Adjust Outputs for Patient Characteristics and Case mix 484
16.4.3 Should Environmental and Organizational Factors Be Used as Inputs? 485
16.4.4 Problems on the Best Practice Frontier: A Physician Example 485
16.4.4.1 The Concept of a Preferred Practice Cone or Quality Assurance Region 487
16.4.4.2 Constant Versus Variable Returns to Scale 488
16.4.4.3 Scale and Scope Issues 489
16.4.5 Analyzing DEA Scores with Censored Regression Models 489
16.5 New Directions: From Productive Efficiency Frontiers to Quality-Outcome Frontiers 492
16.5.1 A Field Test: Combining Outcome Frontiers and Efficiency Frontiers 495
16.6 A Health DEA Application Procedure: Eight Steps 497
16.6.1 Step 1: Identification of Interesting Health-Care Problem and Research Objectives 497
16.6.2 Step 2: Conceptual Model of the Medical Care Production Process 498
16.6.3 Step 3: Conceptual Map of Factors Influencing Care Production 498
16.6.4 Step 4: Selection of Factors 498
16.6.5 Step 5: Analyze Factors Using Statistical Methods 499
16.6.6 Step 6: Run Several DEA Models 499
16.6.7 Step 7: Analyze DEA Scores with Statistical Methods 499
16.6.8 Step 8: Share Results with Practitioners and Write It Up 500
16.7 DEA Health Applications: Do´s and Don´ts 500
16.7.1 Almost Never Include Physicians As a Labor Input 500
16.7.2 Use Caution When Modeling Intermediate and Final Hospital Outputs 501
16.7.3 Do Check the Distribution of DEA Scores and Influence of Best Practice Providers on Reference Sets 503
16.8 A Final Word 504
References 506
Index 510