Accounting and Causal Effects (eBook)

Douglas A. Schroeder has been at The Ohio State University since 1986; previously he was serving as a member of the Purdue University faculty. He has published papers in and served as a reviewer for numerous academic journals. His current scholarly interests involve analysis of accounting choice.

Preface 6
Contents 8
List of Tables 15
List of Figures 23
Introduction 24
1.1 Problematic illustration 25
1.2 Jaynes’ desiderata for scientific reasoning 27
1.3 Overview 30
1.4 Additional reading 31
Accounting choice 32
2.1 Equilibrium earnings management 33
2.2 Asset revaluation regulation 35
2.3 Regulated report precision 37
2.4 Inferring transactions from financial statements 40
2.5 Additional reading 41
Linear models 42
3.1 Standard linear model ( 42
3.2 Generalized least squares ( 44
3.3 Tests of restrictions and 45
(Frisch-Waugh-Lovell) 45
3.4 Fixed and random effects 49
3.5 Random coefficients 54
3.6 Ubiquity of the Gaussian distribution 56
3.7 Interval estimation 59
3.8 Asymptotic tests of restrictions: Wald, 61
statistics 61
3.9 Misspecification and 64
estimation 64
3.10 Proxy variables 66
3.11 Equilibrium earnings management 71
3.12 Additional reading 77
3.13 Appendix 78
Loss functions and estimation 81
4.1 Loss functions 81
4.2 Nonlinear Regression 84
4.3 Maximum likelihood estimation 87
4.4 James-Stein shrinkage estimators 92
4.5 Summary 97
4.6 Additional reading 98
Discrete choice models 99
5.1 Latent utility index models 99
5.2 Linear probability models 100
5.3 Logit (logistic regression) models 100
5.4 Probit models 108
5.5 Robust choice models 114
5.6 Tobit (censored regression) models 116
5.7 Bayesian data augmentation 116
5.8 Additional reading 117
Nonparametric regression 118
6.1 Nonparametric (kernel) regression 118
6.2 Semiparametric regression models 120
6.3 Specification testing against a general nonparametric benchmark 122
6.4 Locally linear regression 124
6.5 Generalized cross-validation ( 125
6.6 Additional reading 126
Repeated-sampling inference 127
7.1 Monte Carlo simulation 128
7.2 Bootstrap 128
7.3 Bayesian simulation 131
7.4 Additional reading 142
Overview of endogeneity 143
8.1 Overview 144
8.2 Selectivity and treatment effects 167
8.3 Why bother with endogeneity? 168
8.4 Discussion and concluding remarks 175
8.5 Additional reading 175
Treatment effects: ignorability 177
9.1 A prototypical selection setting 177
9.2 Exogenous dummy variable regression 178
9.3 Tuebingen-style examples 179
9.4 Nonparametric identification 184
9.5 Propensity score approaches 189
9.6 Propensity score matching 192
9.7 Asset revaluation regulation example 195
9.8 Control function approaches 223
9.9 Summary 224
9.10 Additional reading 224
Treatment effects: 226
10.1 Setup 226
10.2 Treatment effects 227
10.3 Generalized Roy model 229
10.4 Homogeneous response 230
10.5 Heterogeneous response and treatment effects 231
10.6 Continuous treatment 255
10.7 Regulated report precision 258
10.8 Summary 292
10.9 Additional reading 292
Marginal treatment effects 293
11.1 Policy evaluation and policy invariance conditions 293
11.2 Setup 295
11.3 Generalized Roy model 295
11.4 Identification 296
connections to other treatment effects 298
11.5 298
11.6 Comparison of identification strategies 304
11.7 304
estimation 304
11.8 Discrete outcomes 306
11.9 Distributions of treatment effects 309
11.10 Dynamic timing of treatment 310
11.11 General equilibrium effects 311
11.12 Regulated report precision example 311
11.13 Additional reading 318
Bayesian treatment effects 319
12.1 Setup 320
12.2 Bounds and learning 320
12.3 Gibbs sampler 321
12.4 Predictive distributions 323
12.5 Hierarchical multivariate Student t variation 324
12.6 Mixture of normals variation 324
12.7 A prototypical Bayesian selection example 325
12.8 Regulated report precision example 329
12.9 Probability as logic and the selection problem 348
12.10 Additional reading 349
Informed priors 350
13.1 Maximum entropy 351
13.2 Complete ignorance 353
13.3 A little background knowledge 354
13.4 Generalization of maximum entropy principle 354
13.5 Discrete choice model as maximum entropy prior 357
13.6 Continuous priors 359
13.7 Variance bound and maximum entropy 368
13.8 An illustration: Jaynes’ widget problem 372
13.9 Football game puzzle 387
13.10 Financial statement example 388
13.11 Smooth accruals 393
13.12 Earnings management 399
distribution 415
13.13 Jaynes’ 415
13.14 Concluding remarks 418
13.15 Additional reading 418
13.16 Appendix 418
Asymptotic theory 430
A.1 Convergence in probability (laws of large numbers) 430
A.2 Convergence in distribution (central limit theorems) 434
A.3 Rates of convergence 439
A.4 Additional reading 440
Bibliography 441
Index 460