Abstract
In this paper we study the identification of nonparametric models, that is, models in which both the functional forms of the equations and the probability distributions of the disturbances remain unspecified. Identifiability in such models does not mean uniqueness of structural parameters but rather uniqueness of policy-related predictions that such parameters would normally support.
We provide sufficient and necessary conditions for identifying predictions of the type “Find the distribution of Y, assuming thatXis controlled by external intervention,” where Y andXare arbitrary variables of interest. Whenever identifiable, such predictions can be expressed in closed algebraic form, in terms of observed distributions. We also show how the identifying criteria can be tested qualitatively using the graphical representation of the structural model, thus simplifying and generalizing the standard identifiability tests of linear models (e.g., rank and order). Finally, we provide meaningful and precise definitions of effect decomposition for both parametric and nonparametric models.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Alwin, D.F., and Hauser, R.M., “The decomposition of effects in path analysis,”American Sociological Review40, 37–47, 1975.
Angrist, J.D., Imbens, G.W., and Rubin, D.B., “Identification of causal effects using instrumental variables,” Department of Economics, Harvard University, Cambridge, MA, Technical Report No. 136, June 1993. To appear inJASA.
Balke, A., and Pearl, J., “Nonparametric bounds on causal effects from partial compliance data,” Department of Computer Science, University of California, Los Angeles, Technical Report R-199, 1993. Submitted to theJournal of the American Statistical Association.
Balke, A., and Pearl, J., “Counterfactual Probabilities: Computational Methods, Bounds, and Applications,” In R. Lopez de Mantaras and D. Poole (Eds.)Uncertainty in Artificial Intelligence - 10Morgan Kaufmann, San Mateo, 46–54, 1994.
Balke, A., and Pearl, J., “Counterfactuals and policy analysis in structural models,” In P. Besnard and S. Hanks (Eds.)Uncertainty in Artificial Intelligence 11Morgan Kaufmann, San Francisco, 11–18, 1995.
Bollen, K.A.Structural Equations with Latent VariablesJohn Wiley and Sons, New York, 1989.
Bowden, R.J., and Turkington, D.A.Instrumental VariablesCambridge University Press, Cambridge, 1984.
Cox, D.R.Planning of ExperimentsJohn Wiley and Sons, New York, 1958.
Dean, T., and Kanazawa, K., “A Model for Reasoning about Persistence and Causation,”Computational Intelligence5, 142–150, 1989.
Fisher, F.M., “A correspondence principle for simultaneous equation models,”Econometrica38, 73–92, 1970.
] Freedman, D., “As others see us: A case study in path analysis” (with discussion)Journal of Educational Statistics12, 101–223, 1987.
] Frisch, R., “Statistical versus theoretical relations in economic macrodynamics,” League of Nations Memorandum, 1938. (Reproduced, with Tinbergen’s comments, inAutonomy of Economic RelationsOslo: Universitetets Socialokonomiske Institutt, 1948). Cited in M.S. MorganThe History of Econometric IdeasCambridge University Press, Cambridge, 1990.
Galles, D., and Pearl, J., “Axioms of causal relevance,” Department of Computer Science, University of California, Los Angeles, Technical Report R-240, January 1996. Submitted to AAAI-96.
Galles, D., and Pearl, J., “Testing identifiability of causal effects,” In P. Besnard and S. Hanks (Eds.)Uncertainty in Artificial Intelligence 11Morgan Kaufmann, San Francisco, CA, 185–195, 1995.
Geiger, D., “Graphoids: A qualitative framework for probabilistic inference,” UCLA, Ph.D. Dissertation, Computer Science Department, 1990.
Geiger, D., and Pearl, J., “On the logic of causal models,”Proceedings of the 4th Workshop on Uncertainty in Artificial IntelligenceSt. Paul, MN, 1988, pp. 136–147. Also in L. Kanal et al. (Eds.)Uncertainty in Artificial Intelligence4, Elsevier Science Publishers, North-Holland, Amsterdam, pp. 3–14, 1990.
Geiger, D., Verma, T.S., and Pearl, J., “Identifying independence in Bayesian networks.”Networksvolume 20, pages 507–534. John Wiley and Sons, Sussex, England, 1990.
Goldberger, A.S., Structural Equation Models in the Social Sciences, Seminar Press, New York, 1973.
Haavelmo, T., “The statistical implications of a system of simultaneous equations,”Econometrica11, 1–12, 1943.
Lauritzen, S.L., and Spiegelhalter, D.J., “Local computations with probabilities on graphical structures and their applications to expert systems,”Proceedings of the Royal Statistical Society Series B50, 154–227, 1988.
Marschak, J., “Economic measurements for policy and prediction,” in T. Koopmans and W. Hood (Eds.)Studies in Econometric Method Cowles Commission Monograph 14Chapter 1, Yale University Press, New Haven, 1953.
Manski, C.F., “Nonparametric bounds on treatment effects,”American Economic Review Papers and Proceedings80, 319–323, 1990.
Pearl, J.Probabilistic Reasoning in Intelligence SystemsMorgan Kaufmann, San Mateo, 1988.
Pearl, J., “Graphical models, causality and intervention,”Statistical Science8(3) 266–269, 1993.
Pearl, J., “A probabilistic calculus of actions,” in R. Lopez de Mantaras and D. Poole (Eds.)Uncertainty in Artificial Intelligence10, Morgan Kaufmann, San Mateo, pp. 454–462, 1994.
Pearl, J., “Causal diagrams for experimental research,”Biometrika82(4) 669–710, 1995.
Pearl, J., “Causal inference from indirect experiments,”Artificial Intelligence in Medicine Journal7, 561–582, 1995.
Pearl, J., “On the testability of causal models with latent and instrumental variables,” In P. Besnard and S. Hanks (Eds.)Uncertainty in Artificial Intelligence 11Morgan Kaufmann Publishers, San Francisco, 435–443, 1995.
Pearl, J., and Robins, J., “Probabilistic evaluation of sequential plans from causal models with hidden variables,” In P. Besnard and S. Hanks (Eds.)Uncertainty in Artificial Intelligence 11Morgan Kaufmann, San Francisco, 444–453, 1995.
Pearl, J., and Verma, T., “A theory of inferred causation,” in J.A. Allen, R. Fikes, and E. Sandewall (Eds.)Principles of Knowledge Representation and Reasoning: Proceeding of the Second International ConferenceMorgan Kaufmann, San Mateo, 1991, pp. 441–452.
] Pearl, J., Geiger, D., and Verma, T., “The logic of influence diagrams, in R.M. Oliver and J.Q. Smith (Eds.)Influence Diagrams Belief Nets and Decision AnalysisJohn Wiley and Sons, New York, 1990, pp. 67–87.
Pratt, J., and Schlaifer, R., “On the interpretation and observation of laws,”Journal of Economics39, 23–52, 1988.
Robins, J.M., “A new approach to causal inference in mortality studies with a sustained exposure period - applications to control of the healthy workers survivor effect,”Mathematical ModelingVol. 7, 1393–1512, 1986.
Robins, J.M., “The analysis of randomized and non-randomized AIDS treatment trials using a new approach to causal inference in longitudinal studies,” in L. Sechrest, H. Freeman, and A. Mulley (Eds.)Health Service Research Methodology: A Focus on AIDSNCHSR, U.S. Public Health Service, 1989, pp. 113–159.
Robins, J., “Estimation of the time-dependent accelerated failure time model in the presence of confounding factors,”Biometrika79(2), 321–334, 1992.
Robins,.J.M., and Greenland, S., “Identifiability and exchangeability for direct and indirect effects,”Epidemiology3(2), 143–155, 1992.
Robins, J.M., Blevins, D., Ritter, G., and Wulfsohn, M.“G-Estimation of the Effect of Prophylaxis Therapy for Pneumocystis carinii Pneumonia on the Survival of AIDS Patients,”EpidemiologyVol. 3, Number 4, 319–336, 1992.
Rosenbaum, P.R., “The consequences of adjustment for a concomitant variable that has been affected by the treatment,”Journal of the Royal Statistical Society, Series A (General)147(5), 656–666, 1984.
Rosenbaum, P.R., “The role of known effects in observational studies,”Biometrics45, 557–569, 1989.
Rosenbaum, P., and Rubin, D., “The central role of propensity score in observational studies for causal effects,”Biometrica70, 41–55, 1983.
Rubin, D.B., “Formal modes of statistical inference for causal effects,”Journal of Statistical Planning and Inference25, 279–292, 1990.
Shafer, G.The Art of Causal Conjecture. MITPress, Cambridge, 1996. Forthcoming.
Simon, H.A., Models of Discovery: and Other Topics in the Methods of Science, D. Reidel, Dordrecht, Holland, 1977.
Spirtes, P., “Conditional independence in directed cyclic graphical models for feedback,” Department of Philosophy, Carnegie-Mellon University, Pittsburg, PA, Technical Report CMU-PHIL-53May1994.
Spirtes, P., Glymour, C., and Schienes, R.Causation Prediction and SearchSpringer-Verlag, New York, 1993.
Sobel, M.E., “Effect analysis and causation in linear structural equation models,”Psychometrika55(3), 495–515, 1990.
Spiegelhalter, D.J., Lauritzen, S.L., Dawid, P.A., and Cowell, R.G., “Bayesian analysis in expert systems,”Statistical Science8(3), 219–247, 1993.
Strotz, R.H., and Wold, H.O.A., “Recursive versus nonrecursive systems: An attempt at synthesis,”Econometrica28, 417–427, 1960.
Verma, T.S., “Causal networks: Semantics and expressiveness,” in R. Shachter et al. (Eds.)Uncertainty in Artificial Intelligence 4Elsevier Science Publishers, North-Holland, Amsterdam, 1990, 69–76.
Verma, T.S. and Pearl, J., “Equivalence and synthesis of causal models.” InUncertainty in Artificial IntelligenceElsevier Science Publishers, Cambridge, pages 6, 220–227, 1990.
Wermuth, N., “On block-recursive regression equations” (with discussion)Brazilian Journal of Probability and Statistics6, 1–56, 1992.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1997 Springer-Verlag New York, Inc.
About this paper
Cite this paper
Pearl, J. (1997). On the Identification of Nonparametric Structural Models. In: Berkane, M. (eds) Latent Variable Modeling and Applications to Causality. Lecture Notes in Statistics, vol 120. Springer, New York, NY. https://doi.org/10.1007/978-1-4612-1842-5_3
Download citation
DOI: https://doi.org/10.1007/978-1-4612-1842-5_3
Publisher Name: Springer, New York, NY
Print ISBN: 978-0-387-94917-8
Online ISBN: 978-1-4612-1842-5
eBook Packages: Springer Book Archive