Abstract
This article considers the question of how to cope with heterogeneity when studying causal effects. The standard approach in empirical economics is still to use a linear model and interpret the coefficients as the average returns or effects. Nowadays, instrumental variables (IV) are quite popular to account for (unobserved) heterogeneity when estimating these parameters. First the inadequacy of these standard methods is illustrated. Then it is shown why varying-coefficient models have a strong natural potential to model heterogeneity in many interesting regression problems. Moreover, it is straight forward to develop alternative IV specifications in the varying-coefficient models framework. The corresponding modeling and implementation facilities that are nowadays available in R are studied.
Similar content being viewed by others
Notes
A covariate \(X\) is said to be endogenous if the error terms is not mean independent of it. This typically leads to identification problems, biased estimators and invalids causal inference. Else it is called exogenous.
We put bias in quotation marks as the OLS estimate is actually not biased in the statistical sense, it does simply estimate something different than the average slope. Alternatively you may say that not the estimator is wrong but the interpretation that people often attach to it.
The preposition real indicates that the numbers \(E^R\) assigned to the different households—in this example from all-over Canada—refer to the same purchase power.
Simplified, this means that if a household switches from a product to a substitute because of a special discount, he will switch back if this discount is no longer conceded.
The published consumer price index refers to a representative mean basket of goods and services. This basket clearly varies a lot over the different income groups, resulting in different effectively realized inflation rates for each group.
References
Allais M (1953) Le comportement de l’homme rationnel devant le risque: Critique des postulats et axiomes de l’ecole americaine. Econometrica 21(4):503–546
Belitz C, Brezger A, Kneib T, Lang S, Umlauf N (2013) BayesX: Software for Bayesian Inference in Structured Additive Regression Models. Version 2.1
Branca M (2013) An analysis of implementations of varying-coefficient models. Master Thesis, University of Geneva, Switzerland
Cai Z, Fan J, Li R (2000) Efficient estimation and inferences for varying-coeffcient models. J Am Stat Assoc 95(451):888–902
Card D (1995) Using geographic variation in college proximity to estimate the return to schooling. In: Christofides LN, Vanderkamp J, Grant EK, Swidinsky R (eds) Aspects of labour market behavior: essays in honour of John Vanderkamp. University of Toronto Press, Toronto, pp 201–222
Chiang C-T, Rice JA, Wu CO (2001) Smoothing spline estimation for varying coeffcient models with repeatedly measured dependent variables. J Am Stat Assoc 96(454):605–619
Crossley TF, Pendakur K (2010) The commonscaling social cost-of-living index. J Bus Econ Stat 28(4):523–538
Deaton A, Muellbauer J (1980) Economics and consumer behavior. Cambridge University Press, Cambridge
Diewert WE (1998) Index number issues in the consumer price index. J Econ Perspect 12(1):47–58
Fahrmeir L, Kneib T, Lang S (2004) Penalized structured additive regression for space-time data: a bayesian perspective. Stat Sin 14:715–745
Fan J, Zhang W (1999) Statistical estimation in varying coefficient models. Ann Stat 27(5):1491–1518
Fan J, Zhang W (2000) Simultaneous confidence bands and hypothesis testing in varying-coefficient models. Scand J Stat 27(4):715–731
Fan J, Zhang W (2008) Statistical methods with varying coefficient models. Stat Interface 1(1):179–195
González Manteiga W, Lombardía MJ, Martínez Miranda MD, Sperlich S (2013) Kernel smoothers and bootstrapping for semiparametric mixed effects models. J Multivar Anal 114:288–302
Hastie T, Tibshirani R (1990) Generalized additive models. Chapman and Hall, London
Hastie T, Tibshirani R (1993) Varying-coeffcient models. J R Stat Soc Ser B (Methodol) 55(4):757–796
Hayfield T, Racine JS (2008) Nonparametric econometrics: the np package. J Stat Softw 27(5):1–32
Imbens GW, Angrist JD (1994) Identification and estimation of local average treatment effects. Econometrica 62(2):467–475
Kneib T, Fahrmeir L (2006) Structured additive regression for multicategorical spacet-ime data: a mixed model approach. Biometrics 62:109–118
Mammen E, Nielsen JP (1999) The existence and asymptotic properties of a backfitting projection algorithm under weak conditions. Ann Stat 27:1443–1490
Mammen E, Nielsen JP (2003) Generalised structured models. Biometrika 90(3):551–566
Newey WK, Powell JL, Vella F (1999) Nonparametric estimation of triangular simultaneous equations models. Econometrica 67(3):565–603
Park BU, Mammen E, Lee YK, Lee ER (2015) Varying coefficient regression models: a review and new developments. Int Stat Rev 83(1):36–64
Pendakur K, Sperlich SA (2010) Semiparametric estimation of consumer demand systems in real expenditure with partially linear price effects. J Appl Econom 25:420–457
Pendakur K, Scholz M, Sperlich SA (2010) Semiparametric indirect utility and consumer demand. Comput Stat Data Anal 54:2763–2775
Profit S, Sperlich S (2004) Non-uniformity of jobmatching in a transition economy—a nonparametric analysis for the Czech Republic. Appl Econ 6(7):695–714
R Development Core Team (2014) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/
Rigby RA, Stasinopoulos DM (2005) Generalized additive models for location, scale and shape (with discussion). Appl Stat 54:507–554
Roca-Pardinas J, Sperlich SA (2010) Feasible estimation in generalized structured models. Stat Comput 20:367–379
Roy R (1947) La Distribution du Revenu Entre Les Divers Biens. Econometrica 15:205–225
Severance-Lossin E, Sperlich S (1999) Estimation of derivatives for additive separable models. Statistics 33(3):241–265
Sperlich S (2009) A note on nonparametric estimation with predicted variables. Econom J 12:382–395
Sperlich S, Tjøstheim D, Yang L (2002) Nonparametric estimation and testing of interaction in additive models. Econom Theory 18:197–251
Telser L (1964) Iterative estimation of a set of linear regression equations. J Am Stat Assoc 59:845–862
Wood SN (2011) Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J R Stat Soc (B) 73(1):3–36
Wood SN (2006) Generalized additive models: an introduction with R. Chapman and Hall/CRC, London
Wood SN (2004) Stable and efficient multiple smoothing parameter estimation for generalized additive models. J Am Stat Assoc 99(467):673–686
Wood SN (2003) Thin-plate regression splines. J R Stat Soc (B) 65(1):95–114
Wood SN (2000) Modelling and smoothing parameter estimation with multiple quadratic penalties. J R Stat Soc (B) 62(2):413–428
Yang L, Park BU, Xue L, Härdle W (2006) Estimation and testing for varying coefficients in additive models with marginal integration. J Am Stat Assoc 101:1212–1227
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Sperlich, S., Theler, R. Modeling heterogeneity: a praise for varying-coefficient models in causal analysis. Comput Stat 30, 693–718 (2015). https://doi.org/10.1007/s00180-015-0581-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-015-0581-y