Abstract
Two stage least squares is an approach to fit a mathematical or statistical model to data where there are simultaneous or bidirectional cause and effect relationship between the independent variables and the dependent variable. Two stage least squares method is frequently implemented in the case that there is no prior information. On the other hand, in the existence of some prior information, Bayesian estimation methods may be preferable to the classical ones in the simultaneous equations model. From this point of view, ridge regression approach is adapted to Bayesian approach in this paper. In mathematical sense, ridge regression can be derived as a solution of Tikhonov regularization which is a method to overcome ill-posed problems. Therefore, the proposed Bayes estimators resulted from the adaptation of ridge regression mitigate the multicollinearity. Theoretical comparisons among the estimators are performed by means of matrix mean square error criterion and determination of the biasing parameter is also examined in this paper. While doing these comparisons we utilize the matrix theory and Euclidian norm. Theoretical results are tested by a real life data analysis and an extensive Monte Carlo experiment. Eventually, the researchers relevant to mathematics and statistics may prefer the proposed Bayes estimators to the two stage least squares estimator in the existence of multicollinearity.
REFERENCES
T. Kloek and L. B. M. Mennes, ‘‘Simultaneous equations estimation based on principal components of predetermined variables,’’ Econometrica 28, 45–61 (1960).
T. Amemiya, ‘‘On the use of principal components of independent variables in two stage least squares estimation,’’ Int. Econ. Rev. 7, 283–303 (1966).
P. J. Dhrymes, Econometrics (Springer, New York, 1974).
K. R. Kadiyala and J. R. Nunns, ‘‘Estimation of a simultaneous system of equations when the sample is undersized,’’ Unpublished Report (Lafayette, Purdue Univ., Indiana, 1976).
H. D. Vinod and A. Ullah, Recent Advances in Regression Methods (Marcel Dekker, New York, 1981).
S. Toker, S. Kaçıranlar, and H. Güler, ‘‘Two stage Liu estimator in simultaneous equations model,’’ J. Stat. Comput. Simul. 88, 2066–2088 (2018).
N. Özbay and S. Toker, ‘‘Determining the effect of some biasing parameter selection methods for the two stage ridge regression estimator,’’ Sakarya Univ. J. Sci. 22, 1878–1885 (2018).
N. Özbay and S. Toker, ‘‘Multicollinearity in simultaneous equations system: Evaluation of estimation performance of two-parameter estimator,’’ Comput. Appl. Math. 37, 5334–5357 (2018).
S. Toker and N. Özbay, ‘‘Evaluation of two stage modified ridge estimator and its performance,’’ Sakarya Univ. J. Sci. 22, 1631–1637 (2018).
S. Toker, ‘‘Investigating the two parameter analysis of Lipovetsky for simultaneous systems,’’ Stat. Papers 61, 2059–2089 (2020).
S. Toker and N. Özbay, ‘‘The effect of target function on the predictive performance of the two-stage ridge estimator,’’ J. Forecast. 38, 749–772 (2019).
A. Zellner, An Introduction to Bayesian Inference in Econometrics (Wiley, New York, 1971).
A. Zellner, ‘‘Estimation of functions of population means and regression coefficients including structural coefficients: As minimum expected loss approach,’’ J. Econometr. 8, 127–158 (1978).
J. H. Dreze, ‘‘Bayesian limited information analysis of the simultaneous equation model,’’ Econometrica 46, 1045–1075 (1976).
T. Kloek and H. K. van Dijk, ‘‘Bayesian estimates of equation system parameters: An application of integration by Monte Carlo,’’ Econometrica 46, 1–19 (1978).
J. S. Mehta and P. A. V. B. Swamy, ‘‘The existence of moments of some simple Bayes estimators of coefficients in a simultaneous equation model,’’ J. Econometr. 7, 1–13 (1978).
A. Zellner, L. Bauwens, and H. K. van Dijk, ‘‘Bayesian specification analysis and estimation of simultaneous equation models using Monte Carlo methods,’’ J. Econometr. 38, 39–72 (1988).
J. Geweke, ‘‘Using simulation methods for bayesian econometric models: Inference, development and communication,’’ Econometr. Rev. 18, 1–73 (1999).
H. K. van Dijk, ‘‘On Bayesian structural inference in a simultaneous equation model,’’ Econometric Inst. Report EI 2002-10 (Econometr. Inst., 2002).
R. C. Hill and G. Judge, ‘‘Improved prediction in the presence of multicollinearity,’’ J. Econometr. 35, 83–100 (1987).
A. E. Hoerl, ‘‘Application of ridge analysis to regression problems,’’ Chem. Eng. Progress 58 (3), 54–59 (1962).
A. E. Hoerl and R. W. Kennard, ‘‘Ridge regression: Biased estimation for nonorthogonal problems,’’ Technometrics 12, 55–67 (1970).
A. N. Tikhonov, ‘‘On the stability of inverse problems,’’ Dokl. Akad. Nauk SSSR 39, 195–198 (1943).
A. N. Tikhonov, ‘‘Solution of incorrectly formulated problems and the regularization method,’’ Dokl. Akad. Nauk SSSR 151, 501–504 (1963).
E. Maasoumi, ‘‘A ridge-like method for simultaneous estimation of simultaneous equations,’’ J. Econometr. 12, 161–176 (1980).
O. Capps, Jr. and W. D. Grubbs, ‘‘A Monte Carlo study of collinearity in linear simultaneous equation models,’’ J. Stat. Comput. Simul. 39, 139–162 (1991).
S. Toker and S. Kaçıranlar, ‘‘Ridge estimator with correlated errors and two-stage ridge estimator under inequality restrictions,’’ Commun. Stat.-Theory Methods 46, 1407–1421 (2017).
D. E. Giles, ‘‘Bayesian estimation of a possibly mis-specified linear regression model,’’ Econometr. Working Paper EWP1004 (Univ. Victoria, Dep. Econ., 2010).
H. Theil, Principles of Econometrics (Wiley, New York, 1971).
H. M. Wagnar, ‘‘A Monte Carlo study of estimates of simultaneous linear structural equations,’’ Econometrica 26, 117–133 (1958).
J. Kmenta and M. E. Joseph, ‘‘A Monte Carlo study of alternative estimates of the Cobb-Douglas production function,’’ Econometrica 31, 363–390 (1963).
R. A. Summers, ‘‘Capital intensive approach to the small sample properties of various simultaneous equation estimators,’’ Econometrica 33, 1–41 (1965).
R. E. Quandt, ‘‘On certain small sample properties of k-class estimators,’’ Int. Econ. Rev. VI, 92–104 (1965).
J. G. Cragg, ‘‘On the relative small sample properties of several structural equation estimators,’’ Econometrica 35, 89–110 (1967).
J. B. Parker, ‘‘The use of the Monte Carlo methods for solving large-scale problems in neutronics,’’ J. R. Stat. Soc., Ser. A 135, 16–43 (1972).
D. F. Hendry, ‘‘The structure of simultaneous equation estimators,’’ J. Econometr. 4, 51–88 (1976).
S. B. Park, ‘‘Some sampling properties of minimum expected loss (MELO) estimates of structural coefficients,’’ J. Econometr. 18, 295–311 (1982).
J. Johnston, Econometrics Methods, 3rd ed. (McGraw-Hill, New York, 1984).
B. P. Carlin, N. G. Polson, and D. S. Stoffer, ‘‘A Monte Carlo approach to non-normal and non-linear state-space modeling,’’ J. Am. Stat. Assoc. 87, 493–500 (1992).
A. Koutsoyiannis, Theory of Econometrics, 2nd ed. (Macmillan, London, 2001).
D. A. Agunbiade, ‘‘A Monte Carlo approach to the estimation of a just identified simultaneous three-equations model with three multicollinear exogenous variables,’’ Ph.D. Thesis (Univ. Ibadan, Nigeria, 2007, unpublished).
D. A. Agunbiade, ‘‘Effect of multicollinearity and the sensitivity of the estimation methods in simultaneous equation model,’’ J. Mod. Math. Stat. 5 (1), 9–12 (2011).
D. A. Agunbiade and J. O. Iyaniwura, ‘‘Estimation under multicollinearity: A comparative approach using Monte Carlo methods,’’ J. Math. Stat. 6, 183–192 (2010).
B. F. Swindel, ‘‘Good ridge estimators based on prior information,’’ Commun. Stat.-Theory Methods A5, 1065–1075 (1976).
R. W. Farebrother, ‘‘Further results on the mean square error of ridge regression,’’ J. R. Stat. Soc., Ser. B 38, 248–250 (1976).
F. A. Graybill, Matrices with Applications in Statistics (Wadsworth, Belmont, 1983).
F. Akdeniz and S. Kaçıranlar, ‘‘More on the new biased estimator in linear regression,’’ Sankhya: Indian J. Stat., Ser. B 63, 321–325 (2001).
M. R. Özkale and S. Kaçıranlar, ‘‘The restricted and unrestricted two-parameter estimators,’’ Commun. Stat.-Theory Methods 36, 2707–2725 (2007).
H. Güler and S. Kaçıranlar, ‘‘A comparison of mixed and ridge estimators of linear models,’’ Commun. Stat.-Simul. Comput. 38, 368–401 (2009).
S. Kaçıranlar, S. Sakallıoğlu, M. R. Özkale, and H. Güler, ‘‘More on the restricted ridge regression estimation,’’ J. Stat. Comput. Simul. 81, 1433–1448 (2011).
L. R. Klein, Economic Fluctuation in the United States, 1921–1941 (Wiley, New York, 1950).
W. E. Griffiths, R. C. Hill, and G. G. Judge, Learning and Practicing Econometrics (Wiley, New York, 1993).
G. C. McDonald and D. I. Galarneau, ‘‘A Monte Carlo evaluation of some ridge-type estimators,’’ J. Am. Stat. Assoc. 70, 407–416 (1975).
B. M. G. Kibria, ‘‘Performance of some new ridge regression estimators,’’ Commun. Stat.-Simul. Comput. 32, 419–435 (2003).
Funding
This paper is supported by Çukurova University Scientific Research Projects Unit Project no. FBA-2018-9770.
Author information
Authors and Affiliations
Corresponding authors
Additional information
(Submitted by A. I. Volodin)
Selma Toker is a corresponding author
Rights and permissions
About this article
Cite this article
Toker, S., Özbay, N. New Bayesian Approach to the Estimation in Simultaneous Equations Model. Lobachevskii J Math 44, 3872–3888 (2023). https://doi.org/10.1134/S1995080223090421
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1134/S1995080223090421