Abstract
We investigate the computational aspect of estimating discrete-choice games under incomplete information. In these games, multiple equilibria can exist. Also, different values of structural parameters can result in different numbers of equilibria. Consequently, under maximum-likelihood estimation, the likelihood function is a discontinuous function of the structural parameters. We reformulate the maximum-likelihood estimation problem as a constrained optimization problem in the joint space of structural parameters and economic endogenous variables. Under this formulation, the objective function and structural equations are smooth functions. The constrained optimization approach does not require repeatedly solving the game or finding all the equilibria. We use two static-game models to demonstrate this approach, conducting Monte Carlo experiments to evaluate the finite-sample performance of the maximum-likelihood estimator, two-step estimators, and the nested pseudo-likelihood estimator.
Similar content being viewed by others
Notes
An equilibrium p is stable under best-response iteration if ρ[∇ p Ψ(p, x;𝜃)], the spectral radius of the Jacobian mapping ∇ p Ψ(p, x;𝜃), is less than 1. It is easy to check that at \(\boldsymbol {p}_{2}^{*}\), \(\rho [\nabla _{\boldsymbol {p}} \boldsymbol {\Psi }(\boldsymbol {p^{*}_{2}}, \boldsymbol {x}; \boldsymbol {\theta }^{0})] = 1.148\), while at \(\boldsymbol {p}_{1}^{*}\) and \(\boldsymbol {p}_{3}^{*}\), \(\rho [\nabla _{\boldsymbol {p}} \boldsymbol {\Psi }(\boldsymbol {p^{*}_{1}}, \boldsymbol {x}; \boldsymbol {\theta }^{0})] = 0.4099\) and \(\rho [\nabla _{\boldsymbol {p}} \boldsymbol {\Psi }(\boldsymbol {p^{*}_{3}}, \boldsymbol {x}; \boldsymbol {\theta }^{0})] = 0.8398\), respectively.
The derivation of the constrained optimization formulation under a general equilibrium selection mechanism is given in Appendix A.
Pesendorfer and Schmidt-Dengler (2008) proposed an asymptotic least-squares estimator and derived the asymptotically optimal weight matrix. Here, we do not derive the optimal weight matrix and use the identity matrix as the weight matrix in this example.
As noted by Pakes et al. (2007), the pseudo likelihood function is an inappropriate criterion function to use in the second step. The reason that we still use the 2S-PML estimator in our Monte Carlo experiments is to provide a direct comparison on the performance of the 2S-PML to that of the ML estimator and the NPL estimator.
One can improve the performance of the 2S-LS estimator by using the optimal weighting matrix, which has been suggested by Pesendorfer and Schmidt-Dengler (2008).
In our experiment, the NPL algorithm usually converges to the same parameter estimates from starting values \(\widehat {\boldsymbol {P}}_{\text {freq}}\), \(\frac {3\widehat {\boldsymbol {P}}_{\text {freq}}}{4}\), and \( \frac {\widehat {\boldsymbol {P}}_{\text {freq}}}{2}\).
These are Row 117 to 316 in Jia’s data file, XMat97.out.
In the whole date sets with 2065 counties, 33.6 % (694 counties) had Walmart stores only, 5 % (103 counties) had Kmart stores only, and 13.9 % (287 counties) had both Walmart and Kmart stores.
We provide more detailed discussion on ρ[∇ P Ψ(P, X;𝜃 0)] and the condition on δ 0 in Appendix B.
Results for T = 10 and T = 25 provide similar observations to those for T = 5 and T = 50, and are omitted here.
References
Arcidiacono, P., & Miller, R.A. (2011). CCP estimation of dynamic discrete choice models with unobserved heterogeneity. Econometrica, 79, 1823–1868.
Aguirregabiria, V., & Mira, P. (2007). Sequential estimation of dynamic discrete games. Econometrica, 75, 1–53.
Aitchison, J., & Silvey, S.D. (1958). Maximum likelihood estimation of parameters subject to restraints. Annals of Mathematical Statistics, 29, 813–828.
Bajari, P., Benkard, C.L., Levin, J. (2007). Estimating dynamic games of imperfect competition. Econometrica, 75, 1331–1370.
Bajari, P., Hong, H., Krainer, J., Nekipelov, D. (2010). Estimating static models of strategic interactions. Journal of Business and Economic Statistics, 28, 469–482.
Berry, S. (1992). Estimation of a model of entry in the airline industry. Econometrica, 60, 889–917.
Ciliberto, F., & Tamer, E. (2009). Market structure and multiple equilibria in airline markets. Econometrica, 77, 1791–1828.
de Paula, A. (2013). Econometric analysis of games with multiple equilibria. Annual Review of Econonomics, 5, 107–131.
Ellickson, P., & Misra, S. (2008). Supermarket pricing strategies. Marketing Science, 27, 811–828.
Ellickson, P., & Misra, S. (2011). Estimating discrete games. Marketing Science, 30, 997–1010.
Fudenberg, D., & Tirole, J. (1991). In Game theory. Cambridge: MIT Press.
Gourieroux, C., & Monfort, A. (1995). In Statistics and econometric models Vol. 1. Cambridge: Cambridge University Press.
Hotz, J., & Miller, R.A. (1993). Conditional choice probabilities and the estimation of dynamic models. Review of Economic Studies, 60, 497–529.
Jia, P. (2008). What happens when Wal-Mart comes to town: an empirical analysis of the discount retail industry. Econometrica, 76, 1263–1316.
Judd, K., Renner, P., Schmedders, K. (2012). Finding all pure strategy equilibria in dynamic and static games with continuous strategies. Quantitative Economics, 3, 289–331.
Kohlberg, E., & Mertens, J.-F. (1986). On the strategic stability of equilibria. Econometrica, 54, 1003–1038.
Mertens, J.-F. (1989). Stable equilibria–a reformulation: Part I. Definition and basic properties. Mathematics of Operations Research, 14, 575–625.
Mertens, J.-F. (1991). Stable equilibria–a reformulation: Part II. Discussion of the definition, and further results. Mathematics of Operations Research, 16, 694–753.
Misra, S. (2013). Markov chain Monte Carlo for incomplete information discrete games. Quantitative Marketing and Economics, 11, 117–153.
Orhun, A. Y. (2013). Spatial differentiation in the supermarket industry: the role of common information. Quantitative Marketing and Economics, 11, 3–37.
Pakes, A., Ostrovsky, M., Berry, S. (2007). Simple estimators for the parameters of discrete dynamic games, with entry/exit examples. RAND Journal of Economics, 38, 373–399.
Pakes, A., Porter, J., Ho, K., Ishii, J. (2011). In Moment Inequalities and Their Applications. Working paper: Harvard University.
Pesendorfer, M., & Schmidt-Dengler, P. (2008). Asymptotic least squares estimators for dynamic games. Review of Economic Studies, 75, 901–928.
Pesendorfer, M., & Schmidt-Dengler, P. (2010). Sequential estimation of dynamic discrete games: a comment. Econometrica, 78, 833–842.
Rust, J. (1987). Optimal replacement of GMC bus engines: an empirical model of Harold Zurcher. Econometrica, 55, 999–1033.
Rust, J. (2008). Private communication.
Seim, K. (2006). An empirical model of firm entry with endogenous product-type choices. RAND Journal of Economics, 37, 619–640.
Su, C.-L., & Judd, K.L. (2012). Constrained optimization approaches to estimation of structural models. Econometrica, 80, 2213–2230.
Silvey, S.D. (1975). Statistical inference. London: Chapman & Hall.
Sweeting, A. (2009). The strategic timing of radio commercials: an empirical analysis using multiple equilibria. RAND Journal of Economics, 40, 710–742.
Vitorino, M.A. (2012). Empirical entry games with complementarities: an application to the shopping center industry. Journal of Marketing Research, 49, 175–191.
Zhu, T., & Singh, V. (2009). Spatial competition with endogenous location choices: an application to discount retailing. Quantitative Marketing and Economics, 7, 1–35.
Acknowledgments
I thank the editor Greg M. Allenby and two anonymous referees for insightful comments and helpful suggestions. I have benefited greatly from discussions with Panle Jia Barwick, Kenneth L. Judd, Sanjog Misra, Ariel Pakes, Peter E. Rossi, and John Rust, and comments from seminar participants at the 2010 QME Conference, Boston College, and Harvard University. I am very grateful to John Rust for developing and suggesting the first static-game example used in this paper and to Harry J. Paarsch for carefully reading this paper. I acknowledge the financial support from the IBM Faculty Research Fund at The University of Chicago Booth School of Business. All errors are mine.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix A: The constrained optimization formulation for the ML estimator under general equilibrium selection mechanism
In this appendix, we derive the constrained optimization formulation of the maximum likelihood estimation problem under a general equilibrium selection mechanism. To simplify the notation, we only consider the case of one market and drop the superscript m for market index in the notation used in Section 3.
Given the parameter vector 𝜃 and firms’ observed types x, let \({\mathcal E}(\boldsymbol {\theta }, \boldsymbol {x}) = \{ \bar {\boldsymbol {p}}_{k}(\boldsymbol {\theta }) \}_{k}\) denote the set of BN equilibria, where \(\bar {\boldsymbol {p}}_{k}(\boldsymbol {\theta }) = (\bar {p}_{ka}(\boldsymbol {\theta }), \bar {p}_{kb}(\boldsymbol {\theta })) \) is the k-th equilibrium that solves the BN equilibrium (4). For the constrained optimization approach, given 𝜃, an equilibrium selection mechanism \(\boldsymbol {\lambda } = \{\lambda _{k}\}_{k=1}^{|{\mathcal E}(\boldsymbol {\theta }, \boldsymbol {x})|}\) and any given set of probabilities \(\{\boldsymbol {p}_{k} = (p_{ka}, p_{kb})\}_{k=1}^{|{\mathcal E}(\boldsymbol {\theta }, \boldsymbol {x})|}\) (which do not need to be the equilibrium probabilities that solve the BN equilibrium equation), we define the augmented logarithm of the likelihood function of observing the decisions \(\boldsymbol {y} = ({y_{a}^{t}}, {y_{b}^{t}})_{t=1}^{T}\) as
The constrained optimization formulation of the ML estimator is then defined as:
Notice the constrained optimization problem formulated above is not computable because the number of constraints, \(|{\mathcal E}(\boldsymbol {\theta }, \boldsymbol {x})|\), in Eq. 27 depends on the values of decision variables 𝜃. Nonetheless, we can characterize the equivalence in objective value of the two problems (10) and (27) in the following proposition.
Proposition 2
Let \((\bar {\boldsymbol {\theta }}, \bar {\boldsymbol {\lambda }}(\bar {\boldsymbol {\theta }}))\) be a solution of the ML estimation problem defined in Eq. 10 with corresponding equilibrium probabilities \(\{\bar {\boldsymbol {p}}_{k}(\bar {\boldsymbol {\theta }})\}_{k=1}^{|{\mathcal E}(\bar {\boldsymbol {\theta }}, \boldsymbol {x})|}\). Let \((\boldsymbol {\theta }^{*}, \boldsymbol {\lambda }^{*}, \{\boldsymbol {p}^{*}_{k}\}_{k=1}^{|{\mathcal E}(\theta ^{*}, \boldsymbol {x})|})\) be a solution of the constrained optimization problem (27). Then \(\mathbb {L}[ \bar {\boldsymbol {\theta }}, \bar {\boldsymbol {\lambda }}(\bar {\boldsymbol {\theta }}), \{ \bar {\boldsymbol {p}}_{k}(\bar {\boldsymbol {\theta }}) \}_{k}]=\mathfrak {L} \left [ \boldsymbol {\theta }^{*}, \boldsymbol {\lambda }^{*}, \{\boldsymbol {p}^{*}_{k}\}_{k} \right ] \). If the model is identified, then \(\bar {\boldsymbol {\theta }} = \boldsymbol {\theta }^{*}\).
Proof
By definition, \(\mathbb {L}[ \bar {\boldsymbol {\theta }}, \bar {\boldsymbol {\lambda }}(\bar {\boldsymbol {\theta }}), \{ \bar {\boldsymbol {p}}_{k}(\bar {\boldsymbol {\theta }}) \}_{k}] \geq \mathbb {L}[ \boldsymbol {\theta }, \boldsymbol {\lambda }(\boldsymbol {\theta }), \{ \bar {\boldsymbol {p}}_{k}(\boldsymbol {\theta }) \}_{k}]\) for any given 𝜃, λ(𝜃) and \(\{ \bar {\boldsymbol {p}}_{k}(\boldsymbol {\theta })\}_{k}\) that satisfy the BN equilibrium equation p = Ψ(p, x, 𝜃). Since the pair \((\boldsymbol {\theta }^{*}, \{\boldsymbol {p}^{*}_{k}\}_{k=1}^{|{\mathcal E}(\theta ^{*}, \boldsymbol {x})|})\) satisfies the BN equation as the constraints in Eq. 27, it follows that
Conversely, since the tuple \((\bar {\boldsymbol {\theta }}, \bar {\boldsymbol {\lambda }}(\bar {\boldsymbol {\theta }}), \{\bar {\boldsymbol {p}}_{k}(\bar {\boldsymbol {\theta }})\}_{k=1}^{|{\mathcal E}(\bar {\boldsymbol {\theta }}, \boldsymbol {x})|})\) satisfies the constraints in Eq. 27, we have
If the model is identified, the solution is unique, so \(\bar {\boldsymbol {\theta }} = \boldsymbol {\theta }^{*}\).
Assuming only one equilibrium is played in the data, only one component in λ is 1 and all the others are 0. Then the constrained optimization problem (27) is equivalent to
Notice that the constrained optimization problem formulated in Eq. 28 is also not computable because its size (the number of constraints, \(|{\mathcal E}(\boldsymbol {\theta }, \boldsymbol {x})|\) and decision variables {p k } k ) depend on the values of decision variables 𝜃; it also involves finding all the equilibrium solutions \(\{ {\boldsymbol {p}_{k}} \}_{k =1}^{|{\mathcal E}(\boldsymbol {\theta }, \boldsymbol {x})|}\) at any given 𝜃. We present a computationally tractable reformulation of the problem (28) as follows:
Notice that the size (number of decision variables and constraints) of the problem (29) is fixed and does not depend on the value of 𝜃. Moreover, we do not need to find all equilibria at any given 𝜃; only one equilibrium p is needed since only one equilibrium constraint is imposed. The equivalence in the objective value and solutions between the problems (28) and (29) is stated below. We omit the proof since it follows similar arguments used in proving Proposition 2.
Proposition 3
Let \(\left ( \boldsymbol {\theta }^{*}, \{\boldsymbol {p}^{*}_{k}\}_{k=1}^{|{\mathcal E}(\boldsymbol {\theta }^{*}, \boldsymbol {x})|} \right )\) denote the solution to the optimization problem (28) with \(\boldsymbol {p}^{*}_{\bar {k}} \in \{\boldsymbol {p}^{*}_{k}\}_{k} \) being the equilibrium that maximizes the objective function, i.e.,
Then, the vector \(\left ( \boldsymbol {\theta }^{*}, \boldsymbol {p}^{*}_{\bar {k}} \right )\) is also an optimal solution to the optimization problem (29). Conversely, let \(\left ( \boldsymbol {\hat {\theta }}, \boldsymbol {\hat {p}} \right )\) denote a solution of the problem (29). Then the pair \(\left ( \boldsymbol {\hat {\theta }}, \boldsymbol {\hat {p}} \right )\) together with all the other equilibrium probabilities \(\{\boldsymbol {p}_{k}\}_{k} \setminus \boldsymbol {\hat {p}} \) at \(\boldsymbol {\hat {\theta }}\) is an solution to the problem (28).
Finally, the formulation in Eq. 29 can be generalized to the constrained optimization problem for the ML estimator defined in Eq. 18 for the case of observing data in multiple markets.
Appendix B: Creating examples with multiple equilibria
In this appendix, we derive the formula for calculating the spectral radius ρ[∇ P Ψ(P, X;𝜃)] for the model in Section 5. We first derive the formula for calculating the spectral radius for one market and then generalize the formula to the case with multiple markets. Recall that in Section 5, for m = 1, …, M, we have
We represent the BN equilibrium equations above for market m as
where \(\boldsymbol {p}^{m} = ({p_{W}^{m}}, {p_{K}^{m}})\), Ψ = (Ψ W , Ψ K ), and \(\boldsymbol {X}^{m} = (\boldsymbol {x}^{m}, \boldsymbol {x}_{W}^{m}, \boldsymbol {x}_{K}^{m} )\). The Jacobian of the mapping Ψ(p m, X m;𝜃) with respect to p m is
The eigenvalues λ m of \(\nabla _{\boldsymbol {p}^{m}} \boldsymbol {\Psi }( \boldsymbol {p}^{m}, \boldsymbol {X}^{m}; \boldsymbol {\theta })\) satisfy the following conditions:
where I is the identity matrix. Hence, we have
The spectral radius of the Jacobian mapping \(\nabla _{\boldsymbol {p}^{m}} \boldsymbol {\Psi }( \boldsymbol {p}^{m}, \boldsymbol {X}^{m}; \boldsymbol {\theta })\) for market m is then
Next we generalize the derivation above to the case with multiple markets. Recall we represent the BN equilibrium equations for all markets m = 1, …, M as
where \(\boldsymbol {P} = ({p^{m}_{W}}, {p^{m}_{K}})_{m = 1}^{M}\), Ψ = (Ψ W , Ψ K ) and \(\boldsymbol {X} = (\boldsymbol {x}^{m}, \boldsymbol {x}_{W}^{m}, \boldsymbol {x}_{K}^{m} )_{m=1}^{M}\). Observe that the Jacobian matrix ∇ P Ψ(P, X;𝜃) has the block-diagonal structure
Consequently, the spectral radius of the Jacobian mapping ∇ P Ψ(P, X;𝜃) is given as
To create an example with multiple equilibria in some markets, we need the spectral radius ρ[∇ P Ψ(P ∗, X;𝜃 0)] evaluated at structural parameters 𝜃 0 and the corresponding equilibrium P to be greater than 1, which implies \(| \delta ^{0} | \sqrt {{p_{W}^{m}}(1-{p_{W}^{m}}) {p_{K}^{m}} (1-{p_{K}^{m}})} > 1 \) for some m. This, in turns, requires that
since \(\sqrt {{p_{W}^{m}}(1-{p_{W}^{m}}) {p_{K}^{m}} (1-{p_{K}^{m}})} < 1/4\). As a conservative choice, we choose δ 0 = 6 in the second set of true parameter values in Eq. 25.
Rights and permissions
About this article
Cite this article
Su, CL. Estimating discrete-choice games of incomplete information: Simple static examples. Quant Mark Econ 12, 167–207 (2014). https://doi.org/10.1007/s11129-014-9144-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11129-014-9144-8
Keywords
- Structural estimation
- Discrete-choice games of incomplete information
- Constrained optimization
- Multiple equilibria