Abstract
Since the publication of the paper by Anderson and van Wincoop (J Econ Litt 42:691–751, 2004), the estimation of gravity models has turned increasingly structural. The benefit of this development is that empirical models which are based on few parameters explain data on bilateral trade flows relatively well, while being consistent with general equilibrium. The latter permits using the estimated parameters for comparative static analysis. In general, in the cross section such models involve country-pair-specific variables and exporter-specific as well as importer-specific variables. The latter are determined through structural model constraints. The disturbances on this model are typically assumed to be independently if not also identically distributed. This paper illustrates that the assumption of independently distributed disturbances is likely flawed in practice. Ignoring such independence leads to inconsistent test statistics and standard errors of the parameters. We present a structural gravity model which permits the disturbances to be cross-sectionally interdependent in up to three dimensions and illustrate the consequences of doing so for inference.
Similar content being viewed by others
Notes
For a recent reference to the use of GTAP data in structural gravity model estimation, consider Caron et al. (2014).
Notice that the No RTA variable then measures trade costs beyond tariffs that are associated with an absence of RTA membership. Such costs relate to non-tariff barriers to trade and a lack of harmonized institutions, regulations, and legal standards (see Egger and Larch 2010). The inclusion of tariffs together with the No RTA indicator variable should reduce the potential endogeneity bias of the No RTA variable. Moreover, following the arguments in Baier and Bergstrand (2007), we assume that controlling for
multilateral resistance terms (or exporter and importer multilateral trade costs) in the gravity model removes most of the potential endogeneity bias of the No RTA variable.
For three countries, Luxembourg, Malaysia, and Singapore, the calculated figures of domestic shipments turned out negative. For these countries, we imputed estimated values using the predictions of a logistic model with \(\hbox {log}(z_i/z_\mathrm{US})\) as dependent variable, where \(z_i\) is defined as \((\hbox {GDP}_i-X_{i.})/(\hbox {GDP}_i-X_{i.}+X_{.i})\). The explanatory variables comprise \(\hbox {ln\,GDP}_i\), \(\hbox {ln\,X}_{i.}\), and \((X_{i.}-X_{.i})/\mathrm{GDP}_i)\), also taken relative to the USA. The negative values of \(X_{ii}\) are replaced by \((\widehat{z}_i)(\hbox {GDP}_i-X_{i.}+X_{.i})\). Note this imputation guarantees that the estimated domestic sales figures are strictly positive.
Notice that the parameter estimation is based on normalized trade flows \(\widetilde{x}_{ij}\), while the comparative static analysis is done for \(\phi _{ij}\). In the past, authors have often used \(\phi _{ij}\) as the dependent variable in estimation, but there is no need for this, in general.
References
Abadir KM, Magnus JR (2005) Matrix algebra. Cambridge University Press, Cambridge
Anderson JE, van Wincoop E (2003) Gravity with gravitas: a solution to the border puzzle. Am Econ Rev 93:170–192
Anderson JE, van Wincoop E (2004) Trade costs. J Econ Lit 42:691–751
Baldwin RE, Forslid R, Martin P, Robert-Nicoud F (2003) The core-periphery model: key features and effects. In: Heijdra BJ, Brakman S (eds) The monopolistic competition revolution in retrospect. Cambridge University Press, Cambridge UK, pp 213–235
Baier SL, Bergstrand JH (2007) Do free trade agreements actually increase members’ international trade? J Int Econ 71:72–95
Baltagi BH, Egger PH, Pfaffermayr M (2007) Estimating models of complex FDI: are there third-country effects? J Econ 140:260–281
Baltagi BH, Egger PH, Pfaffermayr M (2013) A generalized spatial panel data model with random effects. Econ Rev 32:650–685
Baltagi BH, Egger PH, Pfaffermayr M (2015) Panel data gravity models of international trade. In: Baltagi BH (ed) The Oxford handbook of panel data econometrics. Oxford University Press, Oxford, pp 608–641
Beenstock M, Felsenstein D (2015) Spatial dependence in the econometrics of gravity modeling. In: Patuelli R, Arbia G (eds) Spatial econometric interaction modelling. Springer, Berlin
Behrens K, Ertur C, Koch W (2012) ‘Dual’ gravity: using spatial econometrics to control for multilateral resistance. J Appl Econ 27:773–794
Bergstrand JH, Egger PH, Larch M (2013) Gravity redux: estimation of gravity-equation coefficients, elasticities of substitution, and general equilibrium comparative statics under asymmetric bilateral trade costs. J Int Econ 89:110–121
Cameron AC, Gelbach JB, Miller DL (2011) Robust inference with multiway clustering. J Bus Econ Stat 29:238–249
Caron J, Fally T, Markusen JR (2014) Skill premium and trade puzzles: a solution linking production and preferences. Q J Econ 129:1501–1552
Chen N, Novy D (2012) On the measurement of trade costs: direct vs. indirect approaches to quantifying standards and technical regulations. World Trade Rev 11:401–414
Egger PH, Larch M (2010) An assessment of the Europe Agreements’ effects on bilateral trade, GDP, and welfare. Eur Econ Rev 55:263–279
Egger PH, Larch M, Staub KE, Winkelmann R (2011) The trade effects of endogenous preferential trade agreements. Am Econ J Econ Policy 3:113–143
Egger PH, Nigai S (2015) Structural gravity with dummies only. CEPR discussion papers 10427, CEPR discussion papers
Egger PH, Pfaffermayr M (2011) Structural estimation of gravity models with path-dependent market entry. CEPR discussion papers 8458, CEPR discussion papers
Egger PH, Staub KE (2015) Controlling for multilateral resistance terms in linearized trade gravity equations without spatial econometrics. Unpublished manuscript, ETH Zurich
Haining R, Griffith D, Bennett D (1989) Maximum likelihood with missing spatial data and with an application to remotely sensed data. Commun Stat Theory Methods 18:1875–1894
Harville DA (1977) Maximum likelihood approaches to variance component estimation and to related problems. J Am Stat Assoc 72:320–338
Head CK, Ries J (2001) Increasing returns versus national product differentiation as an explanation for the pattern of US-Canada trade. Am Econ Rev 91:858–876
Helpman E, Melitz M, Rubinstein Y (2008) Estimating trade flows: trading partners and trading volumes. Q J Econ 123:441–487
LeSage JP, Pace RK (2008) Spatial econometric modeling of origin-destination flows. J Reg Sci 48:941–967
LeSage JP, Satici E (2013) A bayesian spatial interaction model variant of the poisson pseudo-maximum like-lihood estimator. Unpublished manuscript available for download at http://papers.ssrn.com/sol3/papers.cfm?abstractid=2372094
LeSage JP, Thomas-Agnan C (2015) Interpreting spatial econometric origin-destination flow models. J Reg Sci 55:188–208
MacKinnon JX, Webb X (2015) Wild bootstrap inference for wildly different cluster sizes. Working papers 1314, Queen’s University, Department of Economics
Magnus JR, Neudecker H (1999) Matrix differential calculus with applications in statistics and econometrics, 2nd edn. Wiley, New York
Mayer T, Zignago S (2011) Notes on CEPII’s distances measures: the GeoDist database. CEPII working paper 2011-25, http://www.cepii.fr/CEPII/en/publications/wp/abstract.asp?NoDoc=3877
Pfaffermayr M (2009) Maximum likelihood estimation of a general unbalanced spatial random effects model—a Monte Carlo study. Spat Econ Anal 4:467–483
Santos Silva JMC, Tenreyro S (2006) The log of gravity. Rev Econ Stat 88:641–658
Santos Silva JMC, Tenreyro S (2015) Trading partners and trading flows: implementing the Helpman-Melitz-Rubinstein model empirically. Oxf Bull Econ Stat 77:93–105
van Garderen KJ, Shah Ch (2002) Exact interpretation of dummy variables in semilogarithmic equations. Econ J 5:149–159
Acknowledgments
Egger acknowledges funding from the Czech Science Foundation through grant number GA ČR P402/12/0982. Both authors acknowledge numerous helpful comments from the participants at the CESifo-ETH conference on Gravity Models held at the ifo Institute at Munich in May 2014 and, especially, from two anonymous referees.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendices
1.1 Appendix A: The variance–covariance matrix of the disturbances given in (9)
1.2 Appendix B: The spatial ML estimator in unbalanced spatial panels
For estimation, one can partition \(\varOmega _\upsilon \) given in (9) and its inverse as
where \({\mathbf {S}}_{f}\) is obtained from skipping all rows of \({\mathbf {I}}_{C^2}\) that refer to observed values.
The likelihood will be maximized numerically using
to obtain the concentrated likelihood
1.3 Appendix C: The estimation of the variance–covariance matrix of the estimated parameters
The upper left diagonal block of the information matrix refers to the estimated vector slope parameters, \(\widehat{\beta }\), and, for balanced data, the estimate of the block is given by
where \(\phi _{\mu }=\frac{\sigma _{\mu }^{2}}{\sigma _{\varepsilon }^{2}}\) and \(\phi _{\lambda }= \frac{\sigma _{_{\lambda }}^{2}}{\sigma _{\varepsilon }^{2}}\). The result for the unbalanced case is analogous. The lower right block of the variance–covariance matrix contains the estimates of the sub-information matrix for the remaining parameters \(\widehat{\gamma }=(\widehat{\sigma }_{\varepsilon }^{2},\widehat{\phi }_{\mu },\widehat{\phi }_{\lambda },\widehat{ \rho }_{M},\widehat{\rho }_{L},\widehat{\rho }_{W})\). Harville (1977) provides a useful general differentiation result to derive this lower block information matrix referring to \(\widehat{\gamma }\):
Using
for the balanced panel model, one obtains
In the presence of missing observations, one can use (see Haining et al. 1989)
to derive estimates of the elements of the information matrix \(J_{rs,o}\) of the observed model. Plugging in the estimated parameters gives the required estimate of the second block of the information matrix. However, the numerical calculation will be involved given the complicated structure of \({\varvec{\Omega }}_{\upsilon ,oo}^{-1}\) and \(\frac{\partial {\varvec{\Omega }}_{\upsilon ,oo}}{\partial \gamma _{r}}\).
1.4 Appendix D
Rights and permissions
About this article
Cite this article
Egger, P., Pfaffermayr, M. A generalized spatial error components model for gravity equations. Empir Econ 50, 177–195 (2016). https://doi.org/10.1007/s00181-015-0980-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00181-015-0980-5