Gravity Model - International Markets, Institutions, and Policy - Example

Gravity Model- International Markets, Institutions and Policy Name: Course: Professor: Institution: City and State: Date: The superiority of theoretically derived gravity model against empirically based gravity model The gravity model has come a long way over the years. Presently, there are numerous forms of gravity models that can be derived based on microeconomics foundations. Notably, one of the highly used empirical works is associated with Anderson and Van Wincoop (2003) (“AvW”). One of the notable features of Anderson van Wincoop model is that it is quite easy to derive since it involves only four steps. The initial step involves developing the model’s consumption side. This side seeks to solve consumers’ problem by formulating a set of demand functions that explain consumer preferences. Next is the production side. With many firms engaging in monopolistic competition, strategic interaction wanes and therefore, firms adopt a single markup pricing rule. The next step is to work out the price pass-through equation which incorporates both foreign and domestic prices. This is attained by adding together variable trade costs and ad-valorem tariffs. Lastly, the model introduces equality of sectors expenditures, that is, incomes and aggregate demand of every sector in order introduce additional algebra to the model with a view of producing AvW gravity equation. The Anderson van Wincoop model is not only treated as the baseline gravity model, but also the most used model by many researchers (Puhani, 2000). This is because it is the only theoretically grounded gravity model. This means that, if a researcher considers the model as irrelevant, for instance the per capita GDP, then a researcher must explicitly justify the approach taken in terms of alternative theoretical assumptions. On the other hand, the gravity equation has been used empirically for trade analysis between countries. In this regard, Rose and van Wincoop (2001) defined it as the pillar of international trade due to its ability to estimate bilateral trade flows. Basically, there has been remarkable progress over the years in seeking to understand the theoretical basis of the equation, and also, in improving its empirical estimation. The present paper seeks to explain why the theoretical gravity model , that is, AvW “Gravity with gravitas” (2003) is superior to the empirically based gravity model. In line with Newton’s law of universal gravitation, a mass of goods, labor or other factors of production at a given origin are attracted to demand for products or labor at a given destination. However, the possibility of flow is negated by the distance between the two locations. In this vein, the analogy between physical force of gravity and trade clash due to lack of a set of parameters for an equation to hold. Nevertheless, traditional gravity allowed coefficients of mass variables and the bilateral distance to fit in a statistically inferred relationship between data mass variables, distance and trade flows (Santos and Tenreyro, 2006). Theoretically derived gravity model is expressed in the form of log-log form such that the parameters include elasticity of trade flow based on explanatory variables. In this vein, countries are considered to engage more in trade as opposed to the distance between them. Further, the model is augmented by political factors, which indicate that traded goods are accorded preferential treatment if they are from a preferred or unilateral source. According to the model, bilateral trade flows are based on variables in the right side of the gravity model. This gives a clear direction that runs from distance to trade and income. Nevertheless, the direction derived is based on the theoretical foundation that argues that the gravity equation is formed from microeconomic model where tastes for differentiated products and income are guaranteed. The gravity model is mainly concerned with bilateral trade between countries. According to Anderson and van Winhoop (2003), gravity is differentiated by its tractable and parsimonious representation of economic interaction between many countries in the world. Theoretically derived gravity model helps to estimate international trade using factors such as spatiality and geography. In this vein, the theory involved helps in predicting trade relative to factor abundances. One of the factor abundance models is Heckscher- Ohlin model (Samuelson, 1952). Anderson and van Winhoop (2003) derived a gravity model that separates traded and non-traded goods. Utility maximization using the income constraints provides details of the traded goods. Holding prices constant and with share relationships, foreign country's imports are derived assuming log-linear functions to obtain a gravity equation for aggregate imports. Anderson model is mainly concerned with the econometric properties of the resultant model, as opposed to the easy interpretability of theoretical implications (Baier and Bergstrand, 2009). Further, Anderson theoretically derived model is anchored on a microeconomic foundation that operates within the framework of general equilibrium of world trade. The model uses utility maximizing consumers with constant elasticity of substitution and profit maximizing firms’ production expertise. This form of analysis provided export and import equations which when combined with the equilibrium condition provide simplified form of equations for prices and quantities. The simplified equation eliminates endogenous variables from the explanatory part of the equation while at the same time prices and income ceases to be explanatory variables of bilateral nature. Theoretical derivation of the gravity model is based on Hecksher-Ohlin model that presents in two fronts. First, are the identical preferences and frictionless trade where consumers and producers have random choices of trading partners and trade impediments are neutralized by unequal factor prices (Martínez-Zarzoso, Nowak-Lehmann and Horsewood, 2009). Notably, it is not hard to provide justification for any form of the gravity equation on the foundations of standard theories. As opposed to empirical derivation, theoretical derivation takes into account various sources and numerous destinations of products. In addition, it indicates the relative attractiveness of the origin- destination combination. Basically, every sale has numerous possible destinations whereas every purchase has multiple sources. According to Millimet and Tchernis (2009), empirical derivation of the gravity model fails to accurately estimate the right border effects. Nevertheless, Aderson and van Wincoop (2003) in their theoretical derivation are able to clearly provide an estimation of structural equation based on non-linear least squares after working out multilateral resistance indices which are a function of observed bilateral distances and dummy variables that represent international borders. Theoretical method also provides an easier method due to incorporation of theories that provide explanations for multilateral price terms that help generate unbiased coefficient estimates. As such, this model as compared to the empirical approach estimates the gravity equation by use of unbiased co-efficient estimates. It is easy to understand the patterns of international trade by theoretically deriving gravity model compared to empirical analysis. This is because the theory supports the prediction of trade patterns based on comparative advantage, which has overtime suffered significant empirical challenges. The theory helps to predict with ease the aggregate preferences for goods demanded within countries. Based on comparative advantage theories, nations with a similar product preference are known to have similar industries. Notably, by using the gravity model, nations that earn similar levels of income engages in trade amongst themselves. According to Portes and Rey (2005), these countries trade in differentiated products due to their factor endowment similarities. The implications of theoretically derived gravity model superiority against the empirically derived gravity model for applied econometric analysis of trade flows Despite theoretical gravity model popularity, it cannot be used to develop econometric models though it can be incorporated with an empirical model to estimate the gains derived from liberalizing trade and also measuring the magnitude of barriers on trade. The problem of bridging economic theories with empirical models has resulted in intra-industry trade that is associated with reciprocal dumping. In this case, many nations involved have segmented their markets into homogeneous goods. As a result, they suffer imperfect competition leading to intra-industry trade that occur when these firms seek to expand their businesses to foreign markets by trading in goods that are not differentiated and those that lack comparative advantage due to lack of specialization. Such countries go contrary to the gravity model by portraying that trade depends on the size of the country. The theoretical prediction of the gravity model negates some empirical testing by suggesting that differentiated goods and specialization gravity equation might not adequately explain the gravity equation. According to Persson (2001), use of reciprocal dumping to assess the home market effect for homogenous and differentiated goods shows a favorable relationship in gravity estimation for differentiated goods but adverse relationship for homogeneous goods. According to Santos-Silva and Tenreyro (2006), theoretically derived model fails to concisely indicate the impact of other variables such as the exchange rate and the price level changes in the gravity model. The empirical model accommodates such additional variables thereby showing their relationship with other variables in the gravity model. According to the outcome of the empirically derived gravity model, the effects of price level vary based on the kind of relationship being examined. For example, if the variable in question is exports, an increase in price will lead to increased level of trade between the concern nations on the part of the importer. For a theoretical model, Anderson and van Wincoop (2003) explain an endogenous change of price level using non-linear systems of equations. However, the empirical models use simpler first order log-linearization to account for changes in world prices. Isard (1954) argues that application of the gravity model in estimating international trade does not always hold for econometric applications. The approach used in estimating international trade involves taking logs of both sides of the model which has proved not applicable especially where the volume of trade between countries equals zero. Similar problems were highlighted by Caruso (2003) who argued that estimating log-linearized models using least squares can cause significant biases. The effect of dynamics is another missing variable in the gravity model. It is always important to take into consideration the effects of dynamics. For instance, in the derivation of the gravity model of trade between countries involving numerous firms, the decision of such firms to sell their products to other firms in foreign markets relies on the firm’s ability to cover costs associated with entering a foreign market. The gravity model does not accommodate the dynamic variable are there may not predict the possible changes within the market. There are various forms of data relating to international trade. However, theoretically derived gravity model fails to accommodate such data. Nevertheless, even with application of such data on empirically derived gravity models, there have been econometric problems because of inconsistency of estimators applied on panel data. According to Melitz (2003), failure to observe a given nation specific effect, the inclusion of a dependent variable that is lagged on the right side of the model may lead to correlation between the error term and such dependant variable hence making the least square estimators inconsistent and biased. The theoretical derived gravity model is to a large extent hinged on theories. However, the empirical derived gravity model relies on figures. As such, the theoretical derived gravity model becomes popular due to its ease in applicability and forecasting future international trade patterns and flows. According to Liu (2009), empirically derived gravity model cannot apply in scenarios with zero data sets. The presence of zeros causes the desegregation causing a majority of bilateral flows with finely grained data to appear as inactive. The log of zero is undefined since it is impossible to raise a number to a given power and end up with zero. Zero trade flows cannot be solved using logarithmic functions. Data presented as zero according to Matyas (1997) may mean the absence of any shipment or may reflect shipment that falls below the threshold. Therefore, the presence of zero in any gravity model only serves to present challenges to an analyst since it becomes difficult to differentiate an economic model and the error term with which to base the econometric inference. References Anderson, J., and van Wincoop, E., 2003. Gravity with Gravitas: A Solution to the Border Puzzle. American Economic Review. Baier, S.L., and Bergstrand, J.H., 2009. Bonus Vetus OLS:A Simple Method for Approximating International Trade-Cost Effects Using the Gravity Equation. Journal of International Economics, 5 (2), pp. 69-84. Caruso, R., 2003. The impact of International Economic Sanctions on Trade. An Empirical Analysis. Peace Economics, Peace Science and Public Policy, 9(2), pp. 14-32. Isard, W., 1954. Location Theory and Trade Theory: Short-Run Analysis. Quarterly Journal of Economics, 68 (9), pp. 305- 322. Liu, X., 2009. GATT/WTO promotes trade strongly: sample selection and model specification, Review of International Economics, 17(3), pp. 428-446. Martínez-Zarzoso, I., Nowak-Lehmann, D.F., Horsewood, N., 2009. Are regional trading agreements beneficial? Static and dynamic panel gravity models. Journal of Economics and Finance, 20(1), pp. 46-65. Matyas, L., 1997. Proper Econometric Specification of the Gravity Model. The World Economy, 20 (5), pp. 363-369. Melitz, M.J. 2003. The impact of trade on intra-industry re-allocations and aggregate industry productivity. Econometrica, 5(6), pp. 1695-1725. Millimet, D. L., and Tchernis, R., 2009. On the specification of propensity scores, with applications to the analysis of trade policies. J Bus Econ Stat, 27(3), pp. 397-415. Persson, T. ,2001. Currency Union and trade: how large is the treatment effect. Economic Policy, 33(2), pp. 433-462. Portes, R., and Rey, H., 2005. The determinants of cross-border equity flows. Journal of International Economics, 65(2), pp. 269-296 Puhani, P.A., 2000. The Heckman Correction for Sample Selection and Its Critique. J of Econ Survey, 14(1), pp. 53-69. Rose, A.K., and van Wincoop. E., 2001. National Money as a Barrier to International Trade: The Real Case for Currency Union. American Economic Review, 91(2), pp. 386-390. Samuelson, P., 1952. The Transfer Problem and the Transport Costs: The Terms of Trade When Impediments Are Absent. Economic Journal, 35(21), pp.45-76. Santos, S., and Tenreyro, S., 2006. The Log of Gravity, the Review of Economics and Statistics, 88(4), pp. 641–658. Santos-Silva, J.M.C., and Tenreyro, S., 2006. The log of gravity. Rev Econ Statist, 88(5), pp.641- 658. Read More