Sunday, October 7, 2012: 10:00 AM
Empirical evidence supporting the co-movement of capital and labor has been demonstrated in recent years for several large countries (Kim 2007, Dolman 2008, Clarke 1995, Kugler and Rapoport 2007). It is unclear from this previous literature whether co-movement is a general phenomenon extending to small, open economies, or idiosyncratic to the specific countries examined. Since the single-equation regression models in these earlier papers are not deduced from a theoretical trade model, it is also unclear whether co-movement is a theoretical prediction of a standard trade model or a specious conclusion based on an erroneously specified empirical model. Moreover, the previous literature lacks uniformity in data sets, assumptions, and estimation procedures, so evidence of co-movement in a larger, consistent data has not been shown.
In the present paper we address these issues in a two good, small country model. Goods are produced via neoclassical production functions, consumers have identical and homothetic preferences, and factors (capital and labor) are fully employed. Prior to migration the country is engaged in free trade in capital and a numeraire good at a given world rental rate. This model predicts that the country’s foreign direct investment (net capital outflow) varies directly with its capital endowment and the world rental rate, but inversely with its labor supply (population). Thus, for a given world rental rate, the co-movement of capital and labor is a theoretical conclusion. Linearizing discrete changes in the variables of the model gives us our estimating equation. Specifically, we regress the country’s domestic investment per capita, growth rate in employment (migration rate) and net change in the world rental rate on the country’s FDI per capita. Our priors are that the estimated coefficients are positive on domestic investment per capita and the world rental rate and negative on the migration rate.
Given the assumptions of the theoretical model, we focus on small developed (OECD) economies. We use the OECD stat data base for labor force flows and FDI inflows (and outflows), the world Penn tables for domestic investment, and the FED’s 3-year treasury rate as a proxy for the world rental rate. We employ annual data for the period 1990-2006. Thus, member countries (Chile, Estonia, Israel, Slovak Republic, and Slovenia) that have recently joined the OECD have been excluded. To control for idiosyncratic migration barriers (language and culture, country restrictions on mobility, etc.), we measure ‘openness’ by dividing total migration by population level. Using the U.S. as a benchmark, if this openness measure was ordinarily greater over the time period versus the U.S., the country was included in our panel as an ‘open’ country. Seventeen member countries are in our panel. To correct for fixed effects among countries and still have sufficient observations, we use the Arellano-Bond robust covariance matrix approach. Preliminary results indicate that co-movement is strongly significant in the panel. Further econometric work is currently in progress to complete the paper.
In the present paper we address these issues in a two good, small country model. Goods are produced via neoclassical production functions, consumers have identical and homothetic preferences, and factors (capital and labor) are fully employed. Prior to migration the country is engaged in free trade in capital and a numeraire good at a given world rental rate. This model predicts that the country’s foreign direct investment (net capital outflow) varies directly with its capital endowment and the world rental rate, but inversely with its labor supply (population). Thus, for a given world rental rate, the co-movement of capital and labor is a theoretical conclusion. Linearizing discrete changes in the variables of the model gives us our estimating equation. Specifically, we regress the country’s domestic investment per capita, growth rate in employment (migration rate) and net change in the world rental rate on the country’s FDI per capita. Our priors are that the estimated coefficients are positive on domestic investment per capita and the world rental rate and negative on the migration rate.
Given the assumptions of the theoretical model, we focus on small developed (OECD) economies. We use the OECD stat data base for labor force flows and FDI inflows (and outflows), the world Penn tables for domestic investment, and the FED’s 3-year treasury rate as a proxy for the world rental rate. We employ annual data for the period 1990-2006. Thus, member countries (Chile, Estonia, Israel, Slovak Republic, and Slovenia) that have recently joined the OECD have been excluded. To control for idiosyncratic migration barriers (language and culture, country restrictions on mobility, etc.), we measure ‘openness’ by dividing total migration by population level. Using the U.S. as a benchmark, if this openness measure was ordinarily greater over the time period versus the U.S., the country was included in our panel as an ‘open’ country. Seventeen member countries are in our panel. To correct for fixed effects among countries and still have sufficient observations, we use the Arellano-Bond robust covariance matrix approach. Preliminary results indicate that co-movement is strongly significant in the panel. Further econometric work is currently in progress to complete the paper.