Bruce A. Blonigen, Ronald B. Davies, and Keith Head critique a 2001 American Economic Review paper by Carr, Markusen, and Maskus (CMM) that supports the "knowledge-capital" model of multinational enterprises (MNEs). The knowledge-capital model combines horizontal and vertical motivations for foreign direct investment (FDI), where horizontal motivations involve proximity to customers to avoid trade costs, and vertical motivations involve locating unskilled labor-intensive production in countries with abundant unskilled labor. CMM uses U.S. affiliate sales data from 1986 to 1994 to support the knowledge-capital model. However, the authors argue that CMM's empirical framework mis-specifies the terms measuring skilled-labor abundance, key variables for vertical motivations. After correcting this specification error, the data no longer support the knowledge-capital model but instead strongly support the horizontal model, which predicts that affiliate activity between countries decreases as absolute differences in skilled-labor abundance widen. The authors also show that this result holds when using data from a wider variety of parent and host countries, including OECD data. They conclude that the data do not support the knowledge-capital model and instead support the horizontal model of MNEs. The authors also note that the CMM paper's results may be misleading due to the way skill differences are measured, as the skill difference term can take both positive and negative values, leading to sign reversals in the estimated coefficients. The authors argue that the correct specification of skill differences leads to coefficient signs that do not support the knowledge-capital model. The paper also discusses the implications of these findings for the broader debate on FDI and labor market effects.Bruce A. Blonigen, Ronald B. Davies, and Keith Head critique a 2001 American Economic Review paper by Carr, Markusen, and Maskus (CMM) that supports the "knowledge-capital" model of multinational enterprises (MNEs). The knowledge-capital model combines horizontal and vertical motivations for foreign direct investment (FDI), where horizontal motivations involve proximity to customers to avoid trade costs, and vertical motivations involve locating unskilled labor-intensive production in countries with abundant unskilled labor. CMM uses U.S. affiliate sales data from 1986 to 1994 to support the knowledge-capital model. However, the authors argue that CMM's empirical framework mis-specifies the terms measuring skilled-labor abundance, key variables for vertical motivations. After correcting this specification error, the data no longer support the knowledge-capital model but instead strongly support the horizontal model, which predicts that affiliate activity between countries decreases as absolute differences in skilled-labor abundance widen. The authors also show that this result holds when using data from a wider variety of parent and host countries, including OECD data. They conclude that the data do not support the knowledge-capital model and instead support the horizontal model of MNEs. The authors also note that the CMM paper's results may be misleading due to the way skill differences are measured, as the skill difference term can take both positive and negative values, leading to sign reversals in the estimated coefficients. The authors argue that the correct specification of skill differences leads to coefficient signs that do not support the knowledge-capital model. The paper also discusses the implications of these findings for the broader debate on FDI and labor market effects.