Productivity, Investment, and Current Accounts : Reassessing the Evidence

The most widely accepted explanation for the inverse association between private investments and current accounts [Glick and Rogoff, 1995] rests on data for manufactures through 1990. Is this consensus robust to revisions to the national accounts and the expansion of information technologies since 1990? To address this question I replicate their results, and I find that post-1990 developments eliminate the support for such a conclusion. I also implement alternative formulations and find, again, a lack of empirical support for their findings. Thus I examine the role of measurement errors and focus on the treatment of the manufacturing sector as representative of the whole economy and the exclusion of the contribution of capital when measuring productivity. Correcting these two measurement errors restores to Glick and Rogoff's conclusion its original strength