GLOBAL FACTOR TRADE: Statistical Tests on Technology and Absorption
We would also like to note, though, that the desire to bring new data sources to bear on the problem has carried a cost. Specifically, the factors available to us for this study are limited to capital and aggregate labor. We would very much have liked to be able to distinguish skilled and unskilled workers, but unfortunately the number of skilled and unskilled workers by industry is not available for most countries.
We would like to note how the reader should think about this factor “aggregate labor,” and why we do not believe this presents too great a problem for our study. There are at least a couple of interpretations that can be given. A first fact about our labor variable is that under most specifications the OECD countries are judged scarce in labor while the ROW is abundant in it. This suggests that one appropriate interpretation is that our variable labor is a very rough proxy for unskilled or semi-skilled labor. Note, though, that in most of our later implementations, labor is converted to efficiency units. If this is an appropriate way to merge skilled and unskilled, then the fact that these OECD countries are scarce in it suggests that this is true, even when we convert all labor to common efficiency units. We have little doubt that if it were possible to distinguish highly-skilled labor separately for our study that the US and some of the other OECD countries would be judged abundant in that factor. If you need that money quickly, you surely need a reliable online lender to give you a hand. Our quick loans cash might be coming your way in just a few hours, if that’s how soon you want it, because we always try to meet our customers in the middle. Apply for a loan with this and see how fast it happens!
These reservations notwithstanding, we believe that there are good reasons to believe that choice of factors does not confer an advantage to us over prior studies. Many of the factors we omit are land or mineral factors, which were the best performers for BLS and Trefler (1995). Hence their omission should only work against us. As we will see below, the factors that we do include exhibit precisely the pathologies (e.g. “mystery of the missing trade”) that have characterized the data in prior studies. Both points argue that selection of factors should not prejudice the results of our study.
In sum, we have constructed a rich new data set with compatible data for 10 OECD countries across a wide range of relevant variables. Importantly, we introduce to this literature direct testing on technology and absorption data of the central economic hypotheses in contest. Finally, although in some respects the available data fell short of our ideal, we do not believe that this introduces any bias toward favorable results.
Statistical Tests on Technology and Absorption
The principal hypotheses that distinguish alternative implementations of HOV concern technology and absorption. In prior work, researchers have selected technological and absorption parameters designed to allow the model of net factor trade to work as well as possible. Little data on technology (only that of the US) and no data on absorption were employed.
By contrast, our principal statistical tests will work directly with the data on technology and absorption. We consider a variety of models of technology and absorption suggested by theory and select a preferred model for each. In this respect, the formal statistical tests in this paper will be complete once we have selected the preferred models.