Modelling investment when relative prices are trending: theory and evidence for the United Kingdom

In recent work, Stacey Tevlin and Karl Whelan argue that aggregate econometric models fail to capture the US investment boom in plant and machinery in the second half of the 1990s, whereas a disaggregated approach does much better. In particular, they show that aggregate models do not capture the increase in replacement investment associated with compositional shifts in the capital stock towards high depreciation rate assets, such as computers. And aggregate models invariably find little or no role for the real user cost, so do not pick up the strong effects of relative price declines on investment in computers. In this paper, a data set for the United Kingdom is constructed in order to investigate the ability of different equations to account for the UK boom in plant and machinery investment in the second half of the 1990s. The findings are similar to those of Tevlin and Whelan, whose analysis is extended in two main ways. First, the failure of the aggregate equations is explained more formally in terms of misspecification when relative prices are trending downwards. Second, the econometric analysis is conducted in a formal cointegration framework. As in the United States, the paper shows that asset-level equations can explain the investment boom in plant and machinery in the second half of the 1990s in the United Kingdom, whereas the aggregate equation fails completely.