consistent with the hypothesis earlier – firms in TEs will do this more frequently as transition progresses and they catch up to the developed market economies. Again, we see a pattern of catching up and convergence. The gap is smaller or nonexistent with respect to upgrading. Interestingly, in 2005 there is virtually no variation across countries in terms of discontinuation of products: 15-17% of firms in Germany, the cohesion EU countries, and all the transition countries, report that they had discontinued a product in previous years. The exception is East Germany, which reports considerably lower levels of restructuring activity than observed in firms in West Germany or anywhere else.
The second pattern in the data is more of a surprise. The raw data show an inverse U- shape pattern of restructuring in the transition countries between 1999 and 2005: restructuring activity is low in 1999, high in 2002, and lower again in 2005. The other measures of restructuring available in the BEEPS surveys produce similar results (with the sole exception of “obtained a new quality accreditation”). To check the robustness of this finding, we estimated a restructuring activity equation (see below) on the full set of three years of data, including all the controls that we use below (firm size, sector, ownership, etc.), in order to remove any effects of changes in the composition of the BEEPS survey samples. The U-shape pattern appears strongly in these regression results (Table VI.2), across all the measures of restructuring we use, including the measure that captures both defensive as well as deep restructuring: discontinuation of product lines in the TEs peaks in 2002. The U-shape pattern extends across transition country regions as well.
This pattern is puzzling. A U-shape pattern is consistent with the literature on restructuring discussed earlier: if the early transition period is dominated by defensive restructuring, the opposite side of the coin is that deep restructuring starts with a lag and then accelerates. But why should the timing of the peak of restructuring activity in the new EU member states coincides with the one in middle income and low income CIS countries? It is widely believed that the more developed transition countries were faster reformers at the start of the transition period, and, hence, more restructuring activity in