Concentration of economic power is usually frowned upon because its social and political effects are substantial and undesirable. On the flip side one can argue that concentration provides the necessary scale to compete in international markets. Researchers use concentration measures like the HHI or Lerner’s Index, but coverage is typically restricted to the specific industries under study. Wouldn’t it be useful to have measures for the universe of industries, and at different levels of aggregation? And then aggregate these to calculate concentration for a country? This would allow us to say, for example, that India is more concentrated than the USA.

We address these questions in our paper. Using Prowess data (in this early draft), we find that one can estimate concentration ratios at different levels of industrial aggregation, ranging from the NIC Section to the NIC Sub-class for the universe of firms. However, these estimates can be unreliable because of heterogeneous data coverage across industries. This is more of a problem at lower levels of aggregation, but higher levels are not unaffected. We also find that country-level ratios lead to overestimates of concentration unless appropriate weights are used.