The Hoberg and Phillips Text Based Industry Classifications have a spatial representation.
All firms have a location in a product market space shaped as a unit sphere.
Competitive product markets are areas of the sphere where many firms are located. Concentrated areas
are sparsely populated.
Some regions of the product space have no firms residing there, as some text descriptions of
products would describe products with no demand, such as the word combination: "eggs", "paint" and "gardening".
The best way to tap the full research power of this product market grid is to use the Text-based Network Industry Classifications (TNIC),
which is a network way of identifying competitors to each firm. Competitors are firms residing in close proximity in product space to each firm based on a continuous measure of similarity.
Another key benefit of TNIC industries is that industry composition is updated annually, and our own research indicates
that the product market space itself thus dynamically changes over time. As a result, static fixed-location FIC
classifications miss out on much of the picture.
Hoberg and Phillips
Data Library
Robert H. Smith School of Business
University of Maryland
Van Munching Hall
University of Maryland
College Park, MD 20742-1815
Welcome to the Hoberg-Phillips Data LibraryData provided by Gerard Hoberg and Gordon Phillips Robert H Smith School of Business |
Text-based Network Industry Classifications (TNIC) data [click here]* These new industry classifications are based on firm pairwise similarity scores from text analysis of firm 10K product descriptions. Competitors are firm centric with each firm having its own distinct set of competitors - analogous to networks or a "Facebook" circle of friends. These new industry classifications are updated annually and offer more research flexibility, and are also more informative, than FIC (fixed industry) classifications such as SIC, NAICS, and the 10-K based FIC classifications below. Our research shows they sharply improve upon SIC and NAICS codes in explaining many different firm specific decisions, including firm profitability, mergers and dividends. These benefits are outlined in the readme file above, and in the Hoberg and Phillips (2010a,b) papers cited therein.Fixed Industry Classifications (like SIC codes) (FIC) data [click here]* This industry classification method is also based on firm pairwise similarity scores from 10K product descriptions. Just like SIC or NAICs industry classifications, this method assumes that membership in industries is transitive - if firm A is in firm B's industry and firm C is in firm B's industry, than firm A and C are in each other's industry. It can thus be used in the same way SIC or NAICS industry controls are used. We include this analog for convenience, as it can be highly useful in robustness analysis, or in analysis for which 10-K based industry fixed effects controls might be helpful.Industry Concentration Data (New Industry Classifications and fitted and actual Census HHI data) [click here]* Herfindahl Concentration data based on New "Hoberg-Phillips" Industries and also Census Data.Papers that develop and use these dataProduct Market Synergies and Competition in Mergers and Acquisitions: A Text-Based Analysis. [Download Paper] Gerard Hoberg and Gordon Phillips, 2010, Review of Financial Studies 23 (10), 3773-3811. Text-Based Network Industries and Endogenous Product Differentiation. [Download Paper] Gerard Hoberg and Gordon Phillips, 2010, University of Maryland Working Paper. Real and Financial Industry Booms and Busts. [Download Paper] Gerard Hoberg and Gordon Phillips, 2010, Journal of Finance 65 (1), 45-86. Product Dynamics, Characteristics and Dividends [Download Paper] Gerard Hoberg, Gordon Phillips and Nagpurnanand Prabhala, 2010, University of Maryland Working Paper. Conglomerate Industry Spanning [Download Paper] Gerard Hoberg and Gordon Phillips, 2010, University of Maryland Working Paper. |
|
|