Overview

This chapter traces BDM–BDS measurement from establishment maps to economic interpretation and model calibration. Different multi-unit shares can change how analysts view entry, concentration, scalable production, and labor resilience, but they do not directly estimate market power or worker outcomes. In the authors’ model exercise, a firing tax equal to one year’s wages produces different BDM- and BDS-based simulated effects, demonstrating sensitivity rather than a real-world forecast or recommendation.

Learning outcomes

Treating branches as firms can overstate independent entry; linking them to parents can make concentration more visible.
Different multi-unit shares support implications about organization, but do not by themselves settle market power or worker outcomes.
Under a one-year-wage firing tax, BDM predicts larger simulated output and employment declines than BDS; these are sensitivities, not forecasts.

Keywords

BDMBDSmeasurementconcentrationresiliencecalibration

From Counting Units to Economic Meaning

The BDM–BDS comparison now has two mechanisms: some employers appear in one frame and not the other, while common firms can be represented with different establishment networks. This chapter asks what those mechanisms change beyond the count itself. The focus is the passage from establishment records to economic interpretation and then to policy modeling. Readers should be able to distinguish a direct observation—such as a higher multi-unit share—from an implication about concentration, labor resilience, or policy effects.

The central warning is that an establishment map is not an economic conclusion. It is an input to one. If branches are treated as independent firms, entry may look stronger. If those branches are linked to common parent organizations, the same activity may look more concentrated and more scalable. When the resulting size distribution is used to calibrate a model, the measurement choice can alter a simulated policy effect. Each step is meaningful, but none should be mistaken for the next step without stating the assumptions that connect them.

When a branch is mistaken for a firm

The linked evidence shows how strongly the unit boundary changes the apparent prevalence of multi-unit organizations. In 2016, the ECF and LBD differ in the share of firms classified as having multiple establishments: 6.7% in the ECF compared with 3.3% in the LBD. They also differ in the share of establishments associated with such firms: 37.9% versus 28.1%. For employment, the corresponding shares are 65.7% and 59.0%.

These figures are not three versions of the same statistic. The first asks how many firms are classified as multi-unit. The second asks how many establishments belong to those firms. The third asks how much employment is located in them. Keeping the denominators visible matters because a modest share of firms can account for a much larger share of establishments and employment. Multi-unit organizations are, by definition, larger in organizational footprint than single-establishment firms, so their representation affects several distributions at once.

The professional-services comparison makes the frame difference even larger. In Professional and Business Services, the ECF classifies 8.7% of firms as multi-unit compared with 2.2% in the LBD. The share of establishments at multi-unit firms is 41.5% in the ECF and 17.2% in the LBD. The share of employment at those firms is 70.2% versus 55.3%. The figures do not prove that one frame has discovered the true level of concentration. They establish that the measured organizational structure changes materially with the frame.

That change affects the story of entry. Suppose a parent firm operates ten local establishments. If an analyst treats each establishment as an independent firm, the market appears to have ten competitors or entrants where the organizational record identifies one parent organization. If the analyst links all ten locations to the parent, the market looks more concentrated.

Published by DeepMox

FAQ

How can firm linkage change the interpretation of competition?

It can make competition look stronger if branches are counted as independent firms and more concentrated if linked to a common parent; the evidence does not directly measure market power.

Why does the establishment-size distribution matter for a policy model?

It matters because the model uses the distribution to infer productivity and persistence parameters, which govern simulated responses to employment-reduction costs.

Are the firing-tax results forecasts of a real policy?

No. They are model outcomes under a particular theoretical structure and calibration, intended to show sensitivity rather than the observed causal effect of an actual tax.

Previous9 minBDM vs BDS: One Firm, Different Establishment Maps

Next10 minU.S. BDM vs BDS: Is There a Startup Boom?

BDM vs BDS: How Measurement Shapes Policy Models

From Counting Units to Economic Meaning

When a branch is mistaken for a firm

FAQ