[DSRP Evidence](https://dsrpevidence.org/)

# A Fallacy of Composition

## Details

**Authors** Caballero, Ricardo

**Year** 1991

**Publisher** National Bureau of Economic Research (Working Paper 3735)

**Kind of work** report

**Discipline** Economics

**Applied** false

[Read it at the publisher](https://doi.org/10.3386/w3735) 
10.3386/w3735

## In authors' words

### What they set out to do (purpose)

To show that statements valid for individual economic units need not hold for the aggregate, and the reverse.

### Who or what was studied (sample)

Macroeconomic models of price rigidity and of factor adjustment with many firms that are not fully synchronized.

### How they did it (methods)

Theoretical analysis using probability theory to constrain the joint behavior of large numbers of units.

### What they found (results)

Asymmetric pricing at the firm level does not necessarily produce asymmetric aggregate price adjustment, and asymmetric aggregate behavior does not require firm-level asymmetry. The same holds for asymmetric factor adjustment costs and aggregate capital and employment responses. Representative-agent models blur the distinction between individual and aggregate statements.

## Commentary

### In short

An economic analysis showing that the behavior of an aggregate cannot be read directly from the behavior of its individual firms, so conclusions at one level may fail at the other.

### Response

The finding is a demonstration that a whole (the aggregate) behaves differently from its parts (the firms) and that the way units are coupled, through partial synchronization, determines the aggregate result, which bears on systems and relationships. Caveats: the paper is a formal argument in economics, only the abstract and a summary were read, and distinction and perspective patterns are not claimed.

**Patterns it shows** S, R

Formal proof or model

**Added** 2026-10-26

**How to cite this** Caballero, Ricardo (1991). A Fallacy of Composition. National Bureau of Economic Research (Working Paper 3735).