1. Supplier switching costs and vertical integration in the automobile industry
Monteverde, Kirk and David J. Teece (1982).
Bell Journal of Economics, 13 (1):
Prepared by He Soung AHN
Sept 9, 2009

2. Hypothesis
Why do U.S. automobile companies take parts production in-house?
Main hypothesis:
Assemblers will vertically integrate when the production process generates specialized and non-patentable know-how

3. An assembler will chose to vertically integrate component production when:
Transaction-specific
know-how
and skills
Costly to switch to an alternative supplier
(high supplier switching costs)
Expose the assembler to opportunistic recontracting
“development activities for new automotive parts”
Testable hypothesis:
“The greater is the applications engineering effort associated with the development of any given automobile component, the higher are the expected appropriable quasi rents, therefore, the greater is the likelihood of vertical integration of production for that component.”

4. Methodology Dependent variable Independent variable
The extent of vertical integration by General Motors and Ford for U.S. production in 1976 for 133 automotive components
Manufactured either in-house or by an external supplier --- Probit model: binary choice
Independent variable
Interested primarily in the influence of applications engineering on vertical integration choices
Here, an engineering cost rating was used

5. 3 control variables:
Distinguishes components that are specific to a company from those that are generic
A dummy variable to control for systematic differences between General Motors and Ford
A set of dummy variables related to systems effect
In the vehicle “system,” the relationships between individual components must be considered because its design and implementation are tightly coordinated
Vertical integration is predicted for components with large systems effects.

7. Results

8. Results
The development effort associated with the design of any given automotive component is shown to be positively related to the likelihood of vertically integrated production of that component
Only components specific to a single assembler will be candidates for vertical integration
General Motors is more integrated into component production than Ford
Only a mild relationship between vertical integration and components with large systems effects exists

9. Conclusion
Transaction cost considerations surrounding the development and deepening of human skills appear to have important ramifications for vertical integration in the automobile industry, thereby supporting the transaction cost paradigm