The answer--about seven percent for a 3-to-2 merger-- comes from a natural experiment constructed from prices of rental cars at different airports.
Table 6.―Predictions from Natural
Experiments Compared with Benchmarks
Number
of Firms
|
Increase
in HHI
|
Predicted
Market-Wide Average Percentage Price Increase
|
||||
Our
Prediction
|
Cournot Benchmark
|
Bertrand
Benchmark
|
||||
2
|
.500
|
16.1
|
20.0
|
12.0/15.6
|
||
3
|
.222
|
7.1
|
8.9
|
4.9/5.4
|
||
4
|
.125
|
4.0
|
5.0
|
2.7/2.7
|
||
5
|
.080
|
2.6
|
3.2
|
1.6/1.7
|
Each row in the table above shows how the estimated merger effects relate to the predicted merger effects from standard theoretical merger models (Cournot and Bertrand). So for example, the first row denotes a 2-to-1 merger (2 firms going to 1, or merger to monopoly) which leads to a 16.1% price increase. The second row is our aforementioned 3-to-2 merger which leads to a 7.1% price increase.
Because the estimated effects are close to those from the models, the authors interpret their results as lending "some support to reliance on model-based tools for predicting merger effects."
DISCLAIMER: I am one of the authors.
Erg. I hate the phrase "natural experiment." Call it what it is, an "example."
ReplyDelete