More Resources

Tenure dependence in consumer-firm relationships: an empirical analysis of consumer departures from automobile insurance firms.


by Israel, Mark
RAND Journal of Economics • Spring, 2005 •

Updated Probability of Type [[tau] [[rho].sup.[epsilon].sub.3] [[rho].sup.[epsilon].sub.4] 1 .141 .146 2 .144 .151 3 .145 .157 4 .146 .162 5 .147 .165 6 .146 .168 7 .145 .171 8 .144 .173 9 .143 .174 10 .141 .176 11 .139 .177 12 .137 .177

Updated Probability

of Type

Departure [[tau] [[rho].sup.[epsilon].sub.5] Probability 1 .208 .071 2 .222 .082 3 .240 .077 4 .257 .067 5 .272 .056 6 .286 .060 7 .301 .053 8 .316 .048 9 .329 .048 10 .342 .039 11 .354 .037 12 .365 .037 All computations are for a base-case consumer. TABLE 8 Cumulative Decline in Departure Probability by Source

Total Tenure Effects Percent of Period Change Alone Change 2 (base) .0824 .0824 N/A 2 [right arrow] 3 -.0056 .0007 -12.5% 2 [right arrow] 4 -.0152 -.0046 30.3% 2 [right arrow] 5 -.0266 -.0141 53.0% 2 [right arrow] 6 -.0225 -.0048 21.3% 2 [right arrow] 7 -.0294 -.0102 34.7% 2 [right arrow] 8 -.0342 -.0137 40.1% 2 [right arrow] 9 -.0344 -.0114 33.1% 2 [right arrow] 10 -.0431 -.0219 50.8% 2 [right arrow] 11 -.0450 -.0230 51.1% 2 [right arrow] 12 -.0450 -.0212 47.1%

Selection Percentage Period Alone Change 2 (base) .0824 N/A 2 [right arrow] 3 -.0062 110.7% 2 [right arrow] 4 -.0114 75.0% 2 [right arrow] 5 -.0158 59.4% 2 [right arrow] 6 -.0195 86.7% 2 [right arrow] 7 -.0226 76.9% 2 [right arrow] 8 -.0253 74.0% 2 [right arrow] 9 -.0277 80.5% 2 [right arrow] 10 -.0297 68.9% 2 [right arrow] 11 -.0316 70.2% 2 [right arrow] 12 -.0332 73.8% All computations are for a base-case consumer. Tenure effects alone are computed by fixing [epsilon] = -.1275 to match the period-2 departure probability from the full model. Selection alone is computed by setting [[alpha].sub.1] ([tau]) = [[alpha].sub.1] (2) for all r [greater than or equal to] 2. TABLE 9 Estimates Without Heterogeneity or Tenure Effects

No No Tenure

Base Model Heterogeneity Effects Tenure Effects Increase: period 2-12 .162 .406 0

[[alpha].sub.2](2) 1.022 .895 1.091

[[alpha].sub.2](3) -.005 .039 0

[[alpha].sub.2](4) .037 .072 0

[[alpha].sub.2](5) .071 .098 0

[[alpha].sub.2](6) -.070 -.037 0

[[alpha].sub.2](7) .040 .062 0

[[alpha].sub.2](8) .027 .047 0

[[alpha].sub.2](9) -.018 .004 0

[[alpha].sub.2](10) .085 .097 0

[[alpha].sub.2](11) .010 .023 0

[[alpha].sub.2](12) -.015 .001 0 Heterogeneity Distribution

[sigma] ([epsilon]) .525 0 .598

[[epsilon].sub.1] -.955 N/A -.949

[[epsilon].sub.2] -.131 N/A -.265

[[epsilon].sub.3] -.039 N/A .123

[[epsilon].sub.4] .150 N/A .211

[[epsilon].sub.5] .806 N/A .841

[[rho].sub.1] .143 N/A .171

[[rho].sub.2] .362 N/A .329

[[rho].sub.3] .141 N/A .111

[[rho].sub.4] .146 N/A .145

[[rho].sub.5] .208 N/A .244 TABLE 10 Firm's Willingness to Pay for Higher Tenure in "Markup" Units

Base Tenure "Full" Tenure Base as Percentage Tenure Effects Effects of Full 3 .16 .48 33.3% 4 .35 .93 37.6% 5 .49 1.28 38.3% 6 .53 1.48 35.8% 7 .69 1.78 38.8% 8 .79 1.99 39.7% 9 .87 2.14 40.7% 10 .99 2.30 43.0% 1I .98 2.33 42.1% 12 .96 2.33 41.2% Computations are for a base-case consumer with [epsilon] set to zero. TABLE 11 Extensions

Extension 1

Standard Category Parameter Estimate Error Alternate company [[delta].sub.0] .039 .112

effect [[delta].sub.1] -.014 .010 Price sensitivity [[rho].sub.v] N/A N/A

error [v.sub.1] N/A N/A

[v.sub.0] N/A N/A

[[sigma].sub.v] ([v.sub.1]) N/A N/A Information [[beta].sub.0] N/A N/A

[[beta].sub.1](1) N/A N/A

[[beta].sub.1](2) N/A N/A

[[beta].sub.1](3) N/A N/A

[[beta].sub.2] N/A N/A

[[beta].sub.3] N/A N/A

[[beta].sub.4] N/A N/A

[[beta].sub.5] N/A N/A

[[beta].sub.6] N/A N/A True tenure effects [[alpha].sub.2](1) 1.168 .009

[[alpha].sub.2](l) -.142 .012

[[alpha].sub.2](l) -.006 .021

[[alpha].sub.2](1) .035 .019

[[alpha].sub.2](1) .067 .022

[[alpha].sub.2](1) -.073 .025

[[alpha].sub.2](1) .039 .030

[[alpha].sub.2](1) .026 .032

[[alpha].sub.2](1) -.020 .037

[[alpha].sub.2](1) .082 .044

[[alpha].sub.2](1) .010 .044

[[alpha].sub.2](1) -.019 .043 Heterogeneity [[epsilon].sub.1] -.950 .171

distribution [[epsilon].sub.2] -.130 .189

[[epsilon].sub.3] .049 N/A

[[epsilon].sub.4] .141 .212

[[epsilon].sub.5] .802 .230

[[rho].sub.1] .150 .091

[[rho].sub.2] .361 .130

[[rho].sub.3] .141 N/A

[[rho].sub.4] .146 .084

[[rho].sub.5] .202 .091

Extension 2

Standard Category Parameter Estimate Error Alternate company [[delta].sub.0] N/A N/A

effect [[delta].sub.1] N/A N/A Price sensitivity [[rho].sub.v] .391 .244

error [v.sub.1] .461 .388

[v.sub.0] -.296 N/A

[[sigma].sub.v] ([v.sub.1]) 1.213 .434 Information [[beta].sub.0] N/A N/A

[[beta].sub.1](1) N/A N/A

[[beta].sub.1](2) N/A N/A


10  11  12  13  14  
COPYRIGHT 2005 Rand, Journal of Economics Reproduced with permission of the copyright holder. Further reproduction or distribution is prohibited without permission.
Copyright 2005, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.
NOTE: All illustrations and photos have been removed from this article.


Browse by Journal Name:
Sponsored Links
Marketplace

Learn how to distribute a press release

All-new 2010 Ford Transit Connect
Today on Entrepreneur


Sign Up for the Latest in:
e-Business & Technology
Franchise News
Business Book Sampler
Starting a Business
Sales & Marketing
Growing a Business

E-mail*
Zip Code*