DO WIRELESS AND WIRELINE SERVICES COMPETE? Preliminary - please do not quote July, 2000 Michael R. Ward University of Illinois Glenn A. Woroch University of California at Berkeley Fariza Ahmad University of Illinois
Overview Motivation regulatory issues hinge in local monopoly Goal estimate substitution patterns Model econometrics not quite right yet Results focus on usage rather than access Implications LECs highest margin services are vulnerable
Motivation (1) Worldwide Growth of Wireless Service What factors drive cellular penetration? What explains regional and national differences? What role has privatization and competition played?
International Wireless as of Late 1998 Country Subscribers (mi.) Population (mil.) Rate Belgium 1.3 10.2 12.8% Denmark 1.5 5.2 29.0% Finland 2.5 5.1 49.5% France 8.1 58.6 13.9% Germany 11.4 83.5 13.6% Italy 16.0 57.5 27.8% Netherlands 2.6 15.6 16.5% Norway 1.9 4.4 42.7% Spain 5.5 39.2 14.1% Sweden 3.6 8.9 40.8% United Kingdom 9.8 58.1 16.8% United States 61.3 270.3 23.6%
U.S. Wireless Market 70 $100 60 Subscribers Avg Bill $90 $80 Millions of Subscribers 50 40 30 20 $70 $60 $50 $40 $30 $/Mth 10 $20 $10 0 $0 Dec-98 Jun-98 Dec-97 Jun-97 Dec-96 Jun-96 Dec-95 Jun-95 Dec-94 Jun-94 Dec-93 Jun-93 Dec-92 Jun-92 Dec-91 Jun-91 Dec-90 Jun-90 Dec-89 Jun-89 Dec-88 Jun-88 Dec-87 Date
Motivation (2) Potential for Wireless to Compete with Landline Service How readily do consumers substitute? Do wireless carriers constrain pricing of incumbent wireline carriers? Under what conditions would consumers drop wireline service for wireless? Will LECs lose even if households keep wireline?
Motivation (3) Policy Issues Does substitutability imply diminishing market power for wireline carriers? Should bundling of wireline and wireless services be permitted? At what point is relaxed regulation advisable? Diminished ability for discriminatory dealing implies lifting long distance line-of business restriction?
Substitutability Both provide voice access to PSTN (dialtone, phone number,etc.) Main difference is mobility Other differences: bandwidth, quality, availability, cost, numbering, bundled features
Complementarity Wireless facilitates intra-household communication Accessibility enhanced relative to second wireline
Modeling Wireless Choice Decision Making Wireless conditional on at least one landline Treated as a household decision (Max U HH ) Unobservable Attributes Access: connection to PSTN Mobility: option value Both provide these attributes to some degree (cordless phones, payphones, cell phones at home)
Factors Affecting Wireless Usage Cost of Service: equipment, activation, monthly and usage fees for both services (employer provided/reimbursed) Demographic factors: Number, age, education and mobility of HH members, HH income, home ownership Vertical Services: Custom calling features, voice mail, connection to fax or PC Complements: Fax, PC, ISP Substitutes: Payphones, pager s, work phone
The Data PNR s Bill Harvesting IV Database (1998) From large national panel, residential phone bills from ~30,000 households Selected by whether submit local phone bill(s) (hence can t model wireline replacement by wireless)
Summary Statistics 1994 1995 1996 1998 Observations 6,336 8,489 7,379 31,041 Wireless Price $42.07 $37.82 $35.25 $41.37 Wireline Price $28.97 $26.95 $26.76 $32.23 Wireless Subscribers 12.3% 13.8% 18.6% 25.1%
Summary Statistics (2) Any Local Toll Not a Wireless Subscriber 44.0% Wireless Subscriber 49.4% Average Local Toll Minutes 59.5 70.0 Any Long Distance 77.7% 81.8% Average Long Distance Minutes 142.4 165.2
The Econometric Model First Stage: Predicting Prices Account for non-linear pricing: Expenditures = " + $ Minutes + ( Minutes 2 " and $ by geographic market Instrumental Variables Use estimates of " and $ to as expected marginal prices for subscription and usage Do this for both Wireless and Local Telephone Bills
Econometric Model (2) Second Stage: Telephone Choices Logit of wireless subscription as function of wireless and wireline access and usage prices Reg of minutes of wireless calls... Same for local toll Same for LEC handled long distance
Econometric Model (3) Interpretation Change in number of customers calling is extensive margin from logit Change in calling amount per customer is intensive margin from regression Need to sum both extensive and intensive margins
Wireless Access and Usage Elasticity Results Access Any Usage given Access Usage Amount given any Usage Price Local Access 0.044-0.876* -0.409* Price Local Usage 0.139* 0.029 0.074 Price Wireless Access -0.030 0.072 0.538* Price Wireless Usage -0.041 0.251* 0.078
Local Toll Usage Elasticities Any Usage Usage Amount Wireless Not Wireless Not Price Local Access -0.916* -0.855* -1.004* -0.814* Price Local Usage -0.160* -0.171* -0.833* -0.678* Price Wireless Access 0.141-0.286 Price Wireless Usage 0.070-0.059
LEC Handled Long Distance Usage Elasticities Any Usage Usage Amount Wireless Not Wireless Not Local Access Price -0.057-0.287* -0.246* -0.206* Local Toll Price 0.224* 0.027-0.063 + -0.037 + Wireless Access Price 0.425 + 0.299 + Wireless Usage Price 0.155 0.132 +
Future Work Fix Econometric Model Links in Access and Usage demand imply more complex model Fariza is working on it New data for 1999 in hand as of last week Deal with household idiosyncratic effects? Other Services? Cable TV, ISP
Research Implications Substitutability is likely to be in usage, not subscription I ve got ISP, Broadband (soon!), wireless and no intention of dropping wireline. Substitutability likely to increase Wireless prices fall, quality improves, features added. Can wireless infuse competition where CLEC don t? Facilities, not resell; Broadly available; Enough licenses?