4-7 Sept 2004ITS Biennial Conference, Berlin1 Usage Substitution Between Mobile and Fixed Telephone in the U.S. Michael R. Ward University of Texas at.

4-7 Sept 2004 ITS Biennial Conference, Berlin 1

Usage Substitution Between Mobile and Fixed Telephone in the U.S.

Michael R. Ward

University of Texas at Arlington

Glenn A. Woroch

University of California, Berkeley

ITS Biennial Conference—Berlin, Germany

4-7 September 2004

4-7 Sept 2004 ITS Biennial Conference, Berlin 2

Fixed-Mobile Substitution• Both provide phone access (dialtone, phone

number) and usage (outgoing local and LD, termination, calling features)– Main differences are mobility and clarity.

• Replacement of fixed line – “Cutting of the cord” has plateaued at about 5% of adults

in U.S., 6% in U.K.– Role of DSL, fax, number portability.

• Shifting of minutes to mobile– Mobile carving into fixed LD, and growing termination.– Growth of mobile calls and mobile-to-mobile calling.

4-7 Sept 2004 ITS Biennial Conference, Berlin 3

Policy Issues• Does mobile service constrain fixed line

pricing?– Should local service rate regulation be relaxed?– Is concern over leveraging fixed network

exaggerated?

• Should mobile count toward universal service?– Should USO subsidies target fixed line only?

• How assess fixed-mobile integration?– Allow bundling of fixed and mobile services?

• Is there a need to promote more competition?– Further unbundling of fixed network, or more

spectrum?

4-7 Sept 2004 ITS Biennial Conference, Berlin 4

The US vs. ElsewhereThe U.S.• Mobile

– Cellular party pays – Rapidly moving to flat rate– Predominantly subscription– Multiple standards– Intense competition– Negligible SMS

• Fixed – Flat rate local service– Short wait times – Some local competition, Intense

long distance competition– DSL, CM for second fixed lines

Europe, Asia, etc. • Mobile

– Calling party pays

– Low recurring, high usage rates

– Explosive growth of pre-paid

– Uniform standards

– Mild domestic competition

– SMS very popular

• Fixed – Measured service

– Long(er) wait times

– Limited local, LD competition

4-7 Sept 2004 ITS Biennial Conference, Berlin 5

The Data• TNST’s ReQuest® Market Monitor Database

– Nationwide, 30K+ per quarter, 10 quarters (3Q99-4Q01).– Survey responses and demographics from omnibus.

• TNST’s Bill Harvesting® Database– Fixed and mobile bill harvesting (25% response rate).– Not a panel but some re-sampling (about 10% of bill

submitters).

• Sampling Problems– Households, not individuals.– Some bias (skews older, white, lower income).– Voluntary bill submission causes data headaches.– Bills mask crucial information.

4-7 Sept 2004 ITS Biennial Conference, Berlin 6

Mobile Access and Usage 1999-2001

• Increased mobile access– Number of cellphones in U.S. increased from 76 to 128

million, while number of all (second-line, CLEC) fixed residential lines has fallen from 190 to 188 million.

– Percent of multiple cellphone HHs went from 15% to 25%.

• Increased mobile usage– Average minutes per line increased from 120 to nearly

300 per month.– LD minutes per HH on mobile increased 400% while

fixed LD per HH usage fell 14%. – Also, steady shift from peak to off-peak usage.

4-7 Sept 2004 ITS Biennial Conference, Berlin 7

Trends in Mobile Pricing

• Big buckets – Larger usage allowances– Longer off peak hours

• Bundling – Mobile to mobile, LD included.

• Expanded service– Large home zones or no roaming charges

4-7 Sept 2004 ITS Biennial Conference, Berlin 8

Summary Statistics

IntraLATAInterLATA Intrastate

InterLATA Interstate

Number of Calls Wireline 148,124 70,410 139,891

Mobile 148,124 34,543 98,371Percent of Calls Wireline 23.9% 67.1% 58.7% Mobile 76.1% 32.9% 41.3%Percent of Expenditures Wireline 42.4% 80.0% 67.4% Mobile 57.6% 20.0% 32.6%

Note: IntraLATA wireline calls only include calls for which toll charges would apply while IntraLATA Mobile calls include all calls placed.

4-7 Sept 2004 ITS Biennial Conference, Berlin 9

Source: ReQuest Survey

4-7 Sept 2004 ITS Biennial Conference, Berlin 10

Source: ReQuest Survey

4-7 Sept 2004 ITS Biennial Conference, Berlin 11

Source: ReQuest Survey

4-7 Sept 2004 ITS Biennial Conference, Berlin 12

Source: ReQuest Survey

4-7 Sept 2004 ITS Biennial Conference, Berlin 13

Source: Bill Harvesting data

4-7 Sept 2004 ITS Biennial Conference, Berlin 14

Source: Bill Harvesting data

4-7 Sept 2004 ITS Biennial Conference, Berlin 15

Source: Bill Harvesting data

4-7 Sept 2004 ITS Biennial Conference, Berlin 16

Demand Estimation• Estimate LA/AIDS demand “system” for each of

IntraLATA, InterLATA Intrastate & Interstate.

• Share = + mln(Pm ) + wln(Pw ) + “U” + • Consistent with utility theory (Hicksian).• Flexible functional form.• Unit of observation is LATA.• Use total calling as “U” constant.

4-7 Sept 2004 ITS Biennial Conference, Berlin 17

Prices

• Wireline easy – linear pricing.

• Mobile hard – non-linear pricing.– Conditional expected price per min. – Versus avg. rev. per min. (ARPM).– Versus simulated two-part pricing.

• Aggregate up to the LATA.

• IVs (LATAs and period dummies) to account for measurement error & possible endogeneity.

4-7 Sept 2004 ITS Biennial Conference, Berlin 18

Source: Bill Harvesting data

4-7 Sept 2004 ITS Biennial Conference, Berlin 19

Results

Coef. Std. Err Coef. Std. Err Coef. Std. ErrLn Wireline Price -0.090** -0.014 0.02 -0.011 -0.064** -0.012Ln Mobile Price 0.138** -0.015 0.124** -0.013 0.142** -0.013Total Duration 0.048** -0.006 0.034** -0.008 0.011 -0.006Quarterly Trend -0.006* -0.003 -0.010** -0.003 -0.027** -0.002Intercept -0.009 -0.046 0.814** -0.056 0.648** -0.055Observations 1450 1349 1512Adjusted R2 0.15 0.14 0.19

Table 2LA/AIDS Estimates with Trend

Dependent Variable: Wireline Share of Minutes

InterLATA InterstateIntraLATA

4-7 Sept 2004 ITS Biennial Conference, Berlin 20

Results

Coef. Std. Err Coef. Std. Err Coef. Std. ErrLn Wireline Price -0.083** -0.014 0.027* -0.011 -0.028* -0.013Ln Mobile Price 0.160** -0.012 0.153** -0.011 0.229** -0.012Total Duration 0.050** -0.006 0.037** -0.008 0.01 -0.006Intercept -0.013 -0.046 0.786** -0.057 0.702** -0.062Observations 1450 1349 1512Adjusted R2 0.14 0.09 0.22

LA/AIDS Estimates without TrendTable 3

Dependent Variable: Wireline Share of Minutes

InterLATA Intrastate InterstateIntraLATA

4-7 Sept 2004 ITS Biennial Conference, Berlin 21

Interpretation

Mobile Usage

Wireline Usage

Mobile Usage

Wireline Usage

Mobile Usage

Wireline Usage

With TrendMobile Price -1.19 0.26 -1.52 0.13 -1.42 0.20Wireline Price 0.19 -1.26 -0.03 -0.99 0.21 -1.10

Without TrendMobile Price -1.23 0.31 -1.73 0.18 -1.69 0.33Wireline Price 0.18 -1.25 0.01 -1.00 0.11 -1.05

IntraLATAInterLATA Intrastate Interstate

Table 4Elasticity Estimates from LA/AIDS Estimation

4-7 Sept 2004 ITS Biennial Conference, Berlin 22

Implications

• Discernible moderate substitution– Cross-elasticities are almost always positive.– Strongest results with best data (interstate).– Cross-elasticities there are 0.20-0.33.– Imply own-elasticity for good of -0.3, -0.7 & -

0.7.– Thought experiment: What if no mobile price

decline? If mobile price 100% higher, then wireline minutes 20-30% higher.

4-7 Sept 2004 ITS Biennial Conference, Berlin 23

Predictions?

• Are mobile usage prices likely to continue falling?– Mobile usage growing for non-price reasons.

– High FC, low VC technology.

– Prices track falling AC.

• More or less cross-elastic?– Reached limits to call substitutability? (cut cord)

– Mobile neophytes becoming experienced.

– Greater incorporation of mobile into lifestyle.

4-7 Sept 2004 ITS Biennial Conference, Berlin 24

Conclusions

• Some Usage Substitution– Surveyed consumers think this is big.– Estimates indicate moderate.

• Policy Implications– Estimated substitution not yet competitive constraint.

Access & interconnection revenues are vulnerable.– Mobile-fixed integration could alleviate inefficient

down-stream decisions of “variable proportions” consumer.

– More spectrum: more competitors or more services?