Outcomes and Alternatives for Universal Telecommunications Service: A Case Study of Texas

Jerry Ellig Joseph RotondiSenior Research Fellow Legal Fellow

Motivation for the Study

Universal service has been a source of significant debate on both the federal and state levels

Previous Mercatus Center research has analyzed outcomes, forgone benefits, and performance measurement for federal universal service programs

Texas Legislature directed the Texas PUC to study and consider reform of Texas USF programs

This study applies methods we used to analyze federal USF in order to analyze effects of Texas USF and proposed reforms

Questions Addressed

Outcomes: How do the High-Cost and Lifeline programs affect rates, service availability, and subscribership?

Forgone Benefits: What does Texas give up in order to fund these programs? (What is the “total social cost” of raising the funds?)

Reform options: How would proposed reforms affect the size of the High-Cost programs and the forgone benefits?

Outcomes

Effect on subscribership =

# of subsidized subscribersx

% price change due to subsidy x

Elasticity of demand

Lifeline Prices

Note: “Local Rate” is a subscriber-weighted statewide average, assuming each company’s average residential rate is the midpoint in its range of residential rates.

Non-Lifeline Texas Lifeline

Federal Lifeline Only

Local Rate $9.76 $9.76 $9.76 Federal Subscriber Line Charge (max.)

$6.50 $0 $0

Federal Lifeline Discount

$0 -$3.50 -$1.75

Texas Lifeline Discount

$0 -$3.50 $0

Subtotal $16.26 $2.76 $8.01 Add-on Charges Texas state/local tax (avg. 9.5%)

$1.54

$0.26

$0.76

Texas USF Assessment

$0.78 $0.13 $0.38

Federal tax (3%) $0.49 $0.08 $0.24 Federal USF Assessment

$0.24 $0.04 $0.12

Total Price $19.32 $3.28 $9.52

Lifeline Effect on Subscribership

Total Effect of Lifeline

Effect of Texas Lifeline

Inputs Lifeline Subscribers (q1) 579,339 579,339 Price without Program (p2) $19.32 $9.52 Change in Price Due to Program (∆p) $16.04 $6.24 Price with Program (p1) $3.28 $3.28 Elasticity of Demand (∆q/q1)/(∆p/p1) -0.05 -0.05 Results Change in Quantity (∆q), subscribers 141,589 55,062 Change in subscribers as a % of Lifeline households

24.4% 9.5%

Change in subscribers as a % of total Texas primary lines

1.1% 0.4%

Total govt. expenditures on Lifeline $94 million $36 million Total expenditure per new subscriber $663 $663 Texas USF expenditures on Lifeline $24 million $24 million Texas USF expenditures per new subscriber N.A. $442

Estimating High-Cost Outcomes Large company residential High-Cost:

Use data in PUC Report, Table 5 Assume change in local rate due to subsidy = average per-line support in each support range Local rate is subscriber-weighted average rate for the large companies, assuming each

company’s average rate is the midpoint of its range of residential rates Change in subscribership is calculated for each subsidy range, then summed

Large company business High-Cost: # of subsidized business lines is estimated. Estimate assumes business lines are distributed among the per-line support ranges in the

same proportion as residential lines, and annual subsidy to business lines totals $65 million Remainder of calculation mimics calculations for large company residential high-cost

Small company High-Cost: Use data in PUC Report, Table 7 Assume change in local rate due to subsidy = average per-line support for each company Each company’s average local rate is assumed to be the midpoint of its residential rates Change in subscribership is calculated for each subsidy range, then summed

High-Cost Effect on Subscribership(Spreadsheet inputs)

Large Company, Residential

Large Company, Business

Large Company, Total

Small Company

Total, Large and Small

Inputs Subsidized Lines 2005 (q1) 1,165,007 225,837 1,390,844 412,025 1,802,869

Weighted average subsidized price (p1) $19.47 $36.95 N.A. $17.09 N.A.

Average subsidized price, range $19.47 $36.95 N.A. $13.66 – $30.29

N.A.

Total Annual Subsidy $367 million $65 million $432 million $80 million $512 million

Avg. monthly Support per Line (∆p), range

$3.73 – $226.39

$0.53 - $212.39

$0.53 – $226.39

$4.00 – $236.00

$0.53 – $236.00

Weighted average unsubsidized price (p2)

$50.69

$65.26

N.A.

$36.23

N.A.

Est. average unsubsidized price, range $23.90- $288.42

$37.59- $289.27

N.A.

$20.79 – $310.67

N.A.

Elasticity of Demand (∆q/q1)/(∆p/p1) -0.02

-0.02

-0.02

-0.02

-0.02

High-Cost Effect on Subscribership

Large Company, Residential

Large Company, Business

Large Company, Total

Small Company

Total, Large and Small

Results Change in Quantity (∆q), lines

37,356 3,459 40,816 9,201 50,017

Change in quantity as a % of subsidized lines

3%

1.5%

2.9%

2.2%

2.8%

Change in quantity as a % of total Texas primary lines

0.28%

0.03%

0.3%

0.07%

0.37%

Subsidy per addl. subscription caused by program

$9,834

$18,663

$10,582

$8,655

$10,228

Comments on Outcomes

Programs targeted to low-income subscribers are much more effective at increasing subscription

Likely due to higher elasticity of demand among low-income households

Small company High-Cost program appears slightly more cost-effective than large-company High-Cost program Small company rates are often lower than large company rates A dollar of subsidy represents a larger % reduction in price This generates a larger change in subscriptions per dollar spent

Results are consistent with previous scholarly research on the effects and cost-effectiveness of federal universal service subsidies

Forgone Benefits

Any resource allocation decision involves a forgone benefit: what was not done because resources were used one way rather than another way

The forgone benefit is larger than the money spent, because the universal service assessment mechanism increases the prices of telecom services

If consumers use less of a service because USF assessments increase its price: Consumers lose value they would have gained if they had used more Telecom firms lose revenues and operating profits Governments lose tax revenues

(Economists call these forgone benefits “deadweight loss.”)

Measuring Forgone Benefits

Forgone Consumer Benefits approximated by

.5 x change in price due to USF assessmentx

Change in quantity due to USF assessment

Forgone Benefits for telecom firms and govt. approximated by

Change in quantity due to USF assessmentx

(Price – marginal cost)

(Formulas are texbook economics. Use 2005 data and 5.65% USF assessment rate)

3 major USF contributors Local

Since demand is very insensitive to price, USF likely causes little change in quantity, and hence little deadweight loss

Long-Distance USF acts like a per-minute surcharge 1% price increase generates 0.7% reduction in minutes

used

Wireless USF acts like a per-minute surcharge 1% price increase generates >1% reduction in minutes

used

Forgone Benefits, Long-Distance

Long-Distance, 2005 Baseline (all figures annual) Inputs Price including taxes and USF (p1) 8.3 cents/minute Per Minute Universal Service Assessment (∆p) 0.4 cents/minute Price without TUSF Assessment (p2) 7.9 cents/minute % Change in Price Due to Assessment (∆p/p1) 4.8 percent Quantity (Q1), conversation minutes 9.9 billion Elasticity of Demand (∆q/q1)/(∆p/p1) -0.7 Change in Quantity (∆q), conversation minutes 331 million Marginal Cost (m) 1.7 cents/minute Results Estimated Universal Service Assessment (∆p⋅ Q1) $39 million Deadweight Loss (.5·∆p·∆q + ∆q·(p2-m)) $21 million Total Social Cost (∆p⋅ Q1 + .5·∆p·∆q + ∆q·(p2-m)) $60 million Deadweight Loss per $ of assessment 53.9 cents

Forgone Benefits, Wireless

Wireless, 2005 Baseline (all figures annual) Inputs Price including taxes and USF (p1) 8.2 cents/minute Per Minute Universal Service Assessment (∆p) 0.20 cents/minute Price without TUSF Assessment (p2) 8 cents/minute % Change in Price Due to Assessment (∆p/p1) 2.4 percent Quantity (Q1), minutes 133 billion Elasticity of Demand (∆q/q1)/(∆p/p1) -1.12 Change in Quantity (∆q), minutes 3.6 billion Marginal Cost (m) 4.5 cents/minute Results Estimated Universal Service Assessment (∆p⋅ Q1) $266 million Deadweight Loss (.5·∆p·∆q + ∆q·(p2-m)) $130 million Total Social Cost (∆p⋅ Q1 + .5·∆p·∆q + ∆q·(p2-m)) $396 million Deadweight Loss per $ of assessment 48.9 cents

Total Forgone Benefits, Texas USF

Totals, 2005 Baseline Estimated U. Service Assessment (∆p⋅ Q1) $618 million Deadweight Loss (.5·∆p·∆q + ∆q·(p2-m)) $151 million Total Social Cost (∆p⋅ Q1 + .5·∆p·∆q + ∆q·(p2-m)) $769 million Deadweight Loss per $ of assessment, wireline 0 Deadweight Loss per $ of assessment, long dist 53.9 cents Deadweight Loss per $ of assessment, wireless 48.9 cents Deadweight Loss per $ of assessment, overall 24.4 cents

Social Cost-Effectiveness(Change in subscribership divided by total social cost)

Texas Lifeline

Large Company High-Cost

Small Company High-Cost

Increase in subscribership caused by program

55,062 40,816 9,201

2005 expenditure $36 million $432 million $80 million Deadweight loss $11 million $106 million $19.5 million Total social cost $47 million $538 million $100 million Expenditure/increase in subscribership

$663 $10,582 $8,655

Social cost/increase in subscribership

$864 $13,168 $10,770

Reform Options Modeled

Turn back the clock (pre-2000 funding via access charges)

Per number charge

Rate equalization on subsidized lines

Increase revenue benchmark

Adjust to reflect cost reductions

Target subsidies to low-income subscribers

Turn Back the Clock

Pre-2000 system: implicit subsidies via high access charges, high long-distance rates, and toll revenue pooling

At 2005 prices and level of conversation minutes, it is mathematically impossible to set an access charge that raises anywhere near the $618 million in USF assessments collected in 2005

At 2000 prices and level of conversation minutes, an access charge of 7.15 cents/conversation minute raises $619 million. But deadweight loss would exceed revenues raised!

Main implication: The PUC was exactly right in 2000: Pre-2000 system was not sustainable

Turn Back the Clock

Intrastate Conversation Minutes

Price Per Minute

Access Charge

Revenues Deadweight Loss

DWL/ $Revenue

2005 19.7 billion $0.05 $0.036 $176 million $254 million $1.43 19.7 billion $0.07 $0.05 $247 million $385 million $1.56 19.7 billion $0.10 $0.072 $353 million $592 million $1.67 2000 26 billion $0.15 $.0715 $619 million $732 million $1.18

Per Number Charge (excl. Lifeline)

Per Number Charge for Texas Universal Service LINES Incumbent Local Exchange Carrier 9,974,507 Competitive Local Exchange Carrier 2,032,641 Cable 260,285 Wireless 15,011,531 TOTAL LINES 27,278,964 Minus Lifeline 579,339 TOTAL EXCL. LIFELINE 26,699,625 2005 Assessment $618,000,000 Charge per number per month $1.93

Effect of Per Number Charge

Deadweight Loss Declines Substantially Effects of a Per Number Universal Service Charge

Baseline Per Number Charge Estimated U. Service Assessment (∆p⋅ Q1) $618 million $612 million Deadweight Loss (.5·∆p·∆q + ∆q·(p2-m)) $151 million $86 million Total Cost (∆p⋅ Q1 + .5·∆p·∆q + ∆q·(p2-m)) $769 million $704 million Deadweight Loss per $ of assessment, wireline 0 0 Deadweight Loss per $ of assessment, long dist 53.9 cents 0 Deadweight Loss per $ of assessment, wireless 48.9 cents 25.1 cents Deadweight Loss per $ of assessment, average 24.4 cents 14 cents Deadweight loss is assumed to be zero for wireline. Deadweight loss for wireless calculated assuming elasticity of demand for wireless subscription = -0.57.

Rate Equalization Scenarios

Within Company: Raise rates on all of a company’s subsidized lines to highest rate charged by that company

Within Program: Raise rates on all subsidized lines in each program to highest rate charged by a company in that program

Across All Companies: Raise rates on all subsidized lines to highest rate charged by any incumbent wireline company in the state

Rate Equalization - Results

Savings range from $46-183 million Effects of Rate Equalization

Large Company Residential

Large Company Business

Small Company

Total

Baseline Expenditure $367 million $65 million $80 million $512 million Within Company Expenditure $338 million $54 million $74 million $466 million Savings $29 million $11 million $6 million $46 million Within Programs $16 maximum $41 maximum $19 maximum Expenditure $296 million $39million $26 million $361 million Savings $71 million $26 million $54 million $151 million Across All Companies $19 maximum $44 maximum $19 maximum Expenditure $268 million $35 million $26 million Savings $99 million $30 million $54 million $183 million

Increase Revenue Benchmark

“Revenue Benchmark” in Large Company program is deducted from per line cost to determine subsidy per line

Alternative yardsticks for setting benchmark National average rates Prices of competitive alternatives Assumed increases in local revenues due to vertical

services, broadband, or video

We model an increase in Large Company revenue benchmark plus an equivalent reduction in small company per-line subsidies

Effects of Increased Revenue Benchmark

Effects of Increased Revenue Benchmark

Large Company Residential

Large Company Business

Small Company

Total

Baseline Expenditure $367 million $65 million $80 million $512 million National Average Rate $14.52 $32.81 $14.52 Expenditure $309 million $50 million $47 million $406 million Savings $58 million $15 million $33 million $106 million VoIP Rate $14.95 $39.99 $14.95 Expenditure $306 million $40 million $45 million $391 million Savings $61 million $25 million $35 million $121 million Unsub.Wireless Rate $29.99 $29.99 $29.99 Expenditure $185 million $54 million $13 million $252 million Savings $182 million $11 million $67 million $260 million Satellite VoIP $49 $49 $49 Expenditure $91 million $23 million $5 million $119 million Savings $276 million $42 million $75 million $393 million Satellite Phone $99.95 $99.95 $99.95 Expenditure $30 million $9 million $139,711 $40 million Savings $337 million $56 million $80 million $473 million New Revenues $3.00 $3.00 $3.00 Expenditure $325 million $59 million $65 million $449 million Savings $42 million $6 million $15 million $63 million $5.00 $5.00 $5.00 Expenditure $303 million $55 million $55 million $413 million Savings $64 million $10 million $25 million $99 million

Adjustments for Cost Reductions

Actual changes in costs since 2000 are unknown

2 kinds of changes may have happened: Increased population density in “borderline” rural areas that have

become outer suburbs or vacation destinations Reduced costs of alternative technologies to serve very rural

areas

We examine results of hypothetical cost reductions to see if closer scrutiny of costs may be warranted Eliminate subsidies for least-subsidized lines Cap subsidies for most-subsidized lines

Hypothetical cost reduction scenarios(Per line subsidies are monthly, expenditures and savings are annual.)

Effects of Assumed Cost Reductions

Large Company Residential

Large Company Business

Small Company

Total

Baseline $367 million $65 million $80 million $512 million Cut lowest subsidy range < $10/line < $6/line < $10/line Expenditure $350 million $64 million $71 million $485 million Savings $17 million $1 million $9 million $27 million Cut 2 lowest subsidy ranges <$20/line <$16/line <$20/line Expenditure $300 million $57 million $25 million $382 million Savings $67 million $8 million $55 million $130 million Cap 2 high subsidy ranges $100/line $86/line $100/line Expenditure $344 million $58 million $79.52 million $482 million Savings $23 million $7 million $480,000 $30 million Cap highest subsidy range $150/line $136/line $150/line Expenditure $357 million $62 million $79.56 million $499 million Savings $10 million $3 million $440,000 $13 million

Target Subsidies to Low-Income Subscribers Exempt Lifeline subscribers from effects of reforms

that increase rates Continue High-Cost subsidies at current levels for any

Lifeline lines, or Increase funding for Lifeline to offset any rate increases

Lifeline subscribers are 4.8% of total subscribers

If they are distributed evenly, then targeting subsidies reduces savings from any reform by 4.8%

Reform Options Not Modeled

Provide all subsidies directly to consumers

Limit per line subsidies based on household income

Eliminate subsidies in deregulated exchanges

Eliminate subsidies where unsubsidized competitors offer service at price comparable to those they offer elsewhere

These would require significant extensions of the model

For more information …

About the Mercatus Center: www.mercatus.org

Universal Service Research: Performance measurement, comments to the FCCOct. 2005:http://www.mercatus.org/repository/docLib/MC_RSP_PIC200507FCCPerfMeasures_051017.pdfJan 2006:http://www.mercatus.org/repository/docLib/MC_RSP_ExPartePIC200602FCCPerfMeasures_060126.pdf

Consequences of federal USF“Costs and Consequences of Federal Telecommunications Regulation,” Federal Communications

Law Journal 58:1 (2006)http://www.mercatus.org/repository/docLib/MC_RSP_RPTJTelecomCostsandConseq_060307.pdf

Intercarrier compensation“Intercarrier Compensation and Consumer Welfare,” Journal of Law, Technology, & Policy (2005)http://www.mercatus.org/repository/docLib/MC_RSP_RPTJIntercarrierComp_060303.pdf