The Effect of State Tax Policies on Where E-Tailers Collect Sales Tax Donald Bruce, Professor of Economics William F. Fox, Professor of Economics LeAnn.

The Effect of State Tax Policies on Where E-

Tailers Collect Sales Tax

Donald Bruce, Professor of EconomicsWilliam F. Fox, Professor of Economics

LeAnn Luna, Associate Professor of Accounting

University of Tennessee – Knoxville, TNCenter for Business and Economic Research

• 45 states with a sales tax• States with sales taxes also have use

taxes– due on taxable items purchased

remotely

• A state can only require sellers with nexus to collect sales taxes

• Use tax compliance is very, very low

Policy Background

• Quill Corp. vs. North Dakota (504 US 298, 1992): – Physical Presence– How much and what type of “physical presence” is

needed?

• Varies by state– Owning/leasing tangible property– Storing goods in public warehouse– Owning display racks– Shipping in-process inventory for processing– Using company-owned vehicles or third-party

distributors– Having a local phone number, bank account, or P.O. Box– Out-of-state employees providing in-state services

What is Nexus? [nek-suhs]

• Affiliate or attributional relationships–Nexus of one entity can extend to affiliates

• “Amazon Laws” or click-through nexus– Small in-state sellers who receive

commissions for having an Amazon link on their websites can create nexus for Amazon

• These efforts to assert nexus are being challenged in the courts

State Creativity

Study Overview• Questions:– How do state tax policies affect the propensity

of firms to establish sales tax nexus in states?– How do nexus decisions impact revenues and

employment?

• How: – We combine a database with operating data for

the nation’s largest e-tailers (and aggregate to the state level) with a panel of state-level policy and economic date to estimate state-level regressions.

• Why e-tailers?5

6

E-Commerce Key Stats

$4.1 trillion – estimated total e-commerce for 2011

13.0% - annual compound growth rate since 2000.

92% - of all transactions are business to business, with the remainder business to consumer.

2/3 – of all e-commerce is in the manufacturing sector

$11.4 billion - Estimated 2012 total state and local revenue loss (Bruce, Fox, Luna).

• Representative firm decides whether to establish nexus in a state

• Depends on net impact of nexus on PROFITS– SALES might increase or decrease

• Home-state effect vs. “tax revolt”• Sign depends on economic and tax structure

– PRICES might increase or decrease• Over-shifting vs. under-shifting of the tax• Sign depends on economic and tax structure

– TRANSACTION COSTS are likely to decline• Tax collection costs offset by transportation cost savings

• Nexus is established when firms stand to gain profit through (a) quantity or price increases and/or (b) transaction cost savings

Theoretical Motivation

• Firm-level nexus decisions are likely to be impacted by state economic and policy environment– Sales tax rate– Breadth of sales tax base– Size of market (personal income)– Additional sales tax compliance costs– Other factors

Link to Empirics

• Internet Retailers Top 500 Guide• Hand-collected data for large sample of online

retailers (drawn for prior research projects)• Indicators for whether sales tax was collected on

a hypothetical order from each sales-taxing state

• 2006: 100 largest firms as of 2005• 2008: 50 largest firms plus 50 others in Top 500• 2010: Any firm in 2006 and 2008 data plus

several more (total of 179)• 2011: Slightly expanded to 182 firms• 2012: Expanded to 297 firms

Nexus Data

10

Amazon.com Inc. (#1)

Nexus

No sales tax

E-Sales $61,090,000,000

11

Walmart.com (#4)

Nexus

No sales tax

E-Sales $7,700,000,000

12

Dell Inc. (#8)

Nexus

No sales tax

E-Sales $3,900,000,000

13

Newegg.com (#14)

Nexus

No sales tax

E-Sales $2,800,000,000

14

L.L. Bean Inc. (#29)

Nexus

No sales tax

E-Sales $1,140,000,000

15

Lulu Lemon Athletica Inc. (#122)

Nexus

No sales tax

E-Sales $197,300,000

16

Harry and David Holdings (#132)

Nexus

No sales tax

E-Sales $171,900,400

17

National Geographic Society (#352)

Nexus

No sales tax

E-Sales $37,260,000

18

Weight Watchers Inter. (#61)

Nexus

No sales tax

E-Sales $486,400,000

19

eMusic.com, Inc. (#215)

Nexus

No sales tax

E-Sales $94,160,000

• Would ideally consider firm-level decisions

• We instead condense firm-level data to state-level aggregates for empirical work

𝑁𝑒𝑥𝑢𝑠𝑖,𝑡 = 𝛽0 + 𝛽1𝑆𝑃𝐼𝑖,𝑡 + 𝛽2(𝑆𝑃𝐼𝑖,𝑡∗𝑅𝑎𝑡𝑒𝑖,𝑡) +

𝛽3 ( 𝑆𝑃𝐼𝑖,𝑡∗𝐵𝑎𝑠𝑒𝑖,𝑡 ) + 𝛽4𝑅𝑎𝑡𝑒𝑖,𝑡 + 𝛽5𝐵𝑎𝑠𝑒𝑖,𝑡+

𝛼𝑖+𝛾𝑡+𝜖𝑖,𝑡

Empirical Strategy

• Nexus 1: Count of companies with nexus in that state

• Nexus 2: Share of total web sales occurring at firms with nexus in that state

• Nexus 3: Nexus 2 pre-multiplied by that state’s share of total national state and local sales tax collections

Nexus Variables

• State Personal Income (SPI, in $b)• State general sales tax rate (%)• Sales tax breadth (collections/rate in $b)• Streamlined Sales Tax Project “age”• Amazon law indicator• Corporate tax policies– Top tax rate, sales factor weight, and

indicators for combined reporting, throwback, LLC/LLP withholding, and LLC/LLP entity-level tax

Control Variables

Summary Statistics

2006 2010 2012

Nexus 1: Count of Companies 49.089 69.556 83.578Nexus 2: Share of Online Sales 65.836 56.353 59.879Nexus 3: Tax-Weighted Share 1.504 1.32 1.52Personal Income ($ billions) 245.978 268.453 296.638Sales Tax Rate (%) 5.327 5.563 5.555Sales Tax Base ($ billions) 93.158 87.245 96.179Years of SSTP Membership 0.356 2.111 3.156Amazon Law (0/1) 0 0.089 0.244

MeanVariable Mean Mean

Contemporaneous ModelsNexus 1 Nexus 2 Nexus 3

Personal Income 0.255 0.180 0.017(6.21)** (3.49)** (4.99)**

Sales Tax Rate* Personal Income -0.006 -0.01 -0.002(1.75) (2.19)* (5.97)**

Sales Tax Base* Personal Income 0.000 0.000 0.000(3.74)** (1.23) (1.00)

Sales Tax Rate 3.739 2.561 0.521(2.21)* (1.21) (3.82)**

Sales Tax Base 0.044 0.109 0.013(0.71) (1.43) (2.59)*

Years of SSTP Membership -0.636 0.23 0.007(2.56)* (0.74) (0.34)

Amazon Law 0.435 4.086 0.274(0.38) (2.84)** (2.96)**

Constant -17.811 15.631 -3.952(1.25) (0.88) (3.43)**

R 2 0.96 0.56 0.48N 225 225 225

• Corporate policy may affect sales tax nexus– Include 8 additional control variables

• Policy endogeneity: nexus drives policy?– Try first-lagged models

• Spatial correlations: nexus in one state related to nexus in neighboring states– Try Spatial Autoregressive models

• “Amazon” effect for Nexus 2– Drop state years with Amazon nexus– Include an indicator variable for Amazon nexus

Potential Concerns

Nexus on Collections and EmploymentSales Tax

CollectionsTotal

EmploymentRetail

EmploymentPersonal Income 3.092 -0.279 -0.137

(1.49) (0.70) (2.89)**Years of SSTP Membership 71.248 6.562 0.641

(1.96) (0.95) (0.77)Amazon Law -104.542 6.422 2.01

(0.60) (0.19) (0.50)Nexus 3 Index -6.899 1.846 0.22

(1.16) (4.67)** (4.66)**Sales Tax Rate 327.041

(1.49)Nexus 3 Index * Sales Tax Rate 2.267

(2.01)*Constant 2,016.35 2,807.26 343.335

(1.58) (30.17)** (30.85)**R 2 0.21 0.36 0.4N 225 225 225

• Firms are more likely to establish nexus in large-market states (i.e., with higher SPI)

• The large-market effect falls as the sales tax rate rises or the base broadens

• Amazon laws have a lagged, positive effect on nexus

• Nexus increases total and retail employment

• Nexus increases revenues at amounts that rise with the sales tax rate.

Conclusions