MAY 1999

NO. 31B A C K G R O U N D P A P E R

BY PATRICK FLEENORSENIOR ECONOMIST

TAX FOUNDATION

How Excise TaxDifferentials AffectCross-Border Sales

Of Beer in theUnited States

About the Tax FoundationIn 1937, civic-minded businessmen envisioned an independent group of researchers who, by

gathering data and publishing information on the public sector in an objective, unbiased fashion,could counsel government, industry and the citizenry on public finance.

Six decades later, in a radically different public arena, the Tax Foundation continues to fulfillthe mission set out by its founders. Through newspapers, radio, television, and mass distributionof its own publications, the Foundation supplies objective fiscal information and analysis topolicymakers, business leaders, and the general public.

The Tax Foundation's research record has made it an oft-quoted source in Washington andstate capitals, not as the voice of left or right, not as the voice of an industry or even of businessin general, but as an advocate of a principled approach to tax policy, based on years of profes-sional research.

Today, farsighted individuals, businesses, and charitable foundations still understand the needfor sound information on fiscal policy. As a nonprofit, tax exempt 501(c)(3) organization, the TaxFoundation relies solely on their voluntary contributions for its support.

©1999 Tax Foundation

* * *

Price: $25Annual subscription for six issues: $60

* * *

Tax Foundation1250 H Street, NW

Suite 750Washington, DC 20005

202-783-2760 Tel202-783-6868 Fax

www.taxfoundation.orgtf@taxfoundation.org

Table of Contents

Executive Summary ................................................................................... 1

Introduction............................................................................................... 3

Packaged Beer Sales by State .................................................................... 4

A Model of Beer Sales by State .................................................................. 6

Net Interstate Importing States ......................................................... 8

Net Interstate Exporting States.......................................................... 10

Canadian Cross-Border Traffic .......................................................... 11

A Model of Demand for Packaged Beer Sales and Its Supply by Source .............................................................. 13

Developing the Model ........................................................................ 13

Estimation ........................................................................................... 15

Results ................................................................................................. 15

Conclusion ................................................................................................. 17

Notes ........................................................................................................... 18

References.................................................................................................. 18

1

Source: Tax Foundation.

Effective Tax Rates on PackagedBeer and the Resulting Cross-Border Purchases: Delaware

and Neighboring States

Thousands of Cases of Beer

1997

Executive SummaryTax competition is the bane of governmentsworldwide that seek higher levels of taxreceipts. On the other hand, tax competi-tion is an effective force protecting taxpay-ers from higher taxes.

State and local governments must com-pete for private sector investment in muchthe same way that companies compete forcustomers. A generally low-tax state willlure private investment away from statesthat impose higher tax burdens. For ex-ample, if a state levies a low sales tax, it willlure customers from neighboring states toits retailers. Similarly, if a state imposes abeer excise that is lower than in surround-ing states, the low-tax state will sell more

beer to the residents of surrounding stateswho live near the border.

In effect, the demonstration in this pa-per of cross-border beer sales is an exampleof tax competition among the states gener-ally and a microcosm of tax competitionamong nations.

From the perspective of sellers and pur-chasers of any product, the market compe-tition that drives down price differentialsamong sellers is stymied when sellers arelocated in different governmental jurisdic-tions with different tax laws. Higher taxesmust be reflected in higher prices, so whensellers in low- and high-tax jurisdictionsserve overlapping markets, cross-bordershopping is the result. (See maps belowshowing the flow of beer across bordersfrom low-tax to high-tax states.) The greater

DE2.8%

NJ7.7%

MD6.4%

PA7.6%

321.9

1,075.9

270.2

Source: Tax Foundation.

Effective Tax Rates on PackagedBeer and the Resulting Cross-Border Purchases: Alabama

and Neighboring States

Thousands of Cases of Beer

1997

GA12.7%

AL21.7%

MS13.3%

TN9.8%

809.7

621.7

235.6

FL16.1%

839.8

2

the price differential, the more likely it isthat individuals living in border areas ofhigh-tax jurisdictions will cross the borderto shop in the low-tax jurisdiction.

The sale of beer in the United States isan interesting case study of this cross-bordershopping phenomenon. By constructing a

sophisticated econometric model thatcalculates demand for beer by state andthen examines its supply by source, we canexamine the patterns of cross-borderactivity.

Briefly, the model estimates the demandfor packaged beer by residents of differentstates by controlling for factors such asabstinence rates, differences inconsumption, and purchases bynonresidents. It then uses sales data toestimate the degree to which this demandis met by cross-border purchases. While themodel accurately estimates shifts in

While cross-border shopping obviouslyaffects sales and product-specific excisetax collections, ... it also affects incomeand property tax collections, licensefees, and a host of other sources ofgovernment revenue.

government revenue collected from stateand local sales taxes and state excise taxes,it does not include other effects of cross-border shopping, such as its effects onrevenue from income taxes, property taxes,and license fees. These unmeasured effectscan match or even exceed the revenuechanges measured by the model, so eventhough the results of this study aresignificant in themselves, governmentpolicymakers should be aware that theeffects of cross-border shopping are evenmore extensive.

The extent of the cross-border shoppingthat occurs in these circumstances can haveimportant fiscal effects on state and localgovernments. While this migration of eco-nomic activity across jurisdictional linesobviously affects sales and product-specificexcise tax collections, the story does notend there. Cross-border shopping affectsincome and property tax collections, licensefees, and a host of other sources of govern-ment revenue.

Policymakers are frequently surprisedby the magnitude of the revenue effectscaused by cross-border shopping. Such sur-prises can be particularly unnerving whenthe government in question is required tomaintain a balanced budget. This studyalerts policymakers to the extent of cross-border shopping and its revenue effects.

3

IntroductionIn a market economy, competition will

tend to drive down and eliminate any pricedifferentials that exist among sellers. How-ever, this price-leveling effect is stymiedwhen sellers are located in different govern-mental jurisdictions with different tax laws.When sellers in low- and high-tax jurisdic-tions serve overlapping markets, cross-bor-der shopping is the result. The greater theprice differential, the more likely it is thatindividuals living in border areas of high-taxjurisdictions will cross the border to shopin a low-tax jurisdiction.

The sale of beer in the United States isan interesting case study of this phenom-enon. The price of beer varies widely acrossthe country. In 1997, for example, theweighted-average price of packaged beerranged from a high of $20.23 per case inAlabama to a low of $16.89 in Oregon.1

Most of this variation is attributable to dif-ferences in state and local sales taxes andbeer excise taxes. More economically sig-nificant than an Alabama-Oregon compari-son are the instances where high- and low-tax jurisdictions border one another.Washington, for example, which levies astatewide 6.5 percent sales tax, additionallocal sales taxes and a $7.172 per barrelbeer excise tax, shares a border with Or-egon, which levies no state or local salestaxes and has a state beer excise of just$2.60 per barrel.

The extent of the cross-border shoppingthat occurs in these circumstances can haveimportant fiscal effects on state and localgovernments. While this migration of eco-nomic activity across jurisdictional linesobviously affects sales and product-specificexcise tax collections, the story does notend there. Cross-border shopping affectsincome and property tax collections, licensefees, and a host of other sources of govern-ment revenue.

The magnitude of cross-bordershopping’s revenue effects often surprisespolicymakers. Such surprises can be par-

ticularly unnerving when the governmentin question is required to maintain a bal-anced budget. This study attempts to alertpolicymakers about the extent of cross-bor-der shopping and its revenue effects.

The prospect of significant, tax-drivencross-border sales is one dimension of alarger issue—tax competition among gov-ernments. In general, state and local gov-

ernments must compete for private sectorinvestment in much the same way that com-panies compete for customers. If a state lev-ies a low tax burden, then it will lure pri-vate investment away from states thatimpose higher tax burdens. For example, ifa state levies a low sales tax, it will lure cus-tomers from neighboring states to its retail-ers. Similarly, if a state imposes a beer excisethat is lower than in surrounding states, thelow-tax state will sell more beer to the resi-dents of surrounding states who live near theborder.

Tax competition is the bane of govern-ments worldwide seeking higher levels oftax receipts. On the other hand, tax com-petition is an effective force protecting tax-payers from higher taxes. In effect, the dem-onstration in this paper of cross-border beersales is a microcosm of the consequencesof tax competition among the states gener-ally, and among nations.

The magnitude of cross-border shopping’srevenue effects frequentlysurprises policymakers.Such surprises can beparticularly unnervingwhen the government inquestion is required tomaintain a balancedbudget.

4

Packaged Beer SalesBy State

The amount of beer sold within the geo-graphic confines of a state is the sum of salesto residents and nonresidents. Sales to resi-dents is a function of the number of indi-viduals in a state who drink beer, their con-sumption levels, and their ability to takeadvantage of lower-priced beer in neighbor-ing jurisdictions. Nonresident sales are afunction of the number of individuals whointentionally travel into the state to takeadvantage of low-priced beer plus sales tononresidents who incidentally purchasebeer while traveling in the state for businessor pleasure.

Looking at the variation in packagedbeer sales by state gives an indication of theeffect these factors have on packaged beersales in a state.2 This variation is illustrated

Source: Tax Foundation.

Var

iation fro

m the A

vera

ge in

Gal

lons

+10

+8

+6

+4

+2

17.6Gal.

–2

–4

–6

–8

–10

NH NV TX NM LA SC DC AZ MS HI ND MT WI FL DE WY AKMO NE SD OH NC GA TN VT RI CO IL AL

Figure 1Variation Among the States in Per Capita Sales of Packaged Beer

1997

MEWV IA AR VA OK IN MI MN KY MA CA OR PA KS ID MD WA CT NJ NY UT

in Figure 1, which ranks states by their percapita packaged beer sales as compared tothe national average for 1997, the latest yearfor which data is available. The state withthe highest per capita packaged beer saleswas New Hampshire, which sold 29.4 gal-lons per capita, 9.6 gallons more than thenational average of 19.8 gallons. While thelatest data from the U.S. Department ofHealth and Human Services shows that therate of alcohol use is relatively high in NewHampshire, cross-border shopping alsoplays a role.

New Hampshire’s lack of state or localsales taxes has long drawn shoppers from sur-rounding states and Canada. Table 1 showsthat the state enjoys a price advantage overall of its neighbors, Maine, Vermont and Mas-sachusetts. Canada’s prices are also higher.Packaged beer sales were also high in Nevada,Texas, New Mexico, and Louisiana. In eachof these cases the states enjoy price advan-tages on most or all of their borders.

5

At the other end of the spectrum wasUtah, which sold just 11.5 gallons of beerper capita, 8.2 gallons less per capita thanthe national average. Table 1 shows thatUtah’s beer prices are higher than in any ofthe neighboring states except New Mexico.While this price differential undoubtedly ac-counts for part of the state’s low beer sales,alcohol sales are generally low in Utah dueto the large fraction of the population whobelong to the Church of Jesus Christ of Lat-ter-Day Saints (Mormons), which stronglydiscourages drinking by its adherents.

New York has the second lowest rate ofbeer sales despite being a popular touristdestination. Tourism invariably means higherconsumption, but in 1997 the sales NewYork lost to border states Pennsylvania, Mas-sachusetts and New Jersey outweighed theextra sales to tourists, driving per capita beersales down to 14.8 gallons, or 4.9 gallons lessthan the national average.

Table 1Variation Among the States in Per Capita

Packaged Beer Sales

1997

Per Capita Variance from AveragePackaged Beer National Average PriceSales (gallons) (gallons) ($ Per Case)

United States 19.8 0.0 $18.20

New Hampshire 29.4 9.6 17.45Nevada 28.6 8.8 17.95Texas 25.7 5.9 18.15New Mexico 25.6 5.8 18.58Louisiana 25.4 5.6 18.62

South Carolina 24.6 4.8 19.44District of Columbia 24.5 4.7 17.78Arizona 24.2 4.4 17.89Mississippi 24.0 4.2 18.84Hawaii 23.6 3.8 19.64

North Dakota 23.5 3.7 17.85Montana 23.5 3.7 17.03Wisconsin 22.9 3.1 17.65Florida 22.4 2.6 19.29Delaware 21.9 2.1 17.08

Wyoming 21.9 2.1 17.50Alaska 21.6 1.8 17.58Missouri 21.3 1.6 17.63Nebraska 21.0 1.2 18.03South Dakota 21.0 1.2 17.99

Ohio 20.9 1.1 18.03North Carolina 20.5 0.7 18.96Georgia 20.3 0.5 18.73Tennessee 20.3 0.5 18.25Vermont 20.2 0.4 18.12

Rhode Island 20.1 0.4 17.99Colorado 20.1 0.4 17.43Illinois 20.1 0.3 17.86Alabama 20.1 0.3 20.23Maine 19.6 –0.2 18.49

West Virginia 19.5 –0.2 18.05Iowa 19.1 –0.7 17.99Arkansas 19.0 –0.8 18.16Virginia 18.9 –0.9 18.09Oklahoma 18.9 –0.9 18.34

Indiana 18.7 –1.1 17.73Michigan 18.7 –1.1 18.11Minnesota 18.4 –1.4 17.90Kentucky 18.1 –1.7 17.80Massachusetts 18.0 –1.8 17.71

California 17.9 –1.9 18.29Oregon 17.9 –1.9 16.89Pennsylvania 17.3 –2.5 17.88Kansas 17.1 –2.7 18.00Idaho 17.1 –2.7 17.82

Maryland 16.7 –3.1 17.67Washington 16.6 –3.2 18.49Connecticut 15.2 –4.6 18.10New Jersey 15.0 –4.7 17.89New York 14.8 –4.9 18.02Utah 11.5 –8.2 18.48

Source: Tax Foundation

6

A Model of BeerSales by State

While a cursory examination of aggre-gate sales data gives some indication of thevarious factors affecting state beer sales, theinformation such an analysis can yield is lim-ited. Much more information can begleaned by constructing a sophisticatedeconometric model that calculates demandfor beer by state and then examines its sup-ply by source. The Tax Foundation has de-

veloped such a model, and it is describedin detail starting on page 13.

Briefly, the model estimates the demandfor packaged beer by residents of differentstates by controlling for factors such as ab-stinence rates, differences in consumption,and purchases by nonresidents. It then usessales data to estimate the degree to whichthis demand is met by cross-border pur-chases. While the model accurately esti-mates shifts in government revenue col-lected from state and local sales taxes andstate excise taxes, it does not include othereffects of cross-border shopping, such as its

Source: Tax Foundation.

Figure 2Effective Tax Rates on Packaged Beer by State

(State & Local Sales Taxes and Beer Excise Taxes)

1997

under 6.0%

7.0% to 8.0%

8.0% to 9.0%

TX9.2%

LA12.0%

MS13.3%

AL21.7%

GA12.7%

FL16.1%

OK10.4% AR

9.3%

CO4.9%

NM11.8%

KS8.3%

MO6.1%

TN9.8%

SC17.0

NC14.1%

VA8.8%KY

7.1%

OH8.4%I N

6.7%I L

7.5%

WI6.2%

IA8.2%

MN7.7%

NE8.5%

SD8.2%

ND7.4%

AZ7.6%

UT11.2%

WY5.3%

MT2.5%

ID7.2%

NV8.0%

CA10.1%

OR1.6%

WA11.3%

MI9.0%

WV8.6%

PA7.6%

NY8.4%

ME11.2%

DC7.0%

MD6.3%

DE2.8%

NJ7.7%

CT8.9%

RI8.2%

MA6.6%

VT9.0%

NH5.0%

9.0% to 10.0%

6.0% to 7.0%

over 14.0%

10.0% to 14.0%

7

effects on revenue from income taxes, prop-erty taxes, and license fees. These unmea-sured effects can match or even exceed therevenue changes measured by the model, soeven though the results of this study are sig-nificant in themselves, governmentpolicymakers should be aware that the ef-fects of cross-border shopping are evenmore extensive.

Figures 2 and 3 show how a disparityin effective beer tax rates on two sides of aborder results in cross-border purchases.This correspondence is clear even thoughthe effective tax rates shown in the maps

are partial; that is, they only include beerexcises and state and local sales taxes.

Table 2 presents data on sales taxes andstate beer excise taxes across the nation,and the results of an econometric analysisof the 1997 packaged beer market is pre-sented in Table 3. The table is divided intothree sections. Data in the first column ofthe first section shows interstate cross-bor-der imports as a percent of packaged beersales in a state. The data in the second andthird columns of the table shows the num-ber of cases of beer imported into the stateby cross-border shoppers and the effect

Source: Tax Foundation.

Figure 3Net Interstate Cross-Border Sales As a Percentage of Total

Packaged Beer Sales by State

1997

over +4.0%

+1.0% to +2.0%

–1.0% to +1.0%

TX+0.3%

LA–1.1%

MS+1.7%

AL–6.0%

GA+1.4%

FL0%

OK–0.7% AR

+0.4%

CO+1.6%

NM–3.6%

KS–1.3%

MO+1.8%

TN+1.3%

SC–2.2

NC0%

VA–0.4%KY

+1.0%

OH–0.7%I N

+3.6%I L

–1.6%

WI+2.0%

IA–1.3%

MN0.8%

NE–0.5%

SD–1.3%

ND–0.3%

AZ+3.7%

UT–1.8%

WY+2.1%

MT+2.4%

ID+1.2%

NV+1.8%

CA–1.0%

OR+9.1%

WA–4.9%

MI–0.3%

WV–0.6%

PA–0.8%

NY–0.3%

ME–2.6%

DC–0.3%

MD+2.7%

DE+22.3%

NJ–1.1%

CT–3.0%

RI–3.0%

MA–0.3%

VT–2.5%

NH+16.8%

–2.0% to –1.0%

+2.0% to +4.0%

under –4%

–4.0% to –2.0%

8

that this type of activity had on governmentrevenue generated by the combination ofstate and local sales taxes and state beerexcises.

Data in the other two sections of thetable provide similar data for net interstateexporting states and for states which exportbeer to Canada.

Net Interstate Importing StatesThe first section of Table 3 shows that

slightly more than half of the states werenet importers of packaged beer during1997. Ranging from 6.7 percent in Alabamato a trivial 0.02 percent in Florida. Inter-state cross-border purchases averaged 1.2percent of packaged beer sales in the im-porting states. As a result of this in-boundcross-border traffic, net importing states lostapproximately $34.7 million in sales and

beer excise tax revenue during 1997. Im-portantly, the data presented in this columnis net imports. It is common for states tolose sales on some borders while pickingthem up on others; so a total import figurewould be deceptive. Maryland, for example,loses sales to Delaware but picks them upon each of its other borders.

The state that lost the most sales to itsneighbors in 1997 was Alabama (see Figure4). During that year interstate cross-border

Table 2Sales and State Beer Excise Taxes by State

1997

Taxes StateState Sales Local Sales Beer Excise(Percent) Authorized ($ Per Barrel)

Alabama 4% yes $ 32.65Alaska - - 10.85Arizona 5 yes 4.96Arkansas 4.625 yes 7.508California 6 yes 6.20

Colorado 3 yes 2.48Connecticut 6 no 6.20Delaware - - 4.85Florida 6 yes 19.84Georgia 4 yes 14.88

Hawaii 4 no 28.21Idaho 5 yes 4.65Illinois 6.25 yes 2.17Indiana 5 no 3.565Iowa 5 yes 5.89

Kansas 4.9 yes 5.58Kentucky 6 no 2.50Louisiana 4 yes 10.00Maine 6 no 10.85Maryland 5 no 2.79

Massachusetts 5 no 3.30Michigan 6 no 6.30Minnesota 6.5 no 2.40Mississippi 7 no 13.23Missouri 4.225 yes 1.86

Montana - - 4.30Nebraska 5 yes 7.13Nevada 6.5 yes 2.79New Hampshire - - 9.30New Jersey 6 no 3.72

New Mexico 5 yes 12.71New York 4 yes 4.96North Carolina 4 yes 16.53North Dakota 5 yes 4.96Ohio 5 yes 5.5552

Oklahoma 4.5 yes 11.25Oregon - - 2.60Pennsylvania 6 yes 3.31Rhode Island 7 no 3.00South Carolina 5 yes 23.808

South Dakota 4 yes 8.50Tennessee 6 yes 3.90Texas 6.25 yes 6.00Utah 4.875 yes 11.00Vermont 5 no 8.215

Virginia 3.5 yes 8.76Washington 6.5 yes 7.172West Virginia 6 no 5.50Wisconsin 5 yes 2.00Wyoming 4 yes 0.62

District of Columbia 5.75 n.a. 2.79

Source: Tax Foundation.

Policymakers should beaware that the effects ofcross-border shopping onincome taxes, property taxes,and license fees can matchor even exceed the revenuechanges in state and localsales and excise taxesmeasured by the model.

9

Table 3Cross-border Activity by State

Net Interstate Lost Net Interstate Gained Canadian GainedImports Net Sales & Exports Net Sales & Exports Total Sales &

(Imports/ Imports Excise Tax (Exports/ Exports Excise (Exports/ Exports Excise TaxSales) (cases) Revenue Sales) (cases) Tax Revenue Sales) (cases) Revenue

United States 1.2% 18,148,201 $34,711,318 2.0% 18,148,201 $18,752,943 2.7% 10,884,966 $14,620,473

Alabama 6.7% 2,506,751 9,038,975 ** ** ** *** *** ***Alaska * * * ** ** ** 3.0% 176,369 168,377Arizona * * * 3.8% 1,847,281 2,344,391 *** *** ***Arkansas * * * 0.4% 84,140 129,382 *** *** ***California 1.0% 3,139,317 5,248,466 ** ** ** *** *** ***

Colorado * * * 1.7% 562,416 453,566 *** *** ***Connecticut 3.3% 859,168 1,273,244 ** ** ** *** *** ***Delaware * * * 22.1% 1,668,020 765,405 *** *** ***Florida 0.02% 22,800 60,917 ** ** ** *** *** ***Georgia * * * 1.5% 1,080,145 2,280,090 *** *** ***

Hawaii * * * ** ** ** *** *** ***Idaho * * * 1.4% 129,389 155,518 2.3% 217,771 261,749Illinois 1.5% 1,617,226 2,002,627 ** ** ** *** *** ***Indiana * * * 3.6% 1,935,264 2,152,560 *** *** ***Iowa 1.4% 324,412 442,787 ** ** ** *** *** ***

Kansas 1.3% 280,765 387,697 ** ** ** *** *** ***Kentucky * * * 1.1% 393,789 463,173 *** *** ***Louisiana 1.1% 498,286 996,523 ** ** ** *** *** ***Maine 2.9% 312,970 584,068 ** ** ** 11.7% 1,245,746 2,324,822Maryland * * * 3.0% 1,259,841 1,320,614 *** *** ***

Massachusetts * * * 0.3% 136,925 149,301 *** *** ***Michigan 0.3% 269,631 401,807 ** ** ** 3.3% 3,040,756 4,531,374Minnesota 0.8% 324,765 416,424 ** ** ** 3.0% 1,259,649 1,615,163Mississippi * * * 1.9% 515,371 1,141,444 *** *** ***Missouri * * * 1.9% 1,037,759 1,046,577 *** *** ***

Montana * * * 2.7% 236,308 97,694 17.5% 1,557,456 643,884Nebraska 0.5% 73,190 102,969 ** ** ** *** *** ***Nevada * * * 2.4% 554,176 738,290 *** *** ***New Hampshire * * * 15.6% 1,877,437 1,551,894 1.8% 219,214 181,202New Jersey 1.1% 717,099 913,064 ** ** ** *** *** ***

New Mexico 4.0% 729,770 1,426,586 ** ** ** *** *** ***New York 0.3% 532,102 741,681 ** ** ** 0.6% 941,083 1,311,747North Carolina * * * 0.05% 33,189 77,515 *** *** ***North Dakota 0.3% 18,980 23,299 ** ** ** 6.9% 460,914 565,800Ohio 0.8% 733,894 1,030,714 ** ** ** *** *** ***

Oklahoma 0.8% 222,005 382,036 ** ** ** *** *** ***Oregon * * * 9.3% 2,389,970 652,314 *** *** ***Pennsylvania 0.8% 721,090 908,080 ** ** ** *** *** ***Rhode Island 2.8% 258,628 354,174 ** ** ** *** *** ***South Carolina 2.5% 896,378 2,526,972 ** ** ** *** *** ***

South Dakota 1.8% 118,627 162,302 ** ** ** *** *** ***Tennessee * * * 1.6% 733,006 1,195,567 *** *** ***Texas * * * 0.3% 657,871 1,005,261 *** *** ***Utah 1.9% 247,602 460,445 ** ** ** *** *** ***Vermont 3.1% 173,097 259,047 ** ** ** 14.0% 769,334 1,151,340

Virginia 0.6% 400,347 586,757 ** ** ** *** *** ***Washington 5.2% 2,059,599 3,853,998 ** ** ** 2.5% 996,674 1,865,014West Virginia 0.5% 81,758 116,432 ** ** ** *** *** ***Wisconsin * * * 2.0% 906,859 936,991 *** *** ***Wyoming * * * 2.8% 109,045 95,394 *** *** ***

District of Columbia 0.2% 7,942 9,230 ** ** ** *** *** ***

* Net Exporter of Beer** Net Importer of Beer*** Not Applicable

10

purchases were equal to 6.7 percent ofpackaged beer sales in the state. This cross-border shopping was driven by the stategovernment’s 4.0 percent sales tax, addi-tional sales taxes levied by local govern-ments, and its highest-in-the-nation $32.65per barrel beer excise tax. This combina-tion of tax rates resulted in Alabama’s los-ing sales to all of the surrounding states.While most lost sales were the result ofcross-border shopping in Tennessee (33.5percent) and Mississippi (32.3 percent), sig-nificant amounts were also lost along itsborders with Georgia (24.8 percent) andFlorida (9.4 percent). In total, Alabama losta whopping $9.0 million in sales tax collec-tions and state beer excises as a result ofthis cross-border traffic.

Washington also lost a significantamount of packaged beer sales to neighbor-ing states in 1997. During that year anamount equal to 5.2 percent of packagedbeer sales in the state was purchased instates along its borders. That’s more than2.0 million cases of beer that came acrossborders into Washington due to the state’s6.5 percent sales tax, additional sales taxeslevied by local governments, and $7.172 perbarrel state beer excise. More than threequarters of the $3.9 million that Washing-ton lost in sales tax and beer excise revenuecan be attributed to cross-border shoppingin Oregon. Table 3 shows that Oregon hasno state or local sales taxes and a modest$2.60 per barrel state beer excise.

The state with the third highest percent-age of lost package sales during 1997 wasNew Mexico where out-of-state vendorssupplied an amount equal to 4.0 percent ofNew Mexico’s packaged beer sales to cross-border shoppers. The motivating factorswere New Mexico’s 5.0 percent state salestax, additional sales taxes levied at the lo-cal level, and the $12.71 per barrel statebeer excise. As was the case with both Ala-bama and Washington, these relatively hightax rates resulted in New Mexico’s losingsales to all neighboring states. While mostof these lost sales occurred along the state’sborder with Colorado (36.1 percent), largeshares of the total were also lost along itsborders with Arizona (32.9 percent) andTexas (29.0 percent). Even along the state’sshort border with Oklahoma the state lostsales. In total, New Mexico’s general salesand beer tax collections were more than$1.4 million lower in 1997 due to cross-bor-der shopping.

Net Interstate Exporting StatesThe second section of Table 3 presents

data on the share of each state’s packagedbeer sales that was purchased by cross-bor-der shoppers from other states in 1997.Slightly less than half of states were net ex-porters, and in-bound customers accountedfor an average of 2.0 percent of total salesin those states. As a result of the incomingSource: Tax Foundation.

Figure 4Effective Tax Rates on PackagedBeer and the Resulting Cross-Border Purchases: Alabama

and Neighboring States

Thousands of Cases of Beer

1997

GA12.7%

AL21.7%

MS13.3%

TN9.8%

809.7

621.7

235.6

FL16.1%

839.8

11

business, the beer exporting states collected$18.8 million in sales taxes and beer excisesfrom the citizens of neighboring states, ineffect “exporting” their tax burdens. Asstated earlier, sales and excise taxes are theonly types of state and local revenue mea-sured by this study. For beer exportingstates, the migration of business activityacross state lines into their states also leadsto higher revenue from income and prop-erty taxes, license fees, and other businessrelated revenue sources.

The highest rate of net beer export was22.1 percent of sales in Delaware (see Fig-ure 2), where vendors situated along the bor-ders effectively marketed the state’s lack ofstate or local sales taxes and its moderate$4.85 per barrel state beer excise. During

1997 cross-border shoppers bought approxi-mately 1.7 million cases of beer in Delaware.Pennsylvanians accounted for 64.5 percentof these sales while residents of New Jerseyand Maryland accounted for 19.3 and 16.2percent.

Residents of New Hampshire’s neigh-boring states crossed into the state and con-stituted a very significant fraction of itspackaged beer sales during 1997. Approxi-mately 15.6 percent of the packaged beersold in New Hampshire, or approximately1.9 million cases, was purchased by cross-border shoppers from Massachusetts, Maineand Vermont. As with Delaware, NewHampshire’s lack of state or local sales taxescoupled with a moderate beer excise, $9.30per barrel, provided the attraction to cross-border shoppers. Residents of Massachu-setts bought 69.0 percent of NewHampshire’s interstate beer exports, andcross-border sales to residents of Maine andVermont accounted for 21.3 and 9.7 per-cent. As a result, New Hampshire collected$1.6 million in state beer excise taxes fromneighboring states.

The state with the third highest percent-age of net beer exports was Oregon, whichmanaged to sell 9.3 percent of its packagedbeer to cross-border shoppers. Like Dela-ware and New Hampshire, Oregon has nostate or local sales taxes. Combining this in-centive with a low $2.60 per barrel beerexcise, Oregon was able to lure cross-bor-der shoppers who bought 2.4 million casesof beer. Most of it (60.8 percent) was soldto residents of Washington. Nevadans ac-counted for 20.6 percent, and the balancewas purchased by residents of Idaho (11.7percent) and California (6.9 percent). As aresult, more than half a million dollars inbeer excise taxes came into Oregon govern-ment coffers from outside the state.

Canadian Cross-Border TrafficThe third section of Table 3 presents

data on international cross-border shoppingalong the U.S.–Canadian border. The analy-sis shows that all of the states along theborder were net beer exporters to Canada.Source: Tax Foundation.

Figure 5Effective Tax Rates on PackagedBeer and the Resulting Cross-Border Purchases: Delaware

and Neighboring States

Thousands of Cases of Beer

1997

DE2.8%

NJ7.7%

MD6.4%

PA7.6%

321.9

1,075.9

270.2

12

Rates of net export ranged from 17.5 per-cent of packaged beer sales in Montana to0.6 percent of sales in New York. On aver-age, Canadian cross-border purchases aver-aged 2.7 percent of sales in states along theborder. The border states sold 10.9 million

cases of beer to Canadian shoppers in 1997,collecting $14.6 million in sales taxes andbeer excise taxes as a result.

Among the states, Montana exported thehighest percentage of its packaged beersales to Canada in 1997—approximately17.5 percent, netting the state more thanhalf a million dollars in beer excise tax col-lections. Adding together its exports toCanada and to its neighboring states, Mon-

tana sold more than 20 percent of its pack-aged beer to cross-border shoppers in 1997.

Vermont also exported a significantamount of beer to Canada in 1997. Duringthat year the state sold 769,334 cases ofbeer, or 14.0 percent of total sales, to cross-border shoppers from Canada, collecting$1.2 million in sales taxes and beer excisetaxes from them.

The state that exported the third high-est percentage of its packaged beer sales toCanada in 1997 was Maine. Approximately11.7 percent of its sales were to Canadianshoppers who paid more than $2.3 millioninto Maine’s government coffers in state andlocal sales and state beer excise taxes.

Some states with high effective tax rateson beer were able to make up for the salesthey lost to neighboring states by exportinglarge quantities of beer to Canada. Figure 5illustrates this shifting of tax burdens fromjurisdiction to jurisdiction by showing theflow of beer into and out of Washington.

Source: Tax Foundation.

Figure 6Effective Tax Rates on Packaged Beer and

Resulting Cross-Border Activity: Washington andNeighboring Jurisdictions

Millions of Cases of Beer

1997

OR1.6%

WA11.3%

ID7.2%

0.2

1.4

1.2

13

A Model of Demandfor Packaged BeerAnd Its Supply bySource

The econometric model developed inthis section builds on several previous stud-ies of cross-border activity. The most sig-nificant of these are the 1977 and 1985 Ad-visory Commission on IntergovernmentalRelations (ACIR) studies of cross-border ac-tivity. A 1993 Tax Foundation analysis con-ducted by the Economic Policy Group ofKPMG Peat Marwick built on these twostudies. Additional Tax Foundation analysesconducted in 1996 and 1998 further devel-oped the methodology used in the 1993study.

All prior Tax Foundation analyses ofcross-border activity have dealt exclusivelywith cigarettes. This research generated agreat deal of interest on the part ofpolicymakers, the media, and the public. Inresponse, the Tax Foundation has decidedto conduct additional analyses by applyingthe methodology that it has developed toother goods which are subject to sales andproduct-specific excise taxes.

Developing the ModelEconomic theory suggests that the de-

mand for a good is a function of its price,consumers’ income levels, and their tastes.Equation 1 is an algebraic representation ofthe demand for beer by residents of differ-ent states.

Equation 1

ci = a – ppi + iyi + [tastes]

Beer consumption by residents of astate is given by ci. The price of beer is givenby pi. The income level of state residents isgiven by yi. The effect of consumer tastesis temporarily represented by [tastes].

The beer consumed in a state is sup-

plied by sales in that state plus any importsfrom other areas. This relationship is givenby Equation 2.

Equation 2

ci = si + NXi

Domestic sales are given by si. There arethree potential sources of net exports NXi:(1) sales that result when out-of-state resi-dents living close to a state border inten-tionally cross the border to purchase beer;(2) sales that result when residents ofCanada intentionally cross the border topurchase beer in border states; and (3) inci-dental package beer purchases by nonresi-dents when in a state on personal or busi-ness travel. The components of net exportsare listed in Equation 3:

Equation 3

NXi = gS[( ti – tj)wij] + kkiS[( ti – tk)wik] +[incidental nonresident purchases]

Tax differentials between states provideindividuals living near state borders with theincentive to shop across the border. Saleswhich result from cross-border shoppingare measured by S[( ti – tj)wij]. The direc-tion and magnitude of this incentive is givenby the tax differential (ti – tj) where ti is theweighted average effective tax rate on pack-aged beer along the border created by stateand local sales and state beer excises in statei and tj is the weighted effective tax rate cre-ated by these taxes along the border in theneighboring state j. The number of individu-als with the ability to avail themselves of theopportunity to cross-border shop and theireffect on sales in a particular state is givenby wij, where:

Equation 4

wij =population of i on ij border

population of i

if i imports from j, or

14

wij =population of j on ij border

population of i

if i exports to j.By summing the variable (ti – tj)wij for

all of the states which surround a state, it ispossible to determine whether the state willexperience a net increase or decrease inpackaged beer sales as a result of cross-bor-der shopping.

A state’s sales are also likely to be af-fected if it shares a border with Canada.Historically, beer prices have been muchhigher in Canada than in the United States.This price differential, due in large part tohigh Canadian sales and excise taxes, hasresulted in large scale cross-border shop-ping. In order to capture this effect the vari-able kiS[( ti – tk)wik] was created. The firstcomponent of this variable, ki, is a binaryvariable which assigns a 1 to all states lo-cated on the U.S.-Canadian border and a 0to those that are not. The second compo-nent, (ti – tk), measures the direction andmagnitude of the incentive to cross-bordershop. It is similar to the (ti – tj) componentdescribed above. The final component, wik,is similar to the wij variable also describedabove. It measures the effect that Canadi-ans living within 50 miles of the U.S.-Cana-dian border will have on beer sales in bor-der states. Sales that result from incidentalpurchases by nonresidents are temporarilygiven by [incidental nonresident pur-chases].

Substituting Equation 3 into Equation 2yields:

Equation 5

ci = si – gS[( ti – tj)wij] –kkiS[( ti – tk)wik] +[incidental nonresident purchases]

Substituting Equation 7 into Equation 1and rearranging terms so that taxable sales,si, is alone on the left hand side of the equa-tion yields:

Equation 6

si = a – ppi + iyi + [tastes] –gS[( ti – tj)wij] – kkiS[( ti – tk)wik] +[incidental nonresident purchases]

Before Equation 6 can be estimated us-ing empirical data it requires several modi-fications. First, federal data shows that boththe fraction of the population who consumealcohol and average consumption levelsvary widely by state. In order to control forthese factors the following modificationswere made to the model. First, the depen-dent variable was created by dividing totaldomestic packaged sales in a state by thenumber of individuals who consume alco-hol. Then several independent variableswere added to control for differences intastes among consumers.

The most significant of these was anaverage consumption variable. The U.S.Department of Agriculture tracks averagebeer consumption by age and region. TheCensus Bureau produces an annual seriesdetailing the composition of state popula-tion by age. By using these two series it waspossible to compute average consumptionlevels by state. However, since the data pro-vided by the USDA is by region and not stateit can be augmented somewhat. Data fromthe U.S. Department of Health and HumanServices shows that alcohol consumptionis positively correlated with education level.In order to control this factor an educationvariable ei was added to Equation 6. It mea-sures the percentage of the states’ popula-tions who have a bachelors or higher de-gree. Similarly, data from the HHS showsthat alcohol consumption is higher amongmales than females and that it is more preva-lent among the young. In order to controlfor these factors the variable mi was created.It measures the percent of a state’s popula-tion that is comprised of males between 21and 40 years of age.

Finally, alcohol consumption varies byregion. In order to control for these differ-ences, binary intercept variables were addedto the model. The country was divided up

15

into four regions. The first included thestates in the New England, Mideast, andGreat Lakes regions of the country as de-fined by the U.S. Bureau of Economic Analy-sis. The second region includes states in thePlains, Rocky Mountain, and Far West re-gions. States in the Southwest region com-prised the third region while those in theSouthwest made up the forth region. Alaskaand Hawaii were not included in the sec-ond region and were assigned their ownindividual binary variables.

Draught beer is a substitute for pack-aged beer. In order to control for differ-ences in draught beer consumption by statethe variable di was added to the model. Itmeasures draught beer sales as a percent-age of total beer sales by state.

Data on incidental nonresident packagebeer purchases does not exist. However, theU.S. Bureau of Economic Analysis producesdata on the composition of gross state prod-uct by state. The fraction of a state’seconomy that was comprised of hotel,amusement, and recreational services, ni,was used as a proxy for this variable. Equa-tion 7 therefore represents an estimableform of Equation 6.

Equation 7

si = a + aR2 + aR3 + aR4 + aAK + aHI – ppi

+ iyi + cci + hei+ mmi – ddi –gS[( ti – tj)wij] –kkiS[( ti – tk)wik] + nni

The signs of the coefficients of the vari-ables in Equation 7 represent ex ante ex-pectations as to the effect that each willhave on packaged beer sales in a state.There is assumed to be an inverse relation-ship between the price of a good and thequantity demanded by consumers. As a re-sult, the sign of the coefficient p is expectedto be negative. Similarly, since packagedbeer is assumed to be a normal good whoseconsumption rises with income, the sign ofi is expected to be positive.

As discussed earlier, the variables ci, ei

and mi were included in Equation 7 to rep-resent differences in the tastes of consum-ers in the states. Federal data suggests thatall should be positively correlated with beerconsumption in a state. Therefore, the co-efficients of all of these variables are ex-pected to have a positive sign.

Draught beer is a substitute for pack-aged beer. Therefore, the coefficient, d, isexpected to be negative.

If the effective tax rate on beer in statei is higher than it is in neighboring jurisdic-tions it will encourage cross-border shop-ping. Since beer purchased in neighboringjurisdictions replaces packaged sales in thatstate the signs of g and k are expected tobe negative.

The variable ni was included in Equa-tion 7 as a proxy for incidental nonresidentpackaged beer consumption. Since pack-aged beer sales in a state should rise withinfluxes of nonresidents the sign of n is ex-pected to be positive.

EstimationEquation 7 was estimated using ordinary

least squares regression analysis. This typeof analysis is very powerful in that it allowsa researcher to identify various factors (for-mally called independent variables) that af-fect a so-called dependent variable and thenestimate the effects that each has on thatvariable while holding the effects of theother variables constant. In Equation 7, perdrinker domestic packaged beer sales (mea-sured in gallons) is the dependent variable.All of the variables listed on the right handside of the equation are independent vari-ables. The data used to estimate this equa-tion was collected for all fifty states and theDistrict of Columbia for the eight year pe-riod from 1990 though 1997.

ResultsTable 4 presents the results of the re-

gression analysis. The first column of thetable lists the independent variables used

16

in the regression. These have been segre-gated by the types of effects they were in-tended to capture. The second column liststhe estimated coefficients of these variables.The coefficients measure the direction andmagnitude of change in the dependent vari-able for each one unit change in an inde-pendent variable while holding the effectsof the other independent variables constant.The t-statistics, listed in column 3, measurethe degree of confidence in each estimatedcoefficient. If the t-statistic is greater than

1.96 in absolute value, it indicates a highlevel of confidence in the estimated coeffi-cient. All of the variables estimated in thisanalysis were statistically significant exceptthose for the price and income variables.However, both the magnitudes and signs ofthese coefficients were as expected. TheR2 statistic at the bottom of Table 4 mea-sures the amount of variation in the depen-dent variable that is explained by the over-all model. In this particular case, the modelexplains 86.0 percent of the variation in perdrinker packaged beer sales among states.

It was hypothesized above that a rise inthe price of beer would lead to a decline insales. The sign of the estimated price coef-ficient p supports this notion. Similarly, itwas hypothesized that beer was a normalgood whose consumption rose with in-come. The sign of the estimated incomecoefficient i supports this notion.

The variables ci, ei and mi were includedin the model to control for differences intastes among consumers in the states. It washypothesized that all of these variableswould be positively correlated with pack-aged beer sales. The positive signs of eachof the estimated coefficients of these vari-ables supports this notion.

It was hypothesized that draught salesreplace packaged beer sales in states. Thesign of the coefficient d supports this notion.

Two variables were included in themodel to capture the effects of interstateand Canadian cross-border shopping. Thefirst of these, S[(ti – tj)wij], was included tocapture the effect of interstate cross-bordershopping. It was hypothesized that stateswith relatively high effective tax rates onpackaged beer would lose sales to them.The negative sign of the estimated coeffi-cient of the cross-border shopping variableg supports this notion.

The model also included variablekkiS[(ti – tj )wik] to estimate the extent ofCanadian cross-border shopping in borderstates. The positive sign of the k supportsthe notion that high effective tax rates onbeer in Canada has encouraged cross-bordershopping by Canadians in the United States.

Table 4Regression Results

Dependent Variable: Per Drinker Packaged BeerSales

Measured in Gallons

Variable Coefficient t-statistic

Intercepts

Intercept (a) –11.1 –1.4

Region 2 (aR2) 2.4 3.1

Region 3 (aR3) 6.3 5.9

Region 4 (aR4) 11.3 10.3

Alaska (aAK) 6.7 3.5

Hawaii (aHI) 8.6 4.6

Price and Income

Price (p) –0.01010 –0.5

Income (i) 0.00003 0.3

Demographic

Consume (c) 0.2 19.3

Education (e) 69.1 8.2

Male2140 (m) 55.9 2.1

Draught (h) –25.8 –3.5

Cross-Border Shopping

Interstate (g) –88.3 –7.4

Canada (k) –119.0 –5.4

Other

Incidental Nonresident

Consumption 111.2 12.0

R2 86.0

Source: Tax Foundation.

17

It was also hypothesized above thatstates which entertained large numbers ofnonresidents would have relatively highpackaged beer sales, all else being equal.The positive sign of the estimated coeffi-cient of the tourist variable, q, supports thisnotion.

ConclusionThe per capita sale of packaged beer

varies widely by state. It has long been sus-pected that these differences are due in partto cross-border shopping. Building on ear-lier work in this area, this study sought toexplain differences in packaged beer salesamong the states. A model of demand forbeer and its supply by source was con-structed. This model was created in a man-ner that allowed it to capture the effects ofboth interstate and Canadian cross-bordershopping on beer sales in the states.

The model was then tested empiricallyusing data from 1990–1997. Cross-bordershopping was found to have significant ef-fects on packaged beer sales in the states.In particular, the study found that in 1997,18.1 million cases of beer, on net, movedfrom low- to high-tax states. Such exportsaccounted for approximately 2.0 percent ofsales in net exporting states and allowedthem to export $18.8 million in sales andbeer excise taxes to their high-tax neigh-bors. In addition, states along the U.S.-Ca-nadian border were able to export 10.9 mil-lion cases of beer and $14.6 million in salesand beer excise taxes to Canada. The studyclearly shows that high sales and excise taxdifferentials lead to significant increases incross-border beer sales.

18

Notes1. Figures given are weighted average

prices of all packaged sales.2. The term ‘packaged beer’ is some-

times used to describe beer sold for off-pre-mises consumption. Here it refers to allcontainer, or non-draught, beer.

ReferencesAdvisory Commission on Intergovern-

mental Relations. Cigarette Bootlegging: AState and Federal Responsibility. Washing-ton, DC: U.S. Government Printing Office,1977.

———. Cigarette Tax Evasion: A SecondLook. Washington, D.C.: US GovernmentPrinting Office, 1985.

Baltagi, Badi H. and Dan Levin. “Estimat-ing Dynamic Demand for Cigarettes UsingPanel Data: The Effects of Bootlegging, Taxa-tion and Advertising Reconsidered.” The Re-view of Economics and Statistics 68:148-55.

Beer Institute. Brewers Almanac. Wash-ington, DC. Various issues.

Coats, R. Morris. “A Note on EstimatingCross-Border Effects of State CigaretteTaxes.” National Tax Journal 68:573-84.

Felgner, Brent H. “An Open Road ForSmuggled Smokes?” U.S. Distribution Jour-nal. April, 1988. pp. 26-32.

Fleenor, Patrick W. How Excise Tax Dif-ferentials Affect Interstate Smuggling andCross-border Sales of Cigarettes in theUnited States. Background Paper No. 26.Tax Foundation. October 1998.

Fleenor, Patrick W. The Effect of ExciseTax Differentials on the Interstate Smug-gling and Cross-Border Sales of Cigarettesin the United States. Background Paper No.16. Tax Foundation. October 1996.

Fujii, Edwin T. “The Demand for Ciga-rettes: Further Empirical and Its Implica-tions for Public Policy.” Applied Economics12:479-89.

Hamilton, James L. “The Demand forCigarettes: Advertising, the Health Scare, andthe Cigarette Advertising Ban.” The Reviewof Economics and Statistics 54:401-11.

Lyon, Herbert L. and Julian L. Simon.“Price Elasticity of Demand for Cigarettesin the United States.” American Journal ofAgricultural Economics 50:881-893.

Rand-McNally. 1994 Road Atlas: UnitedStates, Canada, and Mexico. Skokie, IL:Rand-McNally Publishing Group, 1994.

19

Tax Foundation. Effects of Cross-BorderSales on Economic Activity and State Rev-enues: A Case Study of Tobacco ExciseTaxes in Massachusetts, New York City, andSurrounding Areas. January 1993.

U.S. Department of Agriculture, Agricul-tural Research Service. Food and NutrientIntakes by Individuals in the United Statesby Region. Washington, D.C., U.S. Govern-ment Printing Office. Various issues.

U.S. Department of Health and HumanServices, Substance Abuse and MentalHealth Services Administration. NationalHousehold Survey of Drug Abuse. Washing-ton, D.C., U.S. Government Printing Office.Various issues.

Warner, Kenneth E. “Cigarette ExciseTaxation and Interstate Smuggling: An As-sessment of Recent Activity.” National TaxJournal 35:483-490.

Wasserman, Jeffery, Willard G. Manning,Joseph P. Newhouse, and John D. Winkler.“The Effects of Excise Taxes and Regulationson Cigarette Smoking.” Journal of HealthEconomics 10:43-64.