Revised June 19, 2003

EXPANDING SALES TAXATION OF SERVICES: OPTIONS AND ISSUES
By Michael Mazerov

Summary

PDF of summary
PDF of full report

View Related Analyses

If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.

Sales of tangible goods dominate most states’ sales tax bases; only a few states impose their sales taxes on a broad array of services.  Sales of “non-durable goods” like clothing and light bulbs and “durable goods” like cars and computers generate the vast majority of state sales tax receipts.  According to the Federation of Tax Administrators, a majority of states apply their sales tax to less than one-third of 164 potentially-taxable service categories.  Eight of the 45 states with sales taxes impose them on fewer than 20 service categories.[1]

Most states could improve their sales taxes and their tax systems in general with some expansion of the tax base to include services.  Adding services could make tax systems fairer, more stable, more economically neutral, and easier to administer.  Perhaps most importantly in the current fiscal crisis, adding services to a state’s sales tax base could help close fiscal gaps and avert cuts in important services or other, less sound, tax increases.  There are significant economic, administrative, and legal issues that must be addressed in expanding sales taxation of services.  The barriers are not insurmountable, however, and the benefits from a broader sales tax base outweigh the challenges.

Why Tax Services?

Public finance economists and other tax experts have been urging states for decades to include more services in the sales tax base.[2]  Taxing services meets all of the criteria by which state tax policy options are normally evaluated.

  • Taxing additional services can generate substantial new sales tax revenue.  Table 1 indicates that the annual, nationwide revenue yield from taxing all services purchased by households except health care, education, housing, and a few others would be on the order of $57 billion.  The new revenue from taxing household services would be less than this, since most states do tax services to some extent.  Table 1 suggests, however, that states that do not tax services to any significant degree at present — such as California, Illinois, Michigan, and Virginia — probably could increase their sales tax revenue by 25 to 30 percent if they taxed services purchased by households comprehensively.

Some states have prepared their own estimates of the revenue that could be gained by expanding the sales taxation of services.  Texas estimates that broadening its sales tax to include just five types of services could yield on the order of $600 million in additional sales tax receipts annually.  Michigan estimates that its failure to tax services costs the state $4.8 billion each year.[3]

  • Taxing services broadly is essential if the long-run revenue adequacy of the sales tax is to be maintained.  Household spending has been shifting from goods to services for decades.  The traditional sales tax base, purchases of durable goods plus non-durable goods except groceries (which the majority of states exempt), fell from 39 percent of household consumption in 1970 to 33 percent in 2001.  Over the same interval, consumption of services rose from 31 percent to 44 percent of total household purchases.[4]  Largely to compensate for this trend, states have increased sales tax rates sharply over the last 30 years.  The ability of states to continue raising rates is constrained, however, by such factors as the ease with which consumers can shift their purchases to the Internet — where sales taxes often are not charged.  Assuming that consumption will continue to shift toward services, including services in the tax base therefore will be essential to maintaining sales tax revenues over the long term.
  • Bringing services into the sales tax base may reduce the year-to-year volatility of sales tax collections.  Sales tax bases are dominated by purchases of “big-ticket” durable goods (such as cars, appliances, and furniture), which often decline sharply during economic downturns.  Limited research finds that purchases of some services do not fall as precipitously as durable goods purchases do when the economy slows nor rise as rapidly when the economy is booming.  The research suggests that including more services in the sales tax base could moderate slightly the volatility of sales tax revenues over the course of the business cycle.
  • Expanding the taxation of services will make the sales tax fairer.  The sales tax is intended to be a general tax on consumption.  There is little reason to distinguish between consumption of goods and consumption of services, which in fact can be substitutes for one another.  For example, it is not equitable — it violates the principle of “horizontal equity” — to tax the person who rents a videotape but not the person who watches a pay-per-view movie on cable TV.

Moreover, sales taxes are regressive; that is, they absorb a larger proportion of the income of lower-income taxpayers than of higher-income taxpayers.  In part, this is because higher-income persons do not consume their entire incomes; the portion of their incomes that they save is not subject to sales or other consumption taxes.  Ideally, one could consider reducing the regressivity of sales taxes by expanding the sales tax to services purchased primarily by the affluent.  While worth doing, the revenue gain from expanding solely to services the affluent buy is not substantial.  To broadly tax services, one must include services purchased by low-, moderate- and middle-income households as well as those purchased by high-income people.  In general, a broad expansion of sales taxation to services will not change relative tax burdens.[5]

There is one circumstance, however, in which expansion of the sales tax to services could make a state’s overall tax system somewhat less equitable.  That can occur if taxing services raises the proportion of total state revenue that is derived from the sales tax and thereby lowers the proportion that is derived from more progressive sources such as personal and corporate income taxes.  There are a number of ways this potential to increase the regressivity of the tax system as a whole can be avoided.   The sales tax base expansion can be balanced with other changes in the tax code, such as an increase in the personal income tax.  Alternatively, targeted credits administered through the income tax or rebates of sales taxes paid can be used to mitigate the increased sales tax burden low-income families could experience when a sales tax is broadly expanded to include services.

  • Taxing services can improve the allocation of economic resources.  The failure to tax services while most goods are taxed subtly distorts resource allocation throughout the economy by creating an artificial incentive to purchase services rather than goods.  For example, some consumers may be encouraged to repair older cars and appliances rather than replace them with more energy-efficient and less polluting alternatives — although the effect is probably modest.  The failure to tax services has also contributed to steady increases in sales tax rates, which create their own economic distortions.  Unnecessarily high tax rates on goods resulting from under-taxation of services stimulate such wasteful activities as tax-motivated interstate shopping.
  • Expanding the taxation of services can simplify the process of administering and complying with the sales tax.  Expanding the taxation of services can reduce the effort and costs entailed in enforcing and complying with the sales tax.  When retailers sell taxable goods and services as well as tax-exempt services, it can be difficult and costly for tax administrators and merchants alike to ensure that the proper amount of tax has been collected and remitted.  If all of a retailer’s sales are subject to tax, many accounting burdens and disputes diminish or disappear.

Expanded taxation of services can contribute to the realization of all of these arguably desirable tax policy objectives.  There are tradeoffs involved, however.  For example, moving toward comprehensive taxation of services could strain the administrative capabilities of state tax departments if done too quickly, since many businesses that previously did not collect sales tax would need to be brought into the sales tax system.  Thus, achieving an optimal balancing of these goals requires an understanding of all of the potential effects of expanding taxation of services and careful choices about which services to tax.

Which Services Could States Tax?  Which Services Should They Tax?

As policymakers contemplate which currently-exempt services to begin taxing, it can be useful to think of services as falling into three categories:

  • services primarily purchased by businesses, such as payroll processing and television advertising,
  • services primarily purchased by households, such as diaper service and cable TV, and
  • services frequently purchased by both households and businesses, such as landscaping and pest control.

Economists generally counsel states to forego taxing the first category of services, so-called “business-to-business” sales.  They point out that taxing the goods and services businesses buy to use as inputs into the production of other goods and services often leads to “tax pyramiding.”  Tax pyramiding refers to the situation in which an input is taxed when purchased and then effectively taxed again when its cost is passed through into the price of a taxable good or service into which it has been incorporated.  Tax pyramiding results in the actual sales tax imposed on a particular good or service bought by a household being higher than what is added at the cash register.  Because the sales taxes imposed on inputs are hidden in the selling price of the item, states may tend to rely on this revenue source more than they otherwise would.  Moreover, some research suggests that the hidden sales taxes are even more burdensome for low-income families than the visible sales tax that is imposed on the final sale, because necessities like food and utilities that often are tax-exempt nonetheless can have substantial sales taxes hidden in their prices.

Taxation of business inputs also tends to complicate sales tax administration.  For example, rules need to be developed for taxing services like accounting that are purchased by businesses for company-wide use in multiple states.

The greatest concern of economists regarding sales taxation of services purchased by businesses is that it can distort the allocation of economic resources.  Since services provided to an employer by an employee are rarely subject to sales tax, taxation of business-to-business sales of services can encourage businesses to provide services using their own employees even if they could be produced more efficiently by an independent firm.  In addition, if purchases of services subject to sales tax are major cost items for a business (for example, electricity and telecommunications services for an Internet service provider), a more efficient business that tries to pass those taxes into its prices could lose business to a less efficient competitor located in another state that exempts those inputs from sales taxation.  Alternatively, a business that makes substantial purchases of taxable services might choose to expand in a state that is sub-optimal from an economic efficiency standpoint but that exempts those services from sales tax.

While these arguments against taxing business purchases of services have merit, there are at least two countervailing considerations:

  • State sales taxes already apply to numerous purchases of goods by businesses.  Assuming that the concerns of economists about the distorting effect on resource allocation of taxing business inputs are valid, economic theory implies that the distortion grows as the tax rate increases.  If the choice is between increasing the tax rate at which business-to-business sales of goods are taxed and taxing some business-to-business sales of services in order to hold down the tax rate, the latter could be preferable from an efficient allocation of resources standpoint.
  • In an economy in which a growing number of people run their own businesses, exempting all purchases of goods and services by businesses would open the door to substantial tax evasion.  Business owners could claim purchases of many services — such as telecommunications, hotel rentals, and auto and computer repair — to be for business use when they were actually for personal use.  Preventing this abuse would require that substantial additional resources for tax enforcement be provided to state tax departments.  The costs of preventing tax evasion could exceed the economic benefits of exempting business inputs from taxation.

In broadening their states’ taxation of services, policymakers generally have struck what arguably is a reasonable balance among the resource allocation issues raised by economists, their states’ revenue needs, and practical tax enforcement considerations.  States largely have avoided taxing services purchased almost exclusively by businesses (like advertising and accounting); instead, to the extent they have taxed services, states have targeted household services (like haircuts) or mixed household/business services (like landscaping).  Where a particular industry has made a credible case that taxation of a service in the latter category has an adverse economic impact (for example, telecommunications purchased by financial institutions), elected officials also have been willing to enact industry-specific exemptions.  Such an approach may make more sense than an across-the-board exemption for all services that happen to be purchased by businesses as well as households.

Even if states forego taxing services that are predominantly purchased by businesses, there is a vast array of household services to which the sales tax can be applied.  Appendix I (see PDF of summary) of this report lists over 200 types of services purchased by households or by both households and businesses, organized into 20 broad categories including personal care, home cleaning and maintenance, recreation and travel, and lawn and garden.  As entrepreneurs perceive new profit-making opportunities, new services will continue to be invented.  States can either implement taxation of services in a way that will incorporate newly emerging services or can monitor the evolution of the service sector and update their tax policies accordingly.

The Legal Mechanics of Broadening the Sales Taxation of Services

Legislators can expand the taxation of services in two different ways.  The comprehensive approach is to apply to services the typical language used to tax goods.  Under most state sales tax laws, all sales of goods are taxable unless they are explicitly identified as exempt.  Hawaii, New Mexico, and South Dakota apply this same treatment to sales of services.  Hawaii and New Mexico adopted this approach from the inception of their sales taxes; South Dakota did so to expand the taxation of services well after the sales tax had been enacted.

The remaining states that tax services do so by specifically enumerating taxable services.  The enumeration often can be found in the definition of a “retail sale.”

Each approach to taxing services has its own advantages and disadvantages.  A key advantage of the comprehensive approach, for example, is that newly-developed services are immediately taxable without legislative action.  This is appropriate given the role of the sales tax as a general tax on consumption.  It also ensures that the revenue yield of the sales tax is maintained as new services displace old ones (for example, as Internet e-mail service substitutes for faxing as a means of transmitting documents).

The major disadvantage of the comprehensive approach to expanding the sales taxation of services is that it is likely to bring a large number of services into the sales tax base in one fell swoop.  The services subjected to taxation are likely to include many business-to-business services that policymakers might not wish to tax because of the potentially adverse economic effects discussed previously.  Moreover, state revenue departments may not be equipped to integrate numerous new services and the merchants selling them into their sales tax administration systems in a short period of time.  These factors likely explain why of all the states that have expanded their taxation of services over the years, only South Dakota did so comprehensively.

The greatest challenge facing legislators who choose to extend their state’s sales tax to specifically-enumerated services is defining unambiguously the services they intend to tax.  Many services are technologically complex and industry-specific, and legislators and their staffs cannot be expected to be business experts.  Clear definitions are essential, however, because providers of a newly-taxed service often will look for every legal opportunity to avoid having to add a 4-10 percent sales tax to their prices.  In a current case, for example, America Online is disputing Tennessee’s claim that its Internet access service falls under the state’s definition of taxable telecommunications.

A question often arises as to whether it is preferable to write the law to identify taxable services in broad terms (such as “fees for participant sports”) or specifically (such as “admissions, equipment rental, and other fees for bowling, batting cages, skiing. . .”).  The answer is that states would be well-advised to do both.  (“Taxable sales include admission, equipment rental, and other fees for participant sports, which include but are not limited to bowling, batting cages, skiing . . .”)  Broad definitions can serve as a good “backstop” for more specific listings that may inadvertently omit a particular service.  A broad definition can also provide a basis for taxing a newly-invented service until such time as the legislature has an opportunity to identify it explicitly.  Nonetheless, broad definitions are not sufficient to ensure taxability in the face of a taxpayer determined not to charge sales tax.

One promising approach to enumerating taxable services is to piggyback on standardized lists and definitions developed for other purposes.  The North American Industry Classification System (successor to the well-known SIC system) and the North American Product Classification System (under development) could be referenced for sales taxation purposes.[6]  South Dakota’s sales tax statute already references SIC definitions, for example.  In addition to taking advantage of careful definitional work, using NAICs or NAPCS to establish the state sales tax base could make it easier to gauge the revenue impact of taxing services; state-by-state data on the dollar volume of sales of services are already categorized by NAICS and will be further categorized by NAPCS when it is completed.

Table 1
 Estimated Sales Tax Revenue Yield from Taxing All “Readily-Taxable”* Services
($ millions)

 

Share of 2001 Ntl. Personal Income

Estimated

Purchases of  “Readily-taxable” Services

July 1 2001 Sales

Tax Rate

Revenue from

Taxing Services

FY01 State General

Sales Tax Revenue

Revenue from Services as Percent of Current

Revenue

(1)

(2)

(3)

(4)

(5)

(6)

United States

100.00%

1,148,000

5.000%

57,400

184,315

31%

Alabama

1.26%

14,522

4.000%

581

1,823

32%

Arizona

1.58%

18,164

5.600%

1,017

3,757

27%

Arkansas

0.71%

8,150

5.125%

418

1,772

24%

California

13.00%

149,251

5.750%

8,582

24,298

35%

Colorado

1.70%

19,560

2.900%

567

1,970

29%

Connecticut

1.67%

19,226

6.000%

1,154

3,475

33%

D. of Columbia

0.26%

3,037

5.750%

175

682

26%

Florida

5.47%

62,786

6.000%

3,767

14,862

25%

Georgia

2.78%

31,867

4.000%

1,275

4,874

26%

Hawaii

0.41%

4,697

4.000%

188

1,560

12%

Idaho

0.37%

4,303

5.000%

215

782

28%

Illinois

4.75%

54,528

6.250%

3,408

6,565

52%

Indiana

1.96%

22,473

5.000%

1,124

3,606

31%

Iowa

0.92%

10,569

5.000%

528

1,756

30%

Kansas

0.89%

10,182

4.900%

499

1,745

29%

Kentucky

1.17%

13,404

6.000%

804

2,656

30%

Louisiana

1.26%

14,493

4.000%

580

2,665

22%

Maine

0.40%

4,548

5.000%

227

818

28%

Maryland

2.18%

25,021

5.000%

1,251

3,326

38%

Massachusetts

2.86%

32,833

5.000%

1,642

3,756

44%

Michigan

3.43%

39,369

6.000%

2,362

7,686

31%

Minnesota

1.90%

21,773

6.500%

1,415

4,768

30%

Mississippi

0.72%

8,223

7.000%

576

2,330

25%

Missouri

1.83%

21,021

4.225%

888

2,805

32%

Nebraska

0.57%

6,547

5.000%

327

1,017

32%

Nevada

0.73%

8,329

6.500%

541

2,049

26%

New Jersey

3.76%

43,220

6.000%

2,593

5,759

45%

New Mexico

0.49%

5,603

5.000%

280

1,731

16%

New York

7.89%

90,585

4.000%

3,623

8,449

43%

North Carolina

2.60%

29,795

4.000%

1,192

3,430

35%

North Dakota

0.19%

2,174

5.000%

109

403

27%

Ohio

3.78%

43,356

5.000%

2,168

6,288

34%

Oklahoma

1.00%

11,476

4.500%

516

1,550

33%

Pennsylvania

4.35%

49,932

6.000%

2,996

7,238

41%

Rhode Island

0.37%

4,233

7.000%

296

699

42%

South Carolina

1.17%

13,375

5.000%

669

2,507

27%

South Dakota

0.23%

2,669

4.000%

107

464

23%

Tennessee

1.79%

20,492

6.000%

1,230

4,483

27%

Texas

7.02%

80,626

6.250%

5,039

17,903

28%

Utah

0.63%

7,260

4.750%

345

1,480

23%

Vermont

0.20%

2,319

5.000%

116

356

33%

Virginia

2.69%

30,836

3.500%

1,079

2,992

36%

Washington

2.21%

25,367

6.500%

1,649

6,078

27%

West Virginia

0.48%

5,454

6.000%

327

1,086

30%

Wisconsin

1.82%

20,916

5.000%

1,046

3,613

29%

Wyoming

0.17%

1,924

4.000%

77

406

19%
*”Readily-taxable” services consist of all services consumed by households except: housing, health care, education, transit, banking, insurance, legal and funeral services.  See Methodology Note for description of data sources.

Table 1 Methodology Note:
Estimating the Revenue Yield from Taxation of Services Using Nationwide Data

Federal government statistics can be used to generate a rough, “order-of-magnitude” estimate of the revenue potential of comprehensive state sales taxation of services purchased by households.  The “Gross Domestic Product” (GDP) accounts of the United States published by the Commerce Department include a reasonably detailed breakdown of household purchases of services; these have the advantage of being published each quarter and calculated in such a way that sales of goods by service businesses are not inappropriately counted as sales of services.

Table 1 develops a state-by-state estimate of the potential sales tax revenue that would be generated by comprehensive state sales taxation of services purchased by households.  The estimate assumes that states would tax all household purchases of services other than health care, housing, education, legal, banking, public transit, insurance, and funeral services.  For distributional, practical, and/or political reasons, it seems unlikely that policymakers in many states will seek to extend their states’ sales taxes to these categories of services.[vii]  The remaining potentially-taxable services are labeled “readily-taxable services” in Table 1.

The major shortcoming of using the GDP data on household consumption of services to generate a sales tax revenue estimate is that the data are only available for the entire United States.  Thus, the estimates in Table 1 are based on the assumption that sales of particular services in a state are proportional to the size of the state’s economy — as measured by the state’s share of national personal income reported by the Commerce Department.[viii]  The first column of Table 1 reports the share of national personal income realized by residents of each state.  Column 2 takes $1.148 trillion in 2001 nationwide purchases of readily-taxable services from the GDP accounts and apportions these purchases to each state in proportion to that state’s share of national personal income.  Column 4 calculates the revenue that would have been generated in Fiscal Year 2001 by fully taxing the readily-taxable services, by multiplying the state sales tax rate in effect on July 1, 2001 (column 3) by the amount in column 2.  Column 5 reports actual state sales tax collections in 2001.[ix]  Finally, Column 6 calculates the revenue yield from full taxation of readily-taxable services as a share of actual sales tax collections in that year (Column 4 divided by Column 5).

Table 1 indicates that:

  • Full taxation of “readily-taxable” services could generate sales tax revenue equal to 25-35 percent of current sales tax collections in about three-fourths of the 45 states currently levying a sales tax.
  • The total revenue yield nationally would be approximately $57 billion per year.
  • The annual revenue yield in specific states ranges from more than $8 billion in California to just $77 million in Wyoming.

Of course, most states already tax some household services, and so it would be unreasonable to expect their revenues to increase by the percentages shown in column 6.  Table 2 (see PDF of summary) shows which states tax 40 services purchased predominantly by households, selected from the Federation of Tax Administrators 1997 survey cited above.[x]  Downward adjustments need to be made to the figures in column 6 of Table 1 to take into account existing taxation of services before those figures can be used to gauge the potential gain in revenue from broadening the sales tax base.

In addition, if the effective price of services rose by anywhere from four to ten percent due to the imposition of a sales tax, sales of services — and hence the revenue yield — likely would drop as well initially due to the normal inverse relationship between price and consumer demand.  It is also reasonable to expect that there would be a certain amount of non-compliance by certain service providers even if they were legally required to charge sales tax — for example, by individuals providing services in peoples’ homes on an informal basis.

Even after making downward adjustments for these factors, however, Table 1 suggests that many states could reasonably expect to realize 25-30 percent increases in sales tax revenues if they extended their sales taxes to all “readily-taxable” services.  In particular, states like California, Colorado, Illinois, Massachusetts, Michigan, and Virginia, which tax virtually no services currently, could gain substantial amounts of revenue.

View full PDF of summary

End Notes:

[1] Federation of Tax Administrators, Sales Taxation of Services, 1996 Update, April 1997.  Available at http://www.taxadmin.org/fta/pub/services/services.html.  There have been some changes in state taxation of services since 1996, but most of them are new exemptions, not tax base expansions.  FTA has begun compiling a list of changes in sales taxation of services since 1996 at the indicated Web site.

[2] In a majority of states with sales taxes, some kinds of local governments are also authorized to impose them.  In most instances, the local sales tax base — the group of items subject to tax — is substantially similar or identical to the state sales tax base.  In the interest of readability, this report will refer to “state” sales taxes, but nearly all of the discussion applies to local sales taxes as well (including the sales tax of the District of Columbia).

[3]  Texas estimates that it lost $623 million in sales tax revenue in FY01 by failing to tax new residential construction labor, residential repair and remodeling, barber and beauty services, auto maintenance and repair, and car washes.  Source: Tax Exemptions and Tax Incidence Report, available at www.window.state.tx.us/taxinfo/incidence/table6.html.  The estimated revenue loss from Michigan’s failure to tax services is in the Executive Budget Tax Expenditure Appendix, Fiscal Year 2003, p. 44.

[4]  The remaining household consumption not accounted for in these figures is composed of housing and food for use at home.  The source of the data is the personal consumption expenditures component of the Gross Domestic Product Accounts published by the Commerce Department.

[5]  Such generalized taxation of services would have to exempt health care services for this statement to be fully accurate, however.  Research suggests that imposing sales taxes on health care services in addition to most other household services would increase the regressivity of the sales tax.

[6]  Information about NAICS is available at www.census.gov/epcd/www/naics.html.  Information about NAPCS is available at www.census.gov/eos/www/napcs/napcs.htm. 

[vii]  While it seems unlikely that states would seek to impose their sales taxes on monthly rent or mortgage payments, it would be feasible for states to tax labor charges for the construction, renovation, and repair of residential housing.  The GDP data for housing consumption do not permit a break-out of amounts spent on such labor, and so housing is excluded entirely from the revenue estimates developed in this section.  The exclusion of residential construction labor means that the estimated revenue yield from taxing services shown in Table 1 is conservative.

[viii]  This methodology implicitly assumes that the share of total personal income devoted to consuming “readily-taxable services” is the same in every state.  This is unlikely to be true for at least two reasons.  First, states in which the average household is relatively affluent are likely to devote a below-average share of personal income to consumption of “readily-taxable services” because households will save rather than consume above-average shares of income.  Second, states with relatively high housing costs or health care expenditures (for example, because of a disproportionately elderly population) will have less discretionary income to devote to buying “readily-taxable services.” 

The goal of Table 1 is to develop an order-of-magnitude estimate for sales taxation of services, and the implicit assumption of equal shares of personal consumption devoted to purchases of services seems consistent with that objective.  However, the shortcomings of the assumption should be kept in mind.

[ix]  The FY01 sales tax collections data reported in Column 5 have been generated by Professor John Mikesell of Indiana University by adjusting Census Bureau data.  The state sales tax collections data published by the Census Bureau itself have certain methodological shortcomings that make them unsuitable for a calculation like that in the right-most column of Table 1.  Professor Mikesell’s adjustments to the Census Bureau data are described in John L. Mikesell, “Retail Sales Taxes, 1999-2001: The Recession Hits,” State Tax Notes, February 10, 2003, pp. 489-99; this article is also the source of the data in Column 5.

[x]  Table 2 has updated the results of the 1996 FTA survey with information recently supplied to FTA by the states and posted on its Web site at http://www.taxadmin.org/fta/pub/services/state1.html.  Not all states have provided updates, so it is possible that Table 2 is not a completely accurate reflection of current state taxation of the enumerated services.