Tài liệu Taxation without Representation in Contemporary Rural China ppt

This result is due in part to the fact that prior research generally does not undertake the question of merit directly, and has been unable to find an adequate proxy for claim merit.Palm

PROFESSIONAL LIABILITY SUITS AGAINST TAX

ACCOUNTANTS: SOME EMPIRICAL EVIDENCE

REGARDING CASE MERIT

Donna D Bobek, Richard C Hatfield and Sandra S Kramer 3

AN EMPIRICAL ASSESSMENT OF SHIFTING THE

PAYROLL TAX BURDEN IN SMALL BUSINESSES

AN EMPIRICAL EXAMINATION OF INVESTOR OR DEALER STATUS IN REAL ESTATE SALES

CHARITABLE GIVING AND THE SUPERDEDUCTION:

AN INVESTIGATION OF TAXPAYER PHILANTHROPIC

BEHAVIOR FOLLOWING THE MOVE FROM A TAX

DEDUCTION TO A TAX CREDIT SYSTEM

v

HOW ENGAGEMENT LETTERS AFFECT CLIENT LOSS

AND REIMBURSEMENT RISKS IN TAX PRACTICE

Lynn Comer Jones, Ernest R Larkins and Ping Zhou 95

THE ALTERNATIVE MINIMUM TAX: EMPIRICAL

EVIDENCE OF TAX POLICY INEQUITIES AND A RAPIDLY

INCREASING MARRIAGE PENALTY

John J Masselli, Tracy J Noga and Robert C Ricketts 123

AN EMPIRICAL INVESTIGATION OF FACTORS

INFLUENCING TAX-MOTIVATED INCOME SHIFTING

EXPORT INCENTIVES AFTER REPEAL OF THE

EXTRATERRITORIAL INCOME EXCLUSION

LIST OF CONTRIBUTORS

Timothy O Bisping Department of Economics and Finance, Louisiana

Tech University, USA

Donna D Bobek School of Accounting, University of Central

Florida, USA

Tracy L Bundy School of Professional Accountancy, Louisiana

Tech University, USA

Ted D Englebrecht School of Professional Accountancy, Louisiana

Tech University, USA

Alexander M G Gelardi College of Business, University of St Thomas,

St Paul, MN, USA

Richard C Hatfield Department of Accounting, University of Texas at

San Antonio, USA

Paul D Hutchison Department of Accounting, University of North

Texas, USA

Lynn Comer Jones Department of Accounting and Finance,

University of North Florida, USA

Sandra S Kramer Fisher School of Accounting, University of

Florida, USA

Ernest R Larkins School of Accountancy, Georgia State University,

USA

John J Masselli Area of Accounting, Texas Tech University, USA

Tracy J Noga Department of Accounting, Suffolk University,

USA

Robert C Ricketts Area of Accounting, Texas Tech University, USA

Toby Stock School of Accountancy, Ohio University, USA

vii

Craig G White Area of Accounting, University of New Mexico,

USA

University of New York – Baruch College, USA

EDITORIAL BOARD

EDITORThomas M Porcano

ADVANCES IN TAXATION

EDITORIAL POLICY AND

CALL FOR PAPERS

Advances in Taxation (AIT) is a refereed academic tax journal published

annually Academic articles on any aspect of federal, state, local, or internationaltaxation will be considered These include, but are not limited to, compliance,computer usage, education, law, planning, and policy Interdisciplinary researchinvolving, economics, finance, or other areas also is encouraged Acceptableresearch methods include any analytical, behavioral, descriptive, legal, quantita-tive, survey, or theoretical approach appropriate for the project

Manuscripts should be readable, relevant, and reliable To be readable, manuscriptsmust be understandable and concise To be relevant, manuscripts must be directlyrelated to problems inherent in the system of taxation To be reliable, conclusionsmust follow logically from the evidence and arguments presented Sound researchdesign and execution are critical for empirical studies Reasonable assumptionsand logical development are essential for theoretical manuscripts

AIT welcomes comments from readers.

Editorial correspondence pertaining to manuscripts should be forwarded to:Professor Thomas M Porcano

PROFESSIONAL LIABILITY SUITS

AGAINST TAX ACCOUNTANTS:

SOME EMPIRICAL EVIDENCE

REGARDING CASE MERIT

Donna D Bobek, Richard C Hatfield

and Sandra S Kramer

ABSTRACT

As with most professional service occupations, liability claims are a major concern for accounting professionals Most of the academic research on accountants’ professional liability has focused on audit services This study extends research on accountants’ professional liability by examining liability claims arising from the provision of tax services In addition to a descriptive analysis, the current study explores the role of merit in tax malpractice litigation Hypotheses are developed based on the legal construct of claim merit, which requires the presence of accountant error and damages as a result of that error for a claim to be considered meritorious The hypotheses are tested using logistic and OLS regression of 89 actual claims filed with

an insurer of tax professionals The results suggest that the components

of merit are significant in determining both the presence of compensatory payments to the client and the dollar amount of those payments, although the hypothesized interaction effect is only significant for the dollar amount of compensatory payments.

4 DONNA D BOBEK ET AL.

“principal cause” of the profession’s liability problems (Arthur Andersen & Co

et al., 1992) Using detailed claim files from an accountant insurance company,

we explore this issue in the tax accounting profession

Palmrose (1997) undertook a review of the audit malpractice literature in aneffort to answer the question, “do the merits of a case matter with regards to bring-ing and resolving claims against auditors?”Kinney (1993)asserts that meritoriousclaims against independent auditors require three elements: substandard financialstatements, substandard audits, and compliance with relevant legal standards(e.g detrimental reliance on the financial statements) However,Palmrose (1997)suggests that the low probability of bringing a claim to court actually seversthe theoretical tie between merits and outcome The empirical data drawn fromseveral studies (e.g Dunbar et al., 1995; Palmrose, 1994) suggest that the role

of merit is inconclusive (particularly with regard to outcome) This result is due

in part to the fact that prior research generally does not undertake the question

of merit directly, and has been unable to find an adequate proxy for claim merit.Palmrose (1997)concluded her study with a call for research that examines therole of merit in accountant malpractice claims

Although most research examining the issue of malpractice liability in theaccounting profession deals with auditor liability, the issue of tax professionalliability is also of concern Several different sources have quantified the fact thattax accounting engagements result in more claims brought by clients than anyother area of accounting practice (although audit claims are higher in total costs)

In fact, the AICPA reports that 60% of all accountant malpractice claims in theAICPA Professional Liability Insurance program arise from tax engagements(Anderson & Wolfe, 2001) This is up from 43% ten years ago.Donnelly et al.(1999) note that the frequent enactment of tax law changes and the relativelyrecent inclination of the IRS and the Tax Court to hold practitioners responsiblefor client information has put additional pressure on small and midsize accountingfirms This pressure, they add, has led to “more frequent and more severe mal-practice claims arising from tax planning and preparation” (p 59) Although theoccurrence of tax malpractice claims is quite high, the research regarding this issuehas been limited

This study extends prior research in both auditing and tax litigation Taxresearch in this area is fairly new and has yet to address the important relationshipbetween claim merit and claim outcome And, although some audit research hasdirectly addressed the issue of merit (e.g Carcello & Palmrose, 1994; Dunbar

et al., 1995), audit researchers typically have a difficult time finding an adequateproxy for claim merit We begin by developing a definition of claim merit based

on case law (Anderson, 1991; Rockler vs Glickman, 1978) Specifically, we define

a meritorious case as one that contains both tax professional error and damagesoccurring as a result of that error We also hypothesize that meritorious claimsshould be more likely to result in compensation being paid to the client, as well aslarger payment amounts We examine these hypotheses with data from the files of

an insurance company (the files contain good proxies for these two components

of merit)

As prior audit research has suggested, it appears that a claim does not have tomeet the strict legal criteria of a meritorious claim in order to result in a compen-

satory payment to the client Our results suggest that the existence of either error

on the part of the tax professional or damages incurred by the client is enough to

result in compensatory payments However, there is a fairly large and significantdifference in the magnitude of payments for claims with both error and damagescompared to all other claims, after controlling for the overall size of the claim

In fact, claims with both components of merit resulted in average compensatorypayments that were more than four times larger than other claims in our sample($62,921 vs $15,284).1Thus, we conclude that the effect of the components ofclaim merit, as suggested by case law, are a significant determinant of both the like-lihood of compensatory payments being made, and the amount of those payments.The remainder of this article is organized as follows In the next section, priorresearch regarding professional liability of accountants is discussed; followed by

a definition of claim merit and development of the hypotheses This is followed

by a description of the variables and descriptive data regarding the sample Inthe next section, results are reported and discussed Finally, conclusions andopportunities for future research are discussed

PRIOR RESEARCH

There are two streams of research on which this study draws First, there issome prior research that deals directly with tax accountant liability, althoughthis research does not address the issue of claim merit Second, there is a largerbody of research regarding audit litigation The audit environment shares somekey characteristics with the tax environment For example, both originate from

6 DONNA D BOBEK ET AL.

accounting firms that may have relatively deep pockets Second, the rate at whichtax claims are brought to trial is similarly low (11% for our sample vs 10% forPalmrose (1991)) However, there also are differences The key difference is thattax professionals serve as paid advocates of the client, while auditors work for theshareholders and are required to be independent of the client’s managers Further,

in the current study we focus on tax professionals from small firms, while mostaudit research has examined Big 5 accountants.Russell (2002)presents surveyresults demonstrating that the median firm size of CPAs in private practice focused

on tax work is just one or two professionals, while the average firm size is aroundfour professionals

Prior Tax Research

The literature on tax practitioner liability is in the early stages and primarily scriptive in nature Prior descriptive research has addressed issues such as: whichareas of tax planning/compliance are more likely to result in malpractice claims(e.g.Anderson & Wolfe, 2001; Demery, 1995; Donnelly & Miller, 1990, 1995;Donnelly et al., 1999), and tips for avoiding malpractice claims (e.g.Anderson,1991; Bandy, 1996; Holub, 2001; Kahan, 1999; Yancey, 1996) Areas that wererepeatedly identified as problem areas include S Corp elections, complex areassuch as estate tax and partnership taxation, like-kind exchanges, and filing errors(Anderson & Wolfe, 2001; Donnelly & Miller, 1995; Donnelly et al., 1999) Com-mon tips for avoiding malpractice problems include the use of engagement letters,avoiding “problem” clients and situations (e.g divorce), proper documentation

de-of procedures, proper communications with the client, and adequate malpracticeinsurance coverage (Anderson, 1991; Bandy, 1996; Holub, 2001; Yancey, 1996).Although there has been some research considering factors that influencethe decision to file a claim against a tax professional (Krawczyk & Sawyers,1995;Schisler & Galbreath, 2000), research has not yet addressed which factorsinfluence whether a filed claim will result in the tax professional actually makingcompensatory payments.Krawczyk and Sawyers (1995)report the results of anexperiment that varied the nature of the engagement letter and the magnitude ofthe IRS assessment As hypothesized, the magnitude of the IRS assessment waspositively associated with both the likelihood of filing suit and the dollar amountrequested Further, an engagement letter that included a statement limiting thepreparer’s financial liability to the amount of the fees paid had the expectedeffect of decreasing the dollar amount requested in the suit However, it hadthe surprising effect of increasing the probability of filing a suit Schisler andGalbreath (2000)found that relative to non-involved observers, subjects who were

placed in the perspective of the taxpayer were more likely to hold tax preparersresponsible for bad outcomes, while taking credit for positive outcomes.

Prior Audit Research

Palmrose (1997)reviewed the relevant audit literature in an attempt to answerthe question “does merit matter” in malpractice litigation Her motivation forthe study was to provide input for the ongoing debate over legal reform and thereduction of auditor liability For example, she cites a Statement of Position by theBig Six (Arthur Andersen & Co et al., 1992, p 1) which claims that “the principalcauses of the accounting profession’s liability problems are unwarranted litigationand coerced settlements.” Palmrose’s ultimate conclusion was that the evidence

to date was not conclusive, particularly with regard to the outcome of filed claims.Alexander (1991)provides a rationale for why merit may not be important in secu-rities litigation The involvement of insurance companies, officers and directors asdefendants, and certain rules of law combined to make the likelihood of carryingsuch cases to court very small Alexander argued that once the option of trial isvirtually eliminated, the outcome of malpractice claims is no longer a function

“matter enough” (Palmrose, 1997, p 365), there also has been the issue of finding

an adequate proxy for claim merit

While not satisfactorily addressing the issue of claim merit, prior research onauditor liability has identified a number of factors that were related to litigationoutcomes The first is characteristics of the auditor Research has found thatfirm size, experience, and number of years on the particular engagement aresignificantly related to litigation outcome (Palmrose, 1988; St Pierre & Anderson,1984) Second, characteristics of the client, such as industry membership, financialcondition, market value, variability of return, bankruptcy, and size have beenrelated to audit litigation outcome (Lys & Watts, 1994;Palmrose, 1994; Stice,1991) Third, research also has examined the event that triggered the error search.For example, negative financial signals from the client and the client’s industryand regulatory reviews (e.g SEC action) have been determined to prompt thesearch for errors (St Pierre & Anderson, 1984) Finally, characteristics of theaudit also may influence the outcome (e.g structure of the audit, portion of total

8 DONNA D BOBEK ET AL.

revenues/independence, report type given, error type) Several of these variables(e.g firm size, experience, client relationship) also may be related to tax mal-practice litigation While we consider these variables in our “additional analyses”section of our results, the purpose of the current study is to focus on the definition

of claim merit in a tax setting and to assess its effect on the outcome of taxmalpractice claims

DEFINING MERIT IN A TAX SETTING

The focus of this study is to examine the relationship between the merit of amalpractice claim and the likelihood that compensatory payments are made bythe accounting firm (or their insurance provider) Accountants are held to thesame standards of care as lawyers, doctors, and other professionals (Rockler vs Glickman, 1978) Anderson (1991) details the extensive malpractice case lawprecedent directly related to lawyers and accountants This case law requires both:

(1) an actual breach of duty by the professional; and (2) damages to the client

because of that breach of duty, before there can be a holding of malpractice.2These two factors3should be the hallmarks of a meritorious case and should be

the distinguishing factors between the group of claims for which compensatorypayments are made and the group of claims where there are no compensatorypayments

Breach of duty or error(s) on the part of the tax professional include at leasttwo broad categories First, the tax professional may provide inaccurate planningadvice or may inaccurately complete the tax return Second, the tax professionalmay inappropriately file a tax return or other tax related document (e.g elections).Although legally, the tax professional should not be liable unless he/she makes anerror resulting in damages to the client, the client may incur tax related damagesfor reasons other than error on the part of the accountant For example, the clientmay provide inaccurate or incomplete information to the tax preparer or maynot completely follow the tax professional’s advice/instructions Any resultingdamages would not be due to the work of the tax professional

An error on the part of the tax accountant is only the first requirement for ameritorious malpractice claim The client also must incur financial damages as

a result of the error Unfavorable consequences can originate with the IRS when

it assesses penalties for late or procedurally deficient filings or selects the returnfor audit If the return is audited, additional taxes, penalties and interest may beassessed or the IRS may determine the return is correct as filed While additionaltaxes, penalties and interest assessed by a tax authority are likely to be the majorsource of financial damage, there certainly are other tax-related damages that can

occur as the result of an incorrect tax return For example, missing the required datefor filing an S election could mean a firm had to file as a C corporation If the errorwas found and the proper C corporation return filed, there would be no IRS penaltybut the corporation would still suffer financial damages in the form of an increasedtax liability.

Theoretically speaking, only meritorious claims should result in compensatory

payments to the client However, prior audit research, as well as anecdotal vation, shows that there are other reasons, including the cost involved in defending

obser-a clobser-aim, the low likelihood of the cobser-ase ending up in court, obser-and the uncertobser-aintyinvolved in proving that a claim is not meritorious, that may lead accountants (andtheir insurance company) to make some form of compensatory payment eventhough the merits of the claim are not completely clear Thus, we hypothesize that

claims filed against tax professionals that are meritorious are more likely to result

in compensatory payments by the tax professional (or their insurance company)

to the client Additionally, the compensatory payments should be, on average,larger for claims that are meritorious than for non-meritorious claims This leads

to the following two hypotheses, stated in alternative form:

Hypothesis 1 Claims where both tax professional error and tax-related

damages (i.e claim merit) are present will result in a greater frequency ofcompensatory payments to the client than will claims where both of thesecharacteristics are not present

Hypothesis 2 Claims where both tax professional error and tax-related

damages (i.e claim merit) are present will result in a larger dollar amount ofcompensatory payments to the client, than will claims where both of thesecharacteristics are not present

DATA

An insurance firm that provides liability insurance to accountants in local andregional accounting firms with 1–100 professionals provided access to all the taxmalpractice claim files that were closed during the period of January 1994 to March

1997 All cases were no longer active (dropped, settled, or litigated) by May 1998.The insured accounting firms are required to report to the insurer any situation

in which the accountant thinks a claim may be filed If a claim of malpractice

is filed, the insurance company hires a tax expert to gather the facts, assess thesituation, and to recommend action to the CEO of the insurance firm The claimcan be settled by the insurance company, dropped by the client, or litigated Theinsurance company files included the facts as set out by the tax expert, information

10 DONNA D BOBEK ET AL.

Table 1. Descriptive Data about the Claims

Claim Information

Number of claims in sample 89

Range of dates of incident 1989–1996

Average # of months claim opened 18 months

Who identified the issue

% of total with payments 48%

Mean payment/claim $24,811 $49,047 Legal expenses

% of total with legal expenses 54%

Mean legal expense/claim $13,700 $25,116

a Includes claims where CPA was concerned about a malpractice claim and notified the insurance company, but client never followed up.

about the accounting firm involved, and information about compensatory paymentsand costs paid by the insurance company.4

Tables 1 and 2present descriptive information about the claims, the accountantsand the clients The dates the claims were reported to the insurance companyrange from 1989 to 1996 The average amount of time it took these claims to beclosed was 18 months (considerably shorter than the 4.3 yearsPalmrose (1997)reported for the non-payment auditing claims) Forty-four percent of the claimswere eventually dropped, 42% were settled and only 11% were the result of ajudge or jury verdict Forty-eight percent of the claimants received some amount

of compensatory payment This is similar to the percentage reported for auditclaims by Palmrose (1997) For those 43 claims that resulted in compensatory

Table 2 Characteristics of Client, CPA/Firm and Claim Issue.

Mean Median CPA and firm info

Years experience of CPA 20.6 years 20 years Number of partners in firm 5.7 5 Number of other (non-partner) CPA’s in firm 9.3 7 Number of total employees in firm 29.7 23 Percentage of billing from tax services 38 38 Client info

% of claims with established clients (>3 years) 43

% of claims with engagement letters 54

% of claims with tax-only clients 58

Client entity type

% of claims clearly about a tax issue 70

% of claims arising from a fee or tax dispute 34

% of claims where accountant alleged client provided

erroneous information

17

Partial list of tax issues

Number of claims relating to S corps and partnerships 10

Number of claims relating to estates and trusts 6

Number of claims relating to pensions 4

Number of claims that were not related to federal taxes 14

Number of claims that involved erroneous filings 12

a These categories are not mutually exclusive.

payments, the average payment was $49,047 For the 54% of the claims wherelegal expenses were incurred, the legal expenses averaged $25,116 For the entiresample of claims, the total average cost per claim (compensatory payments andlegal expense) was $38,511, and 35% of this amount was for legal expenses.The typical CPA was from a small CPA firm and had over 20 years ofexperience Approximately 38% of firm billings were from tax services and therewere, on average, 30 employees employed by their firm Most of the clients in oursample were either individual or corporation tax service-only clients Forty-threepercent had been clients of the respective tax professionals for four years ormore Fifty-four percent of the claims reported the presence of an engagementletter

12 DONNA D BOBEK ET AL.

There was a wide array of issues from which the claims arose The taxareas that were identified in the claims were similar to those noted in priordescriptive research, including S Corporations, partnerships, estate and trustissues, non-Federal income tax issues (e.g sales tax, excise tax, payroll taxes andstate income tax), and failure to properly file required tax forms Interestingly,however, a number of claims, at least partially, arose from “non-tax” issues Forexample, 34% of the claims reported either a fee dispute between the accountantand the client, or a client who was disgruntled about the amount of tax they owed.5Fee disputes have been mentioned in prior descriptive research (e.g Anderson,1991; Stimpson, 2001; Yancey, 1996), and are not a new concern.Yancey (1996)notes a case that involved a fee dispute from the 1960s; andStimpson (2001)goes

so far as to recommend that accountants not sue for fees, as this may lead to anincrease in the frequency of malpractice claims The number of claims that weidentify as involving fee disputes seems to confirm that advice Finally, for 17%

of the claims, the accountant alleged that the problem occurred because the clientprovided either erroneous or untimely information

Variables

Hypothesis 1 examines factors that influence whether or not compensatorypayments are made to the client and Hypothesis 2 considers the magnitude ofsuch payments To test Hypothesis 1, the occurrence of compensatory payments,

we use a binary dependent variable (one if compensatory payments occurred, zerootherwise) The independent variables of interest for this analysis are whether

or not the tax professional committed an error in completing and/or filing thetax return and whether or not damages occurred Both of these variables and thesurrogates developed are discussed below

The first independent variable measures whether the tax professional committed

an error The insurance files contained an expert’s determination of whether therewas a clear error by the accountant In a number of cases, the accused accountantreadily admitted to the insurance company investigator that he had made a mistake

in completing the return In other cases, the insurance company’s investigatordetermined that the accountant had made a mistake If either the accountant or theinsurance company investigator determined that a mistake was made in preparingthe tax return, the variable CPA ERROR is coded one In addition, many of theclaims involved cases where the IRS assessed penalties because a tax returnwas filed late (or not filed at all) or filed with procedural errors such as missingsignatures.6If the filing was late or procedurally incorrect and the investigatordetermined that the error was the CPA’s error, this also resulted in the variableCPA ERROR being coded one In all cases where it was not clear that the CPA

made either an error completing or filing the return, the variable CPA ERROR

is coded zero

The second characteristic necessary for a claim to have merit is that damageswere incurred by the client Case law indicates that if there are no damages thatresult from any act of malpractice, then no compensatory payments should beassessed against the professional The variable TAX DAMAGES is a measure

of whether the damages alleged were tax-related damages (such as interest orpenalties) Increased present or future costs that resulted from the cause of action,although more difficult to accurately quantify, also were included as tax-relateddamages.7Other alleged damages such as loss of time, and pain and sufferingwere included as zero values for this variable The theoretical relationship is thattax/financial damages must exist for a case to have merit Accordingly, there is noexpectation about the size of the tax damage, only the existence of tax damage.Further, the data available often made it difficult to quantify the amount of taxdamages (e.g present value of future tax payments) Therefore, TAX DAMAGES

is an indicator variable equal to one if there were tax-related damages and zero ifthere were no tax-related damages Therefore, Hypothesis One will be tested withthe following regression:

Compensatory Payments= ␤0+ ␤1(CPA ERROR)+ ␤2(TAX DAMAGES)

Hypothesis 2 considers the magnitude of compensatory payments made by the taxprofessional To consider this continuous dependent variable, a control for the size

of the claim is necessary since both independent variables are indicator variables.The best proxy we have for a size control variable is the amount of compensationrequested by the client in the original claim (COMP REQUESTED) Hypothesis 2

is tested with the following regression:

Compensatory Payments= ␤0+ ␤1(CPA ERROR)+ ␤2(TAX DAMAGES)

14 DONNA D BOBEK ET AL.

Table 3.

Panel A – Logistic Regression

Compensatory Paymentsa= ␤ 0 + ␤ 1 (CPA ERROR) + ␤ 2 (TAX DAMAGES)

+ ␤ 3 (CPA ERROR × TAX DAMAGES)

Independent Variableb Wald p-Value

Chi-Square (Two-Tailed) Intercept 4.883 0.027

cThis condition had a significantly smaller occurrence of compensatory payments (p = 0.000).

the likelihood of compensatory payments being made to the client will increase.Logistic regression was used to explore the relationship between the independentvariables and the dichotomous dependent variable The results of this logisticregression are reported in panel A ofTable 3

The regression correctly classified 77.5% of the claims, and the model chi-squarestatistic is significant at the 0.000 level The main effects of both independent vari-ables are significant, suggesting that each individual characteristic of claim meritincreases the probability of compensatory payments An error on the part of the

CPA (CPA ERROR) was significant at p = 0.027, while the presence of damages (TAX DAMAGES) was significant at the p < 0.001 level Hypothesis One predicts

a significant interaction effect such that the probability of compensatory payments

occurring are the largest when both CPA error and tax damages are present Panel

A ofTable 3shows a marginally significant interaction coefficient (p = 0.10).

However, the effect is difficult to interpret given the dichotomous nature ofthe variables

To further explore the interaction effect, we report the percentage of claimsreceiving compensatory payments in each of the four possible conditions inPanel B ofTable 3 An examination of these percentages reveals that rather thanboth conditions being necessary for compensatory payments to be made (i.e bothCPA error and tax damages), either factor is sufficient That is, if there is CPA erroronly, tax damages only, or both, then compensatory payments are more likely to

be made This finding is not completely consistent with Hypothesis 1’s predictionthat merit (as defined by case law) is necessary for compensatory payments to bemade However, the components of merit do matter, and when neither component

is present, the percentage of claims with compensatory payments is significantlylower than in the other three conditions

Hypothesis 2

Hypothesis 2 predicts that the existence of a meritorious case will affect themagnitude of compensatory payments made To test this proposition, we use thedollar amount of compensatory payments made as the dependent variable Tocontrol for the size of the claim, we include the amount of the compensatorypayment requested (COMP REQUESTED) in the regression.8The results of thisregression are displayed in panel A ofTable 4 Again, both independent variablesrepresenting merit have significantly positive main effects (CPA ERROR and TAX

DAMAGES significant at p = 0.000 and p = 0.014, respectively) The tion, as predicted by Hypothesis 2, also was significant (p = 0.051) Panel B of

interac-Table 4provides the mean values of compensatory payments for the four possibleconditions An examination of the cell means reveals that the dollar amount ofcompensatory payments is consistent with the prediction of Hypothesis 2 Further,

we do a contrast test comparing the mean compensatory payment where bothCPA error and tax damages existed (cell 4) with the other three conditions This

contrast test is significant (p < 0.05), suggesting that claim merit, as defined by

case law, is important in determining the magnitude of the compensatory payment.Although we do not consider hypotheses regarding the total cost of a malpracticeclaim, it is of interest to know how much (if any) the relationship between claimmerit and cost is weakened by considering the cost to investigate and defend theclaim The total cost of the claim includes not just the amount paid to the client,but also the legal costs As noted earlier, the legal costs are significant and amount

16 DONNA D BOBEK ET AL.

Table 4.

Panel A – Regression Results

Compensatory Paymentsa= ␤ 0 + ␤ 1 (CPA ERROR) + ␤ 2 (TAX DAMAGES)

+ ␤ 3 (CPA ERROR × TAX DAMAGES) + ␤ 4 (COMP REQUESTED)

Independent Variableb Coefficient t-Statistic

No $12,794 (n= 34) $2,250 (n= 4) $11,684 (n= 38)

Yes $19,845 (n= 30) $62,921 c(n= 17) $35,425 (n= 47)

Total $16,099 (n= 64) $51,365 (n= 21)

Note: Model R2= 0.539 (p-value = 0.000).

a The dependent variable, Compensatory Payments, is the dollar amount of compensation paid to the client by the accountant and/or insurance carrier.

b The independent variables are as follows: CPA Error is valued at “1” if an insurance expert determined that there was CPA error involved and zero otherwise; Tax Damages was valued at “1” if there was any cost to the client which was tax related (e.g interest, penalties, additional taxes) and zero otherwise; Compensation Requested is the dollar amount of compensation originally requested by the client.

cThis cell was significantly larger than the other three cells (p < 0.05).

to 35% of the total cost of the claim Results from a regression using the sameindependent variables used to test Hypothesis 2 with total costs as the dependentvariable showed that “Compensation Requested” and “CPA Error” were the only

significant variables (R2= 0.593) The coefficient for CPA Error was very similar

to that reported in Table 4 ($48,135), while the coefficient for “CompensationRequested” increased to 0.180 (or 80% higher than inTable 4) Additionally, TAXDAMAGES and the interaction term were no longer significant We conclude fromthis analysis that while merit is significantly related to the amount of compensatory

payment made to the client, it would appear that the size of the compensation

requested drives the amount of effort that is expended to defend the claim, and thusthe total costs

Additional Analyses

Though not the primary focus of this study, we also consider six other possibleinfluences on malpractice claim outcome We examined four non-merit variablessuggested by prior accounting research: firm size, tax professional experience,client relationship, and engagement letter We examined one legal issue, contrib-utory negligence; and finally we analyzed the effect of the presence of “non-tax”issues that led to claims on the likelihood and magnitude of compensatorypayments Firm size, tax professional experience, and client relationship weresuggested by prior audit research Firm size is measured as number of employees.Experience is measured as the tax professionals’ years of experience We uselength of relationship (client tenure) as our client relationship variable.9 Due

to data limitations, this variable is treated as an indicator variable.10 Prior taxresearch showed an influence on likelihood to sue based on the wording ofthe engagement letter (Krawczyk & Sawyers, 1995) Additionally, a number

of commentators suggest that the use of engagement letters will reduce thelikelihood and/or cost of a malpractice claim (e.g.Bandy, 1996; Stimpson, 2001;Williams, 1997) We include a variable, “engagement letter” that was coded one

if the services were covered by an engagement letter, otherwise it was coded,zero

As noted earlier, a few of the accountants (17%) in our sample indicated thatthe reason the client incurred damages was not because the tax professional hadmade an error, but instead because the client had provided the accountant witherroneous information The legal precedent surrounding contributory or compara-ble negligence is complex and differs somewhat by jurisdiction (Anderson, 1991).However, there is some recognition by the courts that the accountant should not

be held fully responsible if the client’s negligence contributed to the accountant’serror (Steiner Corp vs Johnson & Higgins, 1998) Although our measure may

be influenced by the involved accountant’s bias, it provides the best availablemeasure of the quality of the client-provided information Accordingly, the vari-able CLIENT BAD INFO was set to one if the accountant alleged that the clientprovided incomplete or inaccurate information If no such allegation was made,the variable was set to zero Thus, we use this variable, CLIENT BAD INFO, as

an initial exploration of whether or not contributory negligence is related to theoutcome of tax malpractice claims

Finally, although not explicitly suggested by prior accounting research, we alsoinvestigated the phenomena of tax malpractice claims arising from non-tax issues

We consider a variable labeled FEE/TAX DISPUTE, because as noted earlier, weobserved that there were a number of claims that arose either because the clientwas unhappy with the fees charged by the accountant or with the amount of taxes

18 DONNA D BOBEK ET AL.

he/she had to pay This variable was coded one if there was evidence of either ofthese concerns; otherwise it was coded zero

We performed three analyses with these six variables For the dichotomous

variables, we performed univariate t-tests comparing the percentage of claims

with compensatory payments and the dollar amount of compensatory payments ateach level of the independent variable Second, we added these six variables to thelogistic regression reported inTable 3to see if any of these variables improvedthe predictive ability of this model; and third, we added these six variables

to the regression model reported inTable 4to see if any of these variables weresignificant in explaining the dollar amount of the compensatory payment.The univariate results are reported in Table 5 The means for relationship,CLIENT BAD INFO and FEE/TAX DISPUTE were significantly differentbetween the groups of claims (although the FEE/TAX DISPUTE difference wasnot significant for the dollar amount).Table 5shows that a longer relationship is

Table 5 Additional Variables – Univariate Tests.

Independent Variable a % of Claims with Average Dollar Amount

Compensatory Payment of Compensatory Payment Relationship

Client bad info

a Relationship refers to the length of the accountant/client relationship; Engagement Letter refers to the presence or absence of an engagement letter between the accountant and the client; Client Bad Info refers to whether or not the accountant alleged that the client provided incomplete or inaccurate information; and Fee/Tax Dispute refers to whether or not the claim, at least partially, arose because

of a fee dispute or a client disgruntled about paying taxes.

associated with a greater likelihood of compensatory payments While that is incontrast to prior research which showed that errors were more likely to occur withnew clients (St Pierre & Anderson, 1984), further analysis of the data explains theresult Claims involving clients with long relationships were more likely to haveresulted from an IRS Audit (71% for long relationships vs 28% for shorter rela-tionships), and thus more likely to be deemed to have actual damages incurred bythe taxpayer (71% for long relationships vs 43% for shorter relationships) There-fore, our tentative conclusion is not that clients who have a long tenure are morelikely to sue, but instead, that claims by clients with a long tenure are more likely

to be meritorious

Regarding CLIENT BAD INFO, both the percentage of claims with pensatory payments (13.3% vs 56.8%) and the dollar amount of compensatorypayments ($3,214 vs $29,071) were significantly lower in claims where theaccountant alleged that the client had provided incomplete or erroneous infor-mation This suggests either that these claims were less likely to be meritorious,and/or that the legal concept of contributory negligence served to reduce theliability of the accountant Similarly, when the claim involved a fee or tax dispute,

com-it was much less likely to result in compensatory payment (29% vs 60.3%), andalthough the dollar amount of the claims was lower when the FEE/TAX DISPUTEvariable was “yes,” the difference was not statistically significant This seems toimply that a number of unsuccessful claims arise, not as a result of actual error

on the part of the accountant or damages, but because of general dissatisfaction

on the part of the client For the subset of claims where fees were at issue and

no compensatory payments were made, the cost to settle the claims was still over

$2,000 (and this amount does not include any foregone fees, which often werewaived in these cases)

When these variables were added to the logistic regression fromTable 3, onlythe relationship variable significantly affected the result.11 The p-value of the relationship variable was 0.053, the Cox and Snell R2of model improved to 0.373,and the predictive ability of the model increased to 80.9% However, none of theadditional variables were significant when they were added to the CompensatoryPayment regression fromTable 4

In summary, after considering a number of variables suggested by priorresearch, we conclude that while a few of these variables (i.e relationship,CLIENT BAD INFO, and TAX/FEE DISPUTE) are related to claim outcome,the components of claim merit, (CPA Error and Tax Damages) appear to be theprimary determinants of claim outcome Additionally, one could argue that thepattern of influence of relationship, CLIENT BAD INFO and TAX/FEE DISPUTE

is consistent with all three of these variables being additional proxies for the merit

of the claim

20 DONNA D BOBEK ET AL.

Limitations

This study is only a first step to understanding malpractice claims involving taxprofessionals It is limited by the data set The data set has only 89 claims from oneinsurer In addition, all of the claims came from small firms, thus we were unable totest a “deep pockets” hypothesis, which is a prominent issue in auditing research.Also, with regard to the contrast test done for Hypotheses 1 and 2, small cellsize and the noise of the control for size should be considered when interpretingour results

CONCLUSIONS AND FUTURE RESEARCH

Legal precedent would suggest that no client should be able to successfully sue

his/her tax accountant unless there are both error on the part of the accountant

and damages sustained by the client as a result of that error However, recentaudit research has been unable to find a significant link between case merit andcase outcome The results of our test of merit differ somewhat depending onwhether we consider the occurrence of compensatory payments or the magnitude

of compensatory payments When considering the occurrence of compensatory

payments, the pattern of results suggests that either error or damages are required

to result in compensatory payments being made to the client The hypothesizedinteraction effect is not significant when assessing the presence of compensatorypayments However, it does appear that merit has a significant effect on themagnitude of payments made Compensatory payments made in claims havingboth CPA error and tax damages were significantly larger than payments madefor other claims (more than four times larger) However, it should be noted thatclaims lacking either characteristic still resulted in an average compensatorypayment of $12,794

While these results are mixed, we believe they do represent a positive andsignificant development in the study of merit in accountant malpractice We wereable to clearly specify the determinants of a meritorious malpractice claim based

on legal precedent, develop adequate proxies for these determinants, and identifystatistically and practically significant results In addition, we considered anumber of other possible influences on claim outcome suggested by prior researchand observations from our data set Only one of these additional variables, thelength of the relationship with the client, had a significant effect on the likelihood

of compensatory payments, and none of them had a significant effect on the dollaramount of the payment Additionally, the nature of the effect is not inconsistentwith our findings regarding claim merit

Future research is necessary to better understand the area of tax professionalmalpractice Data from a larger number of claims and a wide range of accountingfirm sizes would enrich the results It is certainly possible, for example, thatmerit would not be such a dominant determinant of compensatory payments

if the accounting firms involved had much deeper pockets Additionally, wenoted that the concept of “tax damages” is not homogenous There is a widevariety of explicit and implicit financial costs (e.g present value of future taxes,unnecessary transaction costs, etc.) that the client may incur when the accountantmakes an error Future research examining the make-up and potential magnitude

of damages suffered by the client could be helpful to both our understanding ofthe effect of damages on claim outcome and to accountants in helping to improvetheir work product Also, it is necessary to investigate the threshold question

of why a claim is brought against a tax accountant Audit and medical researchhave investigated the characteristics of situations that lead to a lawsuit as well

as characteristics of professionals who are sued Finding evidence about taxaccountant litigation in this context might be helpful in reducing the number of taxclaims filed

NOTES

1 This is not merely a size effect Claims for which both tax professional error anddamages occurred had damage payments averaging 49.5% of the amount requested,compared to 19.1% for all other claims (over 21

2 times as much)

2 In addition to breach of duty and damages caused by the breach of duty, there aretwo other technical, although less interesting requirements There must be an “accountant-client” relationship, and the accountant must have a duty to provide some service to theclient as a result of that relationship

3 Case law actually considers the second requirement that we mention, “damages

as a result of the breach of duty” as two separate issues: damages and proximate cause.Proximate cause means that there must be a connection between the loss suffered bythe client (the damages) and the accountant’s error (breach of duty) We do not try todisentangle damages and proximate cause

4 The files included some cases (27 claims) where the accountant reported a potentialclaim to the insurance company but no claim was ever filed because the client did notpursue the matter These files included less information since the insurance company neverhired an expert to evaluate the situation These claims are included with the dropped claims

in the analysis Results are not changed by removing these 27 cases from the analysis

5 Other non-tax issues that we observed included the accountant being caught in themiddle of a dispute between two parties (e.g in a divorce), the client looking for “deeppockets,” and embezzlement by an employee of the client

6 While data are clear as to whether the tax filing was inappropriate, there is not always

a clear indication in the data whether the tax accountant error led to the improper filing orwhether a taxpayer error led to the problem

22 DONNA D BOBEK ET AL.

7 For example, one claim involved accounting for an estate that was done improperlyand the client sued for fees paid to a new accountant to correct the errors Another caseinvolved bad advice given to the client to retire early, who then lost 18 months of wages

A third example involved a rather minor error on the part of the tax professional; however,the result was a substantial delay in the client’s refund, which reduced their ability to payoff other obligations

8 For this analysis, seven claims were omitted Five were omitted because there wasnot a specific dollar request made Two were omitted because the dollar requests wereextreme outliers (e.g a $30 million request for a claim with no actual damages)

9 In addition to the length of the relationship between a client and a tax professional,the breadth of the relationship also could be considered We did have information aboutwhether the tax professional provided services to the client in addition to tax services.However, length of relationship and “other services” were correlated [Pearson correlationcoefficient= 0.420 (p-value = 0.000)], so we include only one of these variables in the

regression (results are the same regardless of which one is used)

10 If there was evidence that the client had been with the tax professional for more than

3 years, client tenure was coded “1,” otherwise it was coded, “0.” While this is a somewhatarbitrary choice, in an audit setting,St Pierre and Anderson (1984)defined a new client

to be a client of three years or less

11 FEE/TAX DISPUTE and CLIENT BAD INFO were marginally significant at 0.109and 0.104, respectively, and the model performed better with their inclusion, than withoutthem

ACKNOWLEDGMENTS

The authors are grateful to CPA Mutual Insurance for access to their files and data.The first author is grateful to the PriceWaterhouseCoopers Foundation and theUCF College of Business for financial assistance Helpful comments from RobinRoberts, Dale Bandy, Jack Kramer, two anonymous reviewers and the editor areappreciated

REFERENCES

Alexander, J C (1991) Do the merits matter? A study of settlements in securities class actions.

Stanford Law Review, 43(2), 497–598.

Anderson, S., & Wolfe, J (2001) Accountants’ liability: Where are claims coming from? The Ohio

CPA Journal, 60(4), 21–24.

Anderson, T (1991) Tax practitioner malpractice litigation: Exposure and risk management The

Ohio CPA Journal, 50(4), 30–35.

Arthur Andersen & Co., Coopers & Lybrand, Deloitte & Touche, Ernst & Young, KPMG Peat

Marwick and Price Waterhouse (1992) The liability crisis in the United States: Impact on the

accounting profession, a statement of position.

Bandy, D (1996) Limiting tax practice liability The CPA Journal, 66(5), 46–50.

Carcello, J V., & Palmrose, Z V (1994) Auditor lititagion and modified reporting on bankrupt

clients Journal of Accounting Research, 32(Suppl.), 1–30.

Demery, P (1995) Horror stories from the files of professional liability insurers The Practical

Accountant, 28(11), 24–35.

Donnelly, W., & Miller, G (1990) Tax practice areas where an accountant is most likely to face

malpractice claims Taxation for Accountants, 44(3), 162–166.

Donnelly, W., & Miller, G (1995) Malpractice claims more likely in certain tax areas Taxation for

Accountants, 54(5), 285–290.

Donnelly, W., O’Callaghan, S., & Walker, J (1999) Top 10 tax claims Journal of Accountancy,

187(2), 57–59.

Dunbar, F C., Juneja, V M., & Martin, D N (1995) Shareholder litigation: Deterrent value, merit

and litigants’ options National Economic Research Associates, Inc (NERA).

Holub, S (2001) Avoiding tax malpractice The Tax Adviser, 32(12), 854–856.

Kahan, S (1999) When an engagement snaps The Practical Accountant, 32(11), 71–75.

Kinney, W R (1993) Auditors’ liability: Opportunities for research Journal of Economics &

Management Strategy, 2(3), 349–360.

Krawczyk, K., & Sawyers, R B (1995) The effect of magnitude of IRS assessment and engagement

letters on tax preparer liability Journal of the American Taxation Association, 17(2), 71–88 Lys, T., & Watts, R (1994) Lawsuits against auditors Journal of Accounting Research, 32(Suppl.),

65–93.

Palmrose, Z (1988) An analysis of auditor litigation and audit service quality The Accounting

Review, 63(1), 55–73.

Palmrose, Z (1991) Trials of legal disputes involving independent auditors: Some empirical evidience.

Journal of Accounting Research, 29(Suppl.), 149–186.

Palmrose, Z (1994) The joint & several vs proportionate liability debate: An empirical investigation

of audit-related litigation Stanford Journal of Law, Business & Finance, 53–72.

Palmrose, Z (1997) Audit litigation research: Do the merits matter? An assessment and directions

for future research Journal of Accounting and Public Policy, 16, 355–378.

Rockler vs Glickman 273 N W 2d 647 (1978, Minnesota).

Russell, R (2002) Independent practitioners make over half their earnings from tax preparation.

Accounting Today, 16(Fall), 6–7.

Schisler, D., & Galbreath, S C (2000) Responsibility for tax return outcomes: An attribution theory

approach Advances in Taxation, 12, 173–204.

St Pierre, K., & Anderson, J A (1984) An analysis of the factors associated with lawsuits against

public accountants The Accounting Review, 59(2), 242–263.

Steiner Corp v Johnson & Higgins 135 F.3d 684; (1998 U.S Tenth Circuit Court of Appeals).

Stice, J D (1991) Using financial and market information to identify pre-engagement factors

associated with lawsuits against auditors The Accounting Review, 66(3), 516–533.

Stimpson, J (2001) 21 tips for managing risk Practical Accountant, 34(10), 36–41.

Williams, S (1997) The importance of engagement letters The National Public Accountant, 42(4),

31–32.

Yancey, W (1996) Managing a tax practice to avoid malpractice claims: Learning from past disasters.

The CPA Journal, 66(2), 12–17.

of millions of small businesses in the United States As a result, this study revisits the issue by surveying 4,431 small businesses in Arkansas, Louisiana and Mississippi (ArkLaMiss) Results indicate, in the ArkLaMiss area, that the largest share of the tax burden is borne by customers When compared to past literature, a relatively larger portion of the incidence of payroll taxes is likely to fall on employees in the ArkLaMiss, as opposed to the burden being borne by firms and customers Also, stronger anti-tax sentiment was noted in the ArkLaMiss as compared to prior literature Little support was found for the proposition that firm size impacts the incidence of taxation On the other hand, statistical analysis indicates that the industry within which a firm operates was influential in the incidence of taxation Moreover, in the sample, the banking/financial industry passed the largest percentage of the tax on to employees, the public accounting profession passed the largest percentage on

to customers, and the legal profession bore the largest share of the tax in the form of reduced profit.

INTRODUCTION

In 1935, The Federal Insurance Contribution Act (commonly called the Social

Security Tax) was passed Even though the act has been in effect for almost 70years, it is still as controversial as it was when enacted Specifically, its criticscharge that the social security system will be bankrupt by 2032 (Elias, 1998); it isthe fastest growing tax, in dollar terms, in the U.S system of taxes (Englebrecht

et al., 2001); it is a regressive tax (Iyer, 1994; Pechman, 1987); it hampers savings(Barro, 1978; Feldstein, 1996); and the tax burden is passed on to either consumers(Pechman et al., 1968) or employees from businesses (Brittain, 1971; Ferrara,1980) In regard to the first four charges, there is ample evidence to support thoseassertions However, it is still uncertain as to the actual payroll tax burden (Brittain,1972; Iyer, 1994) Although one-half of the payroll tax is paid by the employeeand one-half by the employer, this is just the statutorily mandated split That

is, the actual burden may be different due to shifting mechanisms Specifically,the employer may shift the burden in the form of higher prices to consumersand/or in the form of lower wages to employees Consequently, it is the intent

of this article to assess the payroll tax burden on small businesses In this regard,responses were solicited from small business owners to gauge whether businessmanagers believe the payroll tax is shifted forward, backward, or supported bybusiness profits

The remainder of this article is organized as follows First, an insight intothe background literature is provided This is followed by an explanation ofthe research design and data collection method The third section presents theresults and analysis of the study In the final section, conclusions, limitations andsuggestions for future research are provided

BACKGROUND

Payroll Taxes The Federal Insurance Contributions Act (FICA) finances three social insurance programs that most taxpayers think of as Social Security First, the Old Age and Survivors’ Insurance (OASI) program that provides cash benefits to retired work-

ers and their families, and to surviving dependents of deceased workers Second,

An Empirical Assessment of Shifting the Payroll Tax Burden in Small Businesses 27

the Disability Insurance (DI) program which supplies cash benefits to disabled workers and their families Last, the Health Insurance (HI) program, popularly

called Medicare, that provides for both hospital and physician reimbursements.The total FICA tax rate is currently 7.65% This rate is broken into twocomponents: Social Security tax (old age, survivors, and disability insurance) andMedicare tax (hospital insurance) In 2003, the Social Security tax rate was 6.2%with a base amount of $87,000 and the Medicare tax rate was 1.45% with no limit

on the base amount Also, the employer must match the employee’s portion forboth Medicare and Social Security taxes

In 1937, the initial social security rate was only 1% on a base amount of $3,000

Of course, the rate and base amounts have grown since that time

Prior Research

In recent years, studies that dealt with the social security system have concentratedprimarily on the benefit side of the system As a result, little attention has beendirected toward analysis of the effects of payroll taxation.Brittain (1972),Ricketts(1990),Iyer (1994), andEnglebrecht et al (2001)1 are the four most importantinquiries dealing with payroll taxation that have a tangential bearing on the currentstudy However, only one of the studies, EHI (2001), is directly related to thisstudy’s research questions Each of these prior studies is summarized in turn

Brittain’s Study The first extensive study on the effects of payroll taxes was

byBrittain (1972) In addition to evaluating the equity effects of payroll taxes,Brittain’s study contained an in-depth evaluation of financing social securitybenefits through a regressive payroll tax Even though the data consisted of actualindividual tax returns, the study was set in the 1960s when the payroll tax ratewas much lower and the underlying income tax structure was much different fromwhat it is today No equity measures were computed, merely the effective taxrates of various families of different sizes were provided The horizontal equityeffects of the social security taxes also were not investigated Moreover, no effortwas made to isolate the effects of changes in the payroll tax structure on smallbusinesses Notwithstanding its numerous deficiencies, it is considered a seminalstudy in the analysis of payroll taxes and over time has provided invaluableguidance in evaluating the social security system

Ricketts’ Study.Ricketts (1990)investigated the vertical and horizontal equityeffects of the combined impact of payroll and income taxes for the years 1980,

1984, and 1988 The information on the tax liabilities for 1980 and 1984 wasdirectly available, but the 1988 tax liabilities were simulated on the 1984 incomedistribution Only the Suits index was employed as a measure of vertical equity,

while the coefficient of variation was used as a measure of horizontal equity.Ricketts found that the regressive effects of social security taxes dominated theprogressive effects of individual income taxes However, no attempt was made toincorporate the effect on small businesses related to payroll tax increases.

Iyer’s Study The primary objective of Iyer’s study (1994) was to hensively evaluate the horizontal and vertical equity effects of the growth inpayroll taxation between 1984 and 1993 To accomplish this goal, a sample oftaxpayers was collected from the IRS panel of individual taxpayers for the years

compre-1984 through 1988 Also, mean income distributions were generated from theseyears on which the income and payroll tax liabilities were simulated for the years1989–1993 The findings suggested that the payroll tax was a regressive tax forthe period of 1984 through 1993 In fact, the regressive effects of the payroll taxdominated the progressive effects of the income tax In regard to horizontal equity,the results were mixed Moreover, asIyer (1994)pointed out, the incidence of thepayroll tax burden has remained an unresolved controversy

EHI Study This study is very different from the prior payroll tax studies.

Unlike those studies, EHI focused on the issue of shifting mechanisms employed

by small businesses in light of payroll tax increases An innovation of the EHIstudy is that it allowed for the burden of the tax to be borne by labor, consumers,

or firms, whereas past work has typically assumed the burden falls solely on labor.Responses were elicited from 1440 small business owners in the Hampton Roadsarea of Virginia to ascertain whether the payroll tax is shifted by passing it on tothe consumer by way of increased prices, passing it on to the employee by way

of reduced wages, or absorbed by the business in the form of reduced profits Theresulting sample of 182 small business owners in EHI revealed that, in general,small businesses are not likely to shift the employer’s share of the tax burden toemployees That is, the most utilized option in dealing with payroll tax increaseswas to increase prices for their products/services

Past work on payroll tax incidence has largely focused on the burden borne byemployees, due in part to the theoretical focus in this area (it is simply a foregoneconclusion in many studies), and perhaps due to data limitations Through the use

of survey data, EHI expanded on previous work by incorporating the potentialburden of other groups Additionally, the survey data gathered by EHI providednot only new data, but also data from a valuable new source, that being, theopinions of business managers

This current study extends EHI by utilizing a three state area, as opposed tothe limited metropolitan area in Southeast Virginia used by EHI Specifically,the survey is three times larger than the above study Additionally, it expands thetypes of data analysis performed by EHI (2001) and uses measures of verticalequity

An Empirical Assessment of Shifting the Payroll Tax Burden in Small Businesses 29

Theoretical Background

As stated previously, the burden of payroll taxes may be shifted to employers

or consumers rather than being borne entirely by the firm As economists havepointed out extensively, the degree to which either of these occurs depends uponthe nature of the relevant markets In this case, two markets come into play Whendetermining the degree to which labor will bear the burden of a payroll tax, onemust consider the nature of the labor market in which the firm is operating Also,

in determining the degree to which consumers will bear the burden of a payrolltax, one must consider the nature of the product market According to economictheory, the degree to which firms are able to successfully pass a tax on to otherparties depends on the nature of demand and supply elasticities in these markets

In terms of the labor market, if for example, the employer faces marketconditions represented by a situation where the supply of labor is relativelyinelastic, then the firm will find it easier to pass that tax on to employees, and, as

a consequence, employees will bear a relatively larger share of the burden of thetax However, when firms face a relatively elastic supply curve, they will find itmore difficult to pass such a tax on to employees without triggering a large supplyresponse In turn, employees will bear a relatively smaller share of the burden ofthe payroll tax (Gruber, 1997)

The share of a tax that is passed to consumers is analyzed in a similar fashion Tothe extent that the entire burden of the tax is not shifted to employees, the questionarises as to whether firms can pass this increased cost on to consumers In general,there is consensus among economists that market elasticities are key determinants

of a firm’s ability to pass such a tax on to consumers As a result, an increase inpayroll taxes will increase the cost of producing any given level of output for a firm,thereby triggering a supply response by the firm Within the context of the market,this supply response will yield a change in the market equilibrium, the exact nature

of which once again depends on the relative elasticities of supply and demand Inthis case, where supply is relatively more elastic than demand, consumers will end

up bearing a relatively larger share of the tax burden On the other hand, if demand

is relatively more elastic than supply, then the opposite is true

It is clear that elasticities play a key role in determining the burden of taxation.However, the extent to which the tax is actually passed on to labor, for instance,

is not widely agreed upon by economists Even though tentative conclusionsreached by economists suggest that most of the tax will be passed to employees inthe form of lower wages, there is no general consensus (Hamermesh, 1993) In thecurrent study, we take a somewhat different approach and attempt to determine theopinion of firms regarding the incidence of a payroll tax In essence, this question

is equivalent to asking firms to reveal their beliefs regarding labor market and

product market elasticities If, for example, firms were to state that they felt thatmost of a payroll tax would be passed to employees, this would be consistent with afirm believing that the labor demand is relatively more elastic than the labor supply.Although we ask nothing specifically about elasticities in our survey, our resultsimplicitly reveal the firm’s beliefs regarding labor market and product marketelasticities.

RESEARCH METHOD

Data Base

In order to compile the data necessary to analyze payroll tax incidence in the case

of small businesses, a survey was mailed to randomly-selected small businesses

in Arkansas, Louisiana and Mississippi in the Spring of 2002 The questionnaireitself is an augmented version of the one used by EHI (2001) The use of thisquestionnaire provides not only the data necessary for a thorough analysis of therespondents’ perceptions of tax incidence, but it also provides an ideal situationfor comparison to this past work, and the opportunity to examine the associatedimplications for the degree to which previous results can be generalized

SAMPLE SELECTION AND DESCRIPTIVE MEASURES

Small businesses in Arkansas, Louisiana and Mississippi receiving the naire were randomly chosen from those small businesses listed in the CorporateAmerica database (Thomson, 2001) For our purposes, a small business wasdefined as one with fewer than 100 employees and less than $5,000,000 inannual sales In this tri-state region, a total of 9,557 firms met these criteria

question-Of these firms, 4,779 were selected to be surveyed, accounting for 50% of thepopulation Of the total number, 1,181 surveys were mailed to firms in Arkansas,2,386 to Louisiana and 1,212 to Mississippi; these numbers correspond to eachstate’s proportion in the total population A total of 348 surveys were returned asundeliverable (119 from Mississippi, 136 from Louisiana and 93 from Arkansas),leaving a total relevant mailing of 4,431 Of the 4,431 surveys mailed, 413 werereturned in usable form, for an overall response rate of 9.32% The responserates for Louisiana, Mississippi and Arkansas were respectively 10.58%, 7.91%,and 8.13%.2Where possible, the characteristics of the sample were compared tothe characteristics of the population, and it was found that the characteristics of

An Empirical Assessment of Shifting the Payroll Tax Burden in Small Businesses 31

the sample roughly correspond to those of the population, though Louisiana wasslightly over represented As is the case in most surveys, complete anonymity wasguaranteed to all firms An attempt was made to make the survey as concise, yetcomplete, as possible while avoiding any information deemed highly sensitive tothe firm in order to maximize the resulting response rate

The questionnaire itself was nearly identical to that of EHI, which thereforeallows for direct comparison to this research (The questionnaire is presented in theAppendix.) However, the EHI sample covered only one metropolitan area in Vir-ginia, whereas this current work covers the entire Arkansas-Louisiana-Mississippiarea (hereafter referred to as the ArkLaMiss) The first two questions of the surveywere designed to gather the opinions of firms regarding who should have to paypayroll taxes In Part One of the survey, questions three and four deal with firms’perceptions on whether the current payroll tax system is fair Question five is anew question that pertains to the role of information technology in the payroll taxsystem Part Two deals with how an employer would react to a one percentagepoint increase in the payroll tax in terms of changing product/service prices,changing employee pay or changes in profits These questions are designed todirectly measure the likelihood that the incidence of an increase in the payroll taxwill fall on consumers, employees or firms Part Three asks the firms to specificallyallocate a new $1,000 payroll tax among employee pay, profit and product/serviceprice This question is extremely valuable in that firms are asked to provide specificdollar amounts, thereby making the incidence of the tax quantifiable Part Four ofthe questionnaire is designed to gather demographic and other general informationregarding firms This information allows one to examine, for instance, whetherfirms bear more of the burden of a tax in one industry relative to other industries,

or if perhaps the smallest firms bear the largest burden on a relative basis

RESULTS

Our results are based on an analysis of the data collected in the context of economictheory regarding tax incidence The primary purpose here is twofold First, whilemuch of the theoretical work in the literature would suggest that the incidence

of a tax varies with market conditions, little has been done to examine this usingsurvey techniques Two important aspects of these conditions available in ourdataset include the industry within which a firm operates and the size of the firm.Here we examine the incidence of payroll taxes broken down according to industryand firm size in order to ascertain the impact of these market conditions Second,

a direct comparison to the EHI study provides insight into the degree to which wecan generalize from our results, or the degree to which incidence of payroll taxes

Table 1 Classification by Industry.

Industry Total Mississippi Arkansas Louisiana EHI

Freq % Freq % Freq % Freq % Freq % Manufacturing 31 7.6 6 7.2 6 6.7 19 8.0 17 9.3 Banking/Financial 6 1.5 1 1.2 2 2.2 3 1.3 2 1.1 Insurance 12 2.9 2 2.4 4 4.4 6 2.5 1 0.5 Architectural/Engineering 10 2.4 2 2.4 1 1.1 7 3.0 6 3.3 Legal 7 1.7 0 0.0 1 1.1 6 2.5 3 1.6 Public accounting 9 2.2 5 6.0 0 0.0 4 1.7 3 1.6 Construction 42 10.2 6 7.2 9 10.0 27 11.4 21 11.5 Real estate 10 2.4 2 2.4 3 3.3 5 2.1 8 4.4 Communications 5 1.2 2 2.4 0 0.0 3 1.3 1 0.5 Health care 54 13.2 14 16.9 9 10.0 31 13.1 15 8.2 Hotels/Restaurants 47 11.5 9 10.8 10 11.1 27 11.4 16 8.8 Computer software 1 0.2 0 0.0 0 0.0 1 0.4 2 1.1 Marketing/Advertising 5 1.2 1 1.2 2 2.2 2 0.8 4 2.2 Public utilities 3 0.7 0 0.0 2 2.2 1 0.4 0 0.0 Research/Development 1 0.2 0 0.0 0 0.0 1 0.4 0 0.0 Transportation 15 3.7 4 4.8 2 2.2 9 3.8 4 2.2 Wholesale distribution 8 2.0 2 2.4 2 2.2 4 1.7 9 4.9 Retail trade 35 8.5 7 8.4 10 11.1 18 7.6 17 9.3 Government – Federal 1 0.2 0 0.0 1 1.1 0 0.0 3 1.6 Government – State 2 0.5 0 0.0 2 2.2 0 0.0 0 0.0 Government – Local 3 0.7 1 1.2 1 1.1 1 0.4 1 0.5 Non-Profit organization 20 4.9 2 2.4 6 6.7 12 5.1 16 8.8 Other (service & farm) 83 20.2 17 20.5 17 18.9 50 21.1 31 17.0

24 hour 0 0.0 0 0.0 0 0.0 0 0.0 1 0.5 Not reported 3 – 1 – 1 – 1 – 1 – Total reported 410 100 83 100 90 100 237 100 181 100

may vary by region.Tables 1 through 7provide a summary of our results in theaggregate as well as by state, along with a comparison to past research.Tables 8–10provide more detailed statistical analysis

Descriptive Statistics

Table 1deals with the distribution of our sample by industry The implications ofthis table are substantial as the firm’s industry is one of the variables chosen to proxyfor those market conditions suggested by economists to influence the incidence oftaxation In turn, any comparisons across states, or studies, will be impacted by thisdistribution By examining the most prevalent industry classifications, it is apparent

An Empirical Assessment of Shifting the Payroll Tax Burden in Small Businesses 33

Table 2 Classification by Annual Revenue.

Total Arkansas Louisiana Mississippi EHI Freq % Freq % Freq % Freq % Freq % Annual revenue

Number of employees

11–25 135 33.1 31 34.4 73 31.1 31 37.3 44 24.2 26–50 158 38.7 30 33.3 106 45.1 22 26.5 72 39.6

Not reported 5 – 1 – 3 1.3 1 – 0 – Total 408 100.0 90 100.0 235 100.0 83 100.0 182 100.0

that a certain degree of uniformity exists The “other services and farm” is easilythe largest group in each state, as well as in the EHI study After this category, inthe ArkLaMiss sample Hotel/Restaurant, Health Care, Construction, Retail Tradeand Manufacturing are the predominant industries This differs slightly from theEHI study in that the ordering varies, even though the make-up of top six industries

is equivalent Since a slight variation exists, a closer look at the numbers revealsthat the share of the largest categories are fairly consistent across states and the twostudies as the difference in percentage terms is fairly small This examination of themajor categories reveals that while slight variation in industry distribution existsbetween samples, nonetheless, the overall distribution is reasonably consistent.This bodes well not only for a comparison of tax incidence across states but alsoacross studies

Tables 2 and 3 deal with another potential proxy for the market conditionsfaced by a firm All else equal, firm size may very well correspond to the degree ofmarket power a firm enjoys Although this is not a perfect proxy, it can be argued

Table 3 Who Should Pay the Payroll Tax?

Who should pay the payroll tax? Total Arkansas Louisiana Mississippi

Freq % Freq % Freq % Freq % Panel A

Need not pay if number of

employees<5

24 6.0 6 7.0 8 3.4 10 12.0 Need not pay if number of

employees<10

37 9.2 5 5.8 21 9.1 11 13.3 Need not pay if number of

employees<15

9 2.2 0 0.0 9 3.9 0 0.0 Need not pay if number of

employees<35

5 1.2 1 1.2 4 1.7 0 0.0 Need not pay if number of

employees<40

9 2.2 0 0.0 6 2.6 3 3.6 Need not pay if number of

employees<45

3 0.7 1 1.2 1 0.4 1 1.2 Need not pay if number of

employees<60

6 1.5 1 1.2 4 1.7 1 1.2 Need not pay if number of

employees<80

1 0.2 1 1.2 0 0.0 0 0.0 Need not pay if number of

employees<100

50 12.5 16 18.6 26 11.2 8 9.6 Everyone must pay some

payroll tax

191 47.6 42 48.8 114 49.1 35 42.2 Not reported 12 – 5 – 6 – 1 – Total reported 401 100 86 100 232 100 83 100 Panel B

Need not pay if annual

revenues<20 thousand

16 4.0 2 2.3 7 3.1 7 8.4 Need not pay if annual

revenues<40

5 1.3 1 1.1 1 0.4 3 3.6 Need not pay if annual

revenues<60

9 2.3 1 1.1 7 3.1 1 1.2 Need not pay if annual

revenues<80

3 0.8 0 0.0 3 1.3 0 0.0 Need not pay if annual

revenues<100

46 11.6 8 9.2 28 12.3 10 12.0

An Empirical Assessment of Shifting the Payroll Tax Burden in Small Businesses 35

Table 3. (Continued )

Who should pay the payroll tax? Total Arkansas Louisiana Mississippi

Freq % Freq % Freq % Freq % Need not pay if annual

revenues<150

7 1.8 1 1.1 4 1.8 2 2.4 Need not pay if annual

revenues<200

19 4.8 3 3.4 9 3.9 7 8.4 Need not pay if annual

revenues<300

8 2.0 1 1.1 3 1.3 4 4.8 Need not pay if annual

revenues<400

0 0.0 0 0.0 0 0.0 0 0.0 Need not pay if annual

revenues<500 83 20.9 21 24.1 50 21.9 12 14.5

Everyone must pay some

payroll tax

202 50.8 49 56.3 116 50.9 37 44.6 Not reported 15 – 4 – 10 – 1 – Total reported 398 100 87 100 228 100 83 100

that at the very least the probability that a firm is able to influence market pricesincreases with the firm’s size The size distribution of firms as measured by thenumber of employees of the firm and the sales volume of the firm thus becomes

an important variable to consider when making cross-study/state comparisons

As reported inTable 2, in the ArkLaMiss, 41.9% of the firms responding had

25 or fewer employees Of those states in the ArkLaMiss, Mississippi appears tohave the largest percentage of small firms InTable 2, firm size also is measuredaccording to sales revenue These results are generally consistent with those whenfirm size is measured according to the total number of employees

The consistency of these descriptive statistics is relevant for at least two reasons.First, our primary focus is the examination of the impact of payroll taxes in the Ark-LaMiss, so to the extent that the states in the region are homogenous, direct com-parisons are more reliable Secondly, it facilitates comparisons with other studies

Perceptions about the Payroll Tax in the ArkLaMiss

Attitudes of firms regarding who should have to pay payroll taxes are presented inTable 3 Small businesses were asked their opinions on the relationship betweenfirm size and whether or not a firm should have to pay the tax The responses show

a relatively strong sentiment against the tax in general, especially for the smallestbusinesses In the ArkLaMiss, 30.6% of the firms responding felt that firms with

Table 4 Is the Social Security Tax System Fair?

Responses Frequency Percentage Strongly disagree (with above statement) 99 24.1

Strongly disagree (with above statement) 25 28.1

Strongly disagree (with above statement) 57 23.9

Strongly disagree (with above statement) 17 20.7

Note: Responses to the statement “Overall the social security system is fair.”

fewer than 25 employees should not pay any payroll taxes This sentiment wasstrongest in Mississippi Only 47.6% of respondents reported that everyone mustpay some payroll tax These results are closely mirrored when firm size is measured

by annual sales rather than number of employees