My bibliography Save this paper. We access four years of TCMP data, ,,, and The TCMP data allows us to observe income and taxes before and after a tax audit. In order to generate a range of scalar estimates of the redistributive impact of more complete compliance we employ the family of extended Gini and concentration coefficients.
We find that the vertical equity effects are very small of negative; however, there is a considerable amount of horizontal inequity generated by noncompliance and in this sense more complete auditing of tax returns could improve the fairness of the tax system. Most related items These are the items that most often cite the same works as this one and are cited by the same works as this one. Corrections All material on this site has been provided by the respective publishers and authors.
Intuit Lacerte Tax. Intuit ProConnect Tax. Intuit ProSeries Tax. Additional Accounting Solutions. Tax Pro Center. Tax Practice Resources. Call sales: All rights reserved.
Terms and conditions, features, support, pricing and service options subject to change without notice. By accessing and using this page you agree to the Terms and Conditions. Certain personal data have been deleted to prevent identification of individual taxpayers. Years are available as exact copies only. This series contains summary data for partnerships reported at the minor, major, and division industry level.
Data include the number of partnerships, number of partners, business receipts, depreciation, taxes paid deduction, interest paid, payroll, payments to partners, and net income.
This series contains records based on data from a stratified sample of Forms P, the tax return filed by private foundations and, after , by IRS Code section a 1 charitable trusts treated as private foundations. CSV is not a fixed format, and unless your Excel columns are very small or there are not very many of them , you are most likely overflowing the 80 byte lines and going to the next line -- which really, really, really messes up a CSV file.
If you have to make it fixed and CSV files usually are not fixed length , make it or something large enough to hopefully handle the longest combination of column lengths in your data. In the course of completing the tax return, a business taxpayer makes reporting decisions related to receipts, business expenses, and personal subtractions such as itemized deductions.
In m a n y cases, a business taxpayer has the opportunity to shift subtractions between the business expense and the personal deduc- tion ledgers. A business taxpayer's motivations in reporting receipts and business expenses is further complicated because decisions regarding these matters m a y affect the perceived financial viability of the business and allowable deductions for such things as Keogh plans. Handling these deci- sions would require a model far more complex than the one with which we are working.
Further, the model would need to incorporate business expenses and allow for the observed mixing of personal subtractions and business expenses. Fewer business-group taxpayers itemize than do those with similar incomes in non- business groups. Of those who itemize, the business groups claim larger average deductions for state and local taxes and for health expenditure, both of which might be difl3cult or disadvantageous to mix with business expenses, than do the nonbusiness classes of similar incomes.
However, the business groups claim less on average for mortgage expenses and real estate tax deductions, where it might be possible to mix business and nonbusiness expenses.
Tax Compliance In initially collecting the T C M P data and later in selecting tax returns for audit, the IRS separates taxpayers into 12 separate audit groups based on the income level, the complexity of the tax return, and the portion of the income from unincorporated business earnings. Previous work by the IRS and others Witte and Woodbury, ; Dubin and Wilde, shows that compliance behavior varies substantially across these taxpayer groups.
Further, the IRS develops separate audit selection rules for each of these groups. We estimate our model for the four audit classes that itemize deduc- tions and that have income primarily from sources other than unincorpor- ated businesses.
The four groups differ by their total positive income i. Table II contains the defini- tions of these audit classes. E M P I R I C A L M O D E L Based on our theoretical work, we specify a taxpayer's reported income and subtractions as depending on "true" income, the legally allowable sub- tractions from income, the tax and penalty structure, tastes and preferences, and the probability of an audit endogenous ] We chose audits as the measure of enforcement policy since this is the enforcement action that has been of most interest in the literature on tax compliance and since the IRS believes that audits are its most effective instrument for stimulating accurate taxpayer reports Comptroller General of the United States, , p.
Given the limited time we had access to the data and the restrictions imposed by the IRS, we estimated only the income-reporting equation. We focus on reported income rather than reported subtractions since most noncompli- ance takes the form of underreported income rather than overreported sub- tractions. We were not allowed to estimate a structural equation for audits because of IRS fears regarding revelation of its audit selection rules.
We account for the fact that the probability of audit is endogenous by estimating the income- reporting equation using two-stage least squares. Two-stage least squares allows for the endogeneity of audits without estimating a structural equation for audit.
Since the probability of audit can assume only a limited range of values i. Note that if a taxpayer'sreports affect the probabilityof an audit then it is the parameters of the audit probability function rather than just the probability of an audit that affectstaxpayer compliance. With our approach we are examiningthe effectof a parallel shift in the log odds of an audit when evaluated at the optimum reported income and reported subtractions for a taxpayer who makes his or her reportingdecisions to maximize expected utility.
See Tauchen and Witte for further details. A business return is a return that meets one of the following conditions. This greatly simplifies estimation of the model without materially altering results.
We identify the income-reporting equation by excluding from it two IRS variables that affect the probability of audit but do not directly affect reporting behavior. We measure these variables at the district office level because most direct IRS contacts, includ- ing audits, are carried out by district office employees.
Tax Compliance The crowding variable reflects restrictions on the IRS's ability to allo- cate its resources across districts as it would like. Understanding the crowd- ing measure requires some explanation of how the IRS selects most tax returns for audit. For most audited returns, the first step in the selection process is the scoring of all filed returns for audit potential using formulas developed with T C M P data from three or more years ago. The formulas, referred to as DIF formulas, are based on discriminant analysis designed to differentiate those returns that show large discrepancies positive or nega- tive between the taxpayers' reports and the auditors' findings versus those returns that show little or no discrepancy.
The DIF formulas assign a numer- ical value to each return, with higher scored returns having a higher audit priority. The DIF formulas are done separately for each audit class. Because of the distribution of its employees across districts, the IRS does not succeed in auditing all returns with DIF scores above the national cutoff. Well-staffed districts audit returns with DIF scores below the national cutoff, while other districts are unable to audit returns with scores substan- tially above the national cutoff.
As a result, some taxpayers are audited not because they have a DIF score above the national cutoff but because they file in a district that is overstaffed relative to the number of high-DIF score returns Comptroller General, , ; Wilt, The imbalance in staffing results from civil service regulations requiring uniform pay throughout the United States, 8 IRS's policy of no forced transfers, and differential impact across districts of IRS special programs e.
For a discussion, see Comptroller General , p. Some of the confusion about whether the probability of an audit differs across taxpayers is a result of equating the DIF scoring with the probability of an audit. This does not mean that the probability of an audit, conditional on a taxpayer's report, is the same in all districts.
The lowest DIF scores audited and the probability of an audit differ markedly across districts, although the DIF scoring system does not. The systematic differ- ence in the probability of an audit across taxpayers is essential in order to determine the effect of audits on compliance.
Districts for which the crowding variable is below above 1 are those where the lowest DIF scores audited are below above the national cutoffs for most of the audit classes. The workload measure is the number of returns filed in the district divided by the number of full-time equivalent employees in the district. In order to use the IRS crowding and workload measures to identify the income-reporting equation, these variables should not affect a taxpayer's reporting decisions, apart from the effect on the probability of an audit, and should not be correlated with the error term in the income-reporting equa- tion.
The first of these requirements follows directly from any standard theoretical models of taxpayer compliance, including ours.
In these models the only effect of the administrative resources or the audit selection process is through the audit probability function. If the error term in the income-reporting equation is independently distributed across taxpayers, then neither the workload nor the crowding variable is correlated with the error term. If, however, there are regional differences in tax compliance apart from those that are explained by the sociodemographic and income variables in our model, then we might be concerned that the workload variable was correlated with the error term in the income-reporting equation.
This might be the case if the IRS were able to allocate its audit resources across districts in accordance with a consistent policy that allowed it to audit all returns with DIF scores above the nationally set level for that audit class. The sources cited above show clear evidence that the resources are not distributed in this way. Further, on the basis of data from to , Long concludes that the "introduction of TCMP compliance data did not bring about any dramatic restructuring in audit coverage--even when it disclosed regions or return classes with much lower compliance levels which were receiving less attention than more compliant groups" p.
The probability of audit was obtained from IRS administrative records. A taxpayer's probability of audit is the number of returns audited in the taxpayers audit class and district divided by the number of returns filed in that district and audit class.
For example, the probability of audit for a low income taxpayer in the Los Angeles district is the ratio of the number of low-income taxpayers audited in the Los Angeles district to the number of low-income taxpayers in the Los Angeles district. Reported income is available directly from the TCMP and is simply the total of all income reported on the taxpayer's return e.
Some have simply omitted a measure of income from their models e. It is generally conceded that all available measures of income have problems. IRS research indicates that TCMP audits miss large amounts of unreported income and so the auditor's assessment will generally be less than the "true" income. This later measure of income is referred to as Census income, although it is calculated using both TCMP and Census data. As shown in Table III, the two measures of "true" income are quite close except for high- income taxpayers, for whom the Census income is substantially below the auditors' assessment.
The low Census income measure for the high-income taxpayer groups undoubtedly arises from the way in which the Census reports income. The five-digit Census Zip Code data have the number of households in each income range, with the highest income range being open- ended. The logistic methods used for approximating the average income are dependent on the underlying distributional assumptions.
For the high-income group, though, the auditor's assessment appears to provide a better income measure. Both economic theory and empirical research suggest that the effect of income on the taxpayers report will be nonlinear.
An assumption of risk neutrality might justify using only the auditor's finding of income as the measure of true income. For this reason, the model includes both linear and quadratic terms for the two income measures. We use the TCMP auditor's assessment to measure true subtractions. The IRS and researchers e. It is, of course, possible that taxpayers falsify documentation in some instances, but TCMP examiners are considered quite adept at detecting this type of fraudulent practice. Since credits are subtracted directly from the tax liability rather than from income as are adjustments, exemptions, and deductions, we divide the credits by the taxpayers' true marginal tax rate in order to make them comparable to the other subtraction items.
We also use the TCMP data to create binaries for the taxpayers' filing status and for whether one of the taxpayers filing the return was 65 years, of age or older. The filing status binary is equal to 1 if the auditor determines that the return is a joint tax return, and the "age 65" binary is 1 if the auditor determines that at least one of the individuals filing the return is entitled to the additional exemption allowed taxpayers 65 and older.
Variables for the average age of the population and percentage of house- holds headed by females combine data from the TCMP and the Census. The average age variable is dependent on whether the tax return is allowed an "age 65" exemption.
0コメント