There is an extensive literature addressing the determinants of tax evasion behavior. Aside from a variety of principally theoretical models of tax evasion behavior (Falkinger, 1988; Allingham and Sandmo, 1972; Klepper, Nagin, and Spurr, 1991; Das-Gupta, 1994; Pestieau, Possen, and Slutsky, 1994), there are a number of studies of such behavior using (a) questionnaires or experiments (Spicer and Lundstedt, 1976; Friedland, 1982; Spicer and Thomas, 1982; Benjamini and Maital, 1985; Alm, Jackson, and McGee, 1992; Baldry, 1987; De Juan, 1989; Thurman, 1991), or, in a few cases, (b) what De Juan, Lasheras, and Mayo (1994) refer to as "official data" (Clotfelter, 1983; Slemrod, 1985; Pommerehne and Weck-Hannemann, 1989; Erard and Feinstein, 1994). Indeed, the issue of the size of the underground economy, which consists essentially of economic transactions (or income) that are not reported to the government tax-collection authority, was the subject of an entire recent issue of the journal Public Finance/Finances Publiques (Supplement to Volume 49/1994).
It is generally accepted that the size of the underground economy may be affected by income tax rates (Clotfelter 1983; Slemrod, 1985; Pommerehne and Weck-Hannemann, 1989). Clearly, the higher the pertinent marginal tax rate, the greater the benefit (in terms of a reduced tax liability) from not reporting taxable income. It is also commonly found that the greater the risk associated with participating in the underground economy, the less the degree to which economic agents will choose to either not report, or to underreport their taxable income [Friedland (1982), Spicer and Thomas (1985), De Juan (1989), Alm, Jackson, and McKee (1992)].
The contribution of this study is fundamentally empirical. In particular, based on new, improved data (available directly from Professor Edgar Feige) through 1994 on the size of the underground economy in the United States, this study empirically seeks to provide new, updated insight into the determinants of the size of the underground economy. Relying on tax evasion theory [Pestieau, Possen, and Slutsky (1994, 20)], the analysis uses the Feige data and "official data" from the Internal Revenue Service (IRS) to determine the impact of (1) government-tax-rate policies, (2) IRS audit probabilities, and (3) IRS penalty assessments on the relative size of the underground economy in the United States.
In this study, income-tax-rate policies are principally reflected by maximum marginal federal personal income tax rates and federal corporate income tax rates. In addition, unlike most previous studies, in three of the six estimates provided in this study, measures of the social security tax rate are also included. To reflect the probability of IRS audits, past actual audit rate data are used. Finally, unlike in previous related studies, to reflect IRS penalties, here actual IRS data is used that indicate the total penalty assessed by the IRS per dollar of adjusted gross income.
Let the economy consist of n economic agents. These economic agents generate economic value: income, and have a choice as to whether or not to report their income to the tax-collecting authority. To the extent that said income is reported to the tax-collecting authority, a tax liability is incurred.
The relative probability that the representative economic agent will not report its taxable income to the tax authority is an increasing function of the expected gross benefit to the agent of not reporting income, eb, and a decreasing function of the expected gross costs to the agent of not reporting income, ec. Thus, the ratio of the probability of not reporting income, pnr, to the probability of reporting income, (1 - pnr), is described for the representative economic agent by:
pnr/(1 - pnr) = f(eb, ec), [f.sub.eb] [greater than] 0, [f.sub.ec] [less than] 0 
In turn, the expected gross benefits from not reporting income are anticipated to be an increasing function of the income tax rate (Cagan, 1958; Bawley, 1982; Tanzi, 1982 and 1983; Clotfelter, 1983; Slemrod, 1985; Pyle, 1989). …