During the past two or three years, the center of attention on the underground economy has gradually moved from the pages of newspapers into the pages of scholarly reviews. Reflecting this scholarly interest, several books have been published1 and conferences have been organized. The reason for studying the underground economy is self-evident because of the possible influence on economic policies of the distortion of official estimates of such macroeconomic variables as the national accounts, the employment rate, and the rate of inflation if the underground economy is large.2
There is still disagreement on the definition of the underground economy—and, perhaps even more, on its measurement—but progress is being made. As the present article is not intended to be a survey, it does not review that progress. Rather, using a method developed by the author a few years ago (and by now applied to a large number of countries), it presents yearly estimates for the underground economy in the United States for the period 1930–80.
Researchers have shown their resourcefulness by developing several alternative approaches to the measurement of the underground economy. Some have attempted a direct measurement of the various activities that make up this phenomenon (Internal Revenue Service (1979); Simon and Witte (1980)). Some have used questionnaires to elicit answers from persons interviewed as to whether they had participated in these activities either as buyers or as sellers (Isachsen, Klovland, and Strøm (1982); Hansson (1982)). Some have used employment statistics to attempt to estimate the number of those who were officially or unofficially unemployed, while they were actually working “underground” (Pettenati (1979); Contini (1982)). Others have attempted to estimate the underground economy from the differences between the consumption and the income side of the national accounts (Macafee (1980)); still others have utilized the information on income and consumption contained in household budget studies (Dilnot and Morris (1982)). Finally, there has been the monetary approach, of which the method used in this article is a variant.
I. The Monetary Approach: A Brief Review
The monetary approach can itself be distinguished by (a) the fixed-ratio variant; (b) the currency-denomination variant; and (c) the currency-equation variant.
The fixed-ratio variant relies on two key assumptions. First, it assumes basically that there is a monetary ratio that, except for the effect of the underground economy, would have remained constant over time. Second, it assumes that there was a golden period in the past when no underground economy existed. This monetary ratio is C/D for Guttman (1977) and MV/GNP for Feige (1979), where C is currency in circulation, D is demand deposits, M is money supply (made up of currency and demand deposits), V is transactions (not income) velocity of money,3 and GNP is the officially estimated gross national product. Through the changes over time in these ratios, Guttman and Feige estimate the size of the underground economy. In Guttman’s method the estimation is indirect, in the sense that he first calculates “excessive” currency in circulation (induced by the existence of underground activities) and then, by making assumptions about its in come velocity, estimates the underground economy. In Feige’s method the estimation is direct, as the change in the MV/GNP ratio over time allows a direct estimation of underground GNP. The golden period, when the underground economy was zero, is the late 1930s for Guttman and is either the late 1930s or the mid-1960s for Feige.
The currency-denomination variant is based on the assumption that underground economic activities are associated mainly with the use of bills of certain denominations. In the United States, bills of $100 and above have been considered to be used mainly, and serious proposals have been made to withdraw them from circulation (Henry (1976)). Thus, by analyzing the change in the composition of total currency holdings in circulation, some analysts have attempted to measure the underground economy (Henry (1975)).
The fixed-ratio approach, regardless of whether it is used in Guttman’s version or in Feige’s (more sophisticated) version, is open to the criticism that there is no obvious reason why those ratios should remain constant over long periods. Although, at least for the United States, there may be “great ratios” in the economy that, for some poorly understood reasons, remain relatively unchanged (Klein and Kosobud (1961)), it is not clear why Guttman’s and Feige’s ratios should be among them. Another criticism is that the results are sensitive to the choice of the initial period. For these two reasons, there has been considerable reluctance to accept the high estimates of the underground economy that have been derived by Guttman (1977) and the even higher ones that have been derived by Feige (1979).
The currency-denomination approach is open to the criticism that—with real growth and relatively high rates of inflation—it would be expected that the composition of currency holdings would change over the years. Over time, large bills do not appear to be as large as they once were. When the normal daily wage was a few dollars, a $100 bill was a small fortune; but today, when the wage of many workers exceeds $100 a day, $100 no longer reflects a large purchasing power. Furthermore, as the author has argued elsewhere, many $100 bills are probably being used for transactions or for storing wealth in foreign countries. An unknown but probably high amount of currency substitution (with U.S. currency replacing the currencies of many countries) has certainly been taking place over the years. This substitution, or “dollarization,” is likely to have been particularly significant in countries that are experiencing high rates of inflation.4
Before moving to the more substantive part of this paper, the author has provided, in a summary fashion, some of the basic data that have been utilized in the monetary approach.
Chart 1 depicts the behavior of total currency on a per capita basis for the period 1933–80. It shows that a remarkable increase took place over the period and that, by 1980, there was almost $700 of currency outstanding for each person living in the United States. The increase was particularly significant in the years 1941–45 and the period after the mid-1960s. Chart 1 also shows the behavior of currency in denominations smaller than $100, equal to $100, and greater than $100. The remarkable increase in per capita currency, combined with the high per capita figure of currency in circulation and the large increase in $100 bills, has given credence to claims about a phenomenal growth in underground economic activities.
Chart 1.U.S. Per Capita Currency in Circulation, Expressed in Current Prices, 1933–80
Chart 2, however, gives a somewhat modified picture simply by adjusting for changes in prices. In fact, when per capita currency in circulation is expressed in 1972 prices, it appears that, despite the substantial growth in real income since World War II, per capita currency was much higher during World War II than it was in 1980. Between that early period and the early 1960s, real per capita holdings of currency fell by almost one half. The holdings then started rising again, with the increase attributed mostly to $100 bills. But as a proportion of per capita income in real terms, the holdings have continued to fall up to the present time.
Chart 2.U.S. Per Capita Currency in Circulation, Expressed in Current Prices, 1933–80
In addition to total currency per capita and to total $100-denominated currency per capita, the ratio that has attracted most attention in connection with underground economic activity is that of currency to demand deposits. That ratio, shown as C/D in Chart 3, has fluctuated considerably over the years. It rose sharply during the Great Depression, when eroding confidence in the banking system induced individuals to substitute currency for demand deposits, and it increased sharply during World War II for a variety of reasons detailed in Cagan’s well-known study (1958). From 1945 to the early 1960s, the C/D ratio fell but it then rose again. The behavior of the C/D ratio indicates that (a) this ratio can hardly be taken as constant and (b) if one bases estimates of the underground economy on that ratio, the results obtained will be highly influenced by the beginning period.
Chart 3.Ratio of U.S. Currency in Circulation to Demand Deposits and toM1andM2, 1930–80
Chart 3 also shows that, while the C/D ratio fell between 1945 and 1961 and then rose, the C/M2 ratio fell continuously from 1945 to the early 1970s and then remained roughly constant. This pattern seems to indicate that the increase in the C/D ratio during the past two decades may be due to shifts by individuals from checking accounts into other financial assets (included in M2) that paid interest rather than to sharp increases in currency holdings. In other words, the C/D ratio may have risen because of a fall in D rather than an increase in C. (See also Garcia and Pak (1979).) This conclusion is strongly supported by Chart 4, in which various measures of the money supply are shown as percentages of GNP. This figure shows a substantial fall in D/GNP, while the fall in C/GNP appears to be quite gradual over the post-World War II period.
Chart 4.Various Measures of U.S. Money Supply as Proportion of Gross National Product, 1930–80
II. Description of Method and Data
The method that is used here to derive the yearly estimates of the underground economy in the United States has been explained in detail in an earlier study in which estimates for 1976 were derived.5 The approach consists in specifying a demand-for-currency equation to be able to infer the effect of a change in the tax level on that demand. The key assumptions are, first, that underground economic activities are the direct consequence of high taxes,6 and, second, that currency is used mainly for carrying out such transactions or for storing wealth. In the equation, the ratio of currency holdings (C) to money (defined as M2) is the dependent variable. The independent variables are real per capita income (Y), the rate of interest paid on time deposits (R), the ratio of wages and salaries in national income (WS/NI), and an income tax variable (T). Thus,
where ϵ is an error term.
Once equation (1) has been estimated for the period 1930–80, it can be utilized to estimate currency holdings by making the assumption that the tax variable assumes a value of zero. Once currency holdings at zero taxes are estimated, they can be used to determine the extent of the underground economy by multiplying excessive currency by the income velocity of money, as discussed later in more detail.
The expected sign for both per capita income and the interest rate is negative, while the expected sign for both taxes and the ratio of wages and salaries in national income is positive.
A detailed discussion of why these signs are expected can be found in Tanzi (1980 b). Briefly, as the level of taxation rises, individuals are encouraged to engage in tax-evading activities that are facilitated by the use of currency, as this practice leaves no traces; therefore, the use of currency rises. Because wages are often paid in currency, especially for daily workers, and other types of income (interest, dividends, etc.) are almost always paid by check, an increase in the ratio of wages in total income paid will require more currency. On the other hand, economic development, as proxied by per capita real income, is assumed to lead to the replacement of currency by checks, thus leading to a fall in C/M2. Equally, as an increase in the rate of interest increases the opportunity cost of currency holdings, C/M2 and R can be expected to be negatively correlated because M2 includes time deposits.
The data used are presented in Table 1, which covers the period 1929–80, with currency and M2 based on June 30 of each year. R is the interest rate paid on time and savings deposits. TW and T are two alternative measures of the tax variable: TW is a weighted average tax rate on interest income and is described in detail in Tanzi (1980 b); T is the ratio of total income tax payments after credit to adjusted gross income.
Currency and M2 are based on June 30 each year.
R = interest rate paid on time and savings deposits.
TW = weighted average tax rate.
T = income tax after credit over adjusted gross income.
Currency and M2 are based on June 30 each year.
R = interest rate paid on time and savings deposits.
TW = weighted average tax rate.
T = income tax after credit over adjusted gross income.
Using the data shown in Table 1, equation (1) was solved alternatively with both measures of the tax variable. A logarithmic specification has been used; given that specification, to solve the equations for a situation in which the tax variable is assumed to be zero, that variable has been replaced in each case by the yearly figure shown in the table plus one. In other words, T and TW have been replaced for each year by 1 + T and 1 + TW, respectively.
The estimated equations are as follows:7
In both cases, the
The estimates of the underground economy are derived as follows. For each year, the predicted level of the currency ratio C/M2 can be calculated by using the preceding regression equations. Then, given the actual figure of the M2 for that year, the predicted level of currency holdings (Ĉ) can be calculated.8 Next, the equations are solved in the same way, assuming that the tax variable is zero while the coefficients of the other variables remain unchanged. The resulting value of currency is then defined as
The difference between predicted currency holdings (Ĉ) and the currency holdings predicted from the equation under the assumption of zero taxes (
Chart 5 shows (a) actual currency holdings for the period 1930–80, C; (b) currency holdings predicted by equations (1′) and (1″), Ĉ1 and Ĉ2, respectively; (c) currency holdings, when taxes are assumed to fall to zero, predicted by equations (1′) and (1″),
Chart 5.Actual and Predicted Values of U.S. Currency Holdings Under Alternative Average Tax Rates, 1930–80
|Year||C||Ĉ||C – Ĉ|
|Year||C||Ĉ||C – Ĉ||Ĉ – |
The yearly estimates of the underground economy are shown in Tables 4 and 5. These two tables again use the weighted average tax rate (Table 4) and the average tax rate (Table 5). The first column of these tables is equivalent to the last column of Tables 2 and 3 (the estimates of illegal money). Multiplying column (1) by column (3), which yields the income velocity of money, produces the estimates of the underground economy, shown in column (4). These estimates are also given as proportions of GNP in column (5).
Table 4 shows that the underground economy grew from about $45 billion in 1970 to more than $159 billion in 1980. The figure for 1976 is remarkably close to the well-known estimate of $100 billion by the Internal Revenue Service (1979). Column (5) shows that, as a porportion of GNP, the underground economy exceeded 6 percent in 1980. The estimate for 1974 ($71 billion) is in the range of the other direct estimate made by Simon and Witte (1980, p. 74) of between $70 billion and $75 billion of income from legal sources. Perhaps even more significant is the fact that a clear upward trend has been noticeable since the mid-1960s, when the underground economy averaged about 3.8 percent of GNP. This trend, while not as dramatic as that reported by some news accounts or as estimated by Feige (1979), is nevertheless disturbing, especially as it seems to have accelerated in recent years. As a proportion of GNP, the underground economy reached its peak in 1944, when it was close to 8 percent. This figure is consistent with results reported by other researchers (for example, Cagan (1958) and Molefsky (1982)). From 1944 to the mid-1960s, the share of the underground economy in GNP fell by one half. Table 5 presents the same categories of results but is calculated by using the average tax rate rather than the weighted average tax rate.
The results from Tables 4 and 5 are presented in Chart 6, which shows that the estimates based on the average tax rate are somewhat lower than the others. In 1980 the two estimates varied by about ½ percent of GNP. In general, the behavior of the two lines in Chart 6 is quite similar, except that one is consistently above the other. The only period when the two lines diverge in behavior is during World War II. In fact, while the use of the weighted average tax rate gives the highest figure for the underground economy in 1944, the use of the average tax rate gives the highest figure in 1942.
Chart 6.Size of U.S. Underground Economy Under Alternative Average Tax Rates, 1930–80
Tables 4 and 5 also show estimates of income tax evasion—column (6). For 1980, income tax evasion was estimated at about $15 billion in Table 4 and at about $11 billion in Table 5. These figures were derived by assuming that incomes in the underground economy would have been taxed at the same average rate as incomes in the regular economy. This is, of course, a strong assumption. If incomes in the underground economy had been marginal incomes and thus would have been taxed at marginal tax rates, the estimates of tax evasion would have been much higher. This is undoubtedly true for many providers of underground services. They may hold regular jobs, on which they pay normal taxes, and may provide underground services during weekends or evenings. Therefore, if they had declared these additional incomes, they would have paid higher taxes than they actually paid on their reported incomes. However, many of the participants in underground economic activities may not declare any income at all, and some (for example, illegal aliens) might pay taxes that are lower than the average. Furthermore, some of these activities might not take place at all if they could be controlled and taxed. In conclusion, without ascertaining each activity and determining whether the activity would exist if it had to move above ground, it is impossible to establish what the precise loss in tax revenue from the underground activities would be. Therefore, the assumption that has been made probably yields the most reasonable estimates that can be made in view of the information available.10
IV. Concluding Remarks
In the present article, yearly estimates for the underground economy for the period 1930–80 have been calculated. The main conclusions can be summarized as follows: (1) In 1980 the underground economy, expressed as a percentage of GNP, was somewhere between 4.5 percent and 6.1 percent. (2) The only other period in which it may have been higher than that was during World War II. (3) It has been increasing since the mid-1960s. (4) Over the period 1965–80, it has grown by more than 2 percentage points of GNP, or by almost 50 percent. The recent trend is disturbing as it seems to have accelerated in recent years, especially since the mid-1970s. This trend was probably influenced by the substantial increase in marginal tax rates over the period 1975–80, caused by inflation and the absence of significant tax cuts.11 It is not possible at this point to speculate on whether the tax cuts enacted in 1981 have reversed the trend.
The results obtained in this paper should not be taken as precise measures of the underground economy; they are, at best, broad indications of trends and of orders of magnitude because they are sensitive to the assumptions made as well as to data revisions. However, it is comforting to realize that for 1974 and 1976 they are of the same order of magnitude as direct estimates made by Simon and Witte (1980) and by the Internal Revenue Service (1979).
Perhaps a word on what has been measured here is necessary. The estimates attempt to measure the incomes that were generated through the excessive use of currency and that presumably were not reported to the tax authorities. Whether these incomes were or were not measured by the national accounts authorities cannot be determined. Presumably, part of these incomes not only evaded the tax net but also may have escaped the attention of the national accounts authorities—but, how large this part was cannot be assessed with the information at hand. It should also be emphasized that this measure of the underground economy does not measure “incomes” from criminal activities. But, as noted earlier, if these incomes were discovered, they would vanish and the “legal” incomes that would be generated by the resources so released would probably be small.12 It is then not clear to what extent those illegal incomes represent a loss to the tax authorities or whether they should be measured in the national accounts.
BladesDerek“The Hidden Economy and the National Accounts,”OECD Economic Outlook: Occasional Studies (ParisJune1982) pp. 28–45.
BurkettLarry and WilliamProctorHow to Prosper in the Underground Economy: A Completely Legal Guide to the Hidden Multibillion-Dollar Cash Economy (New York1982).
CaganPhillipThe Demand for Currency Relative to Total Money Supply National Bureau of Economic Research Occasional Paper 62 (New York1958).
ClotfelterCharles T.“Tax Evasion and Tax Rates: An Analysis of Individual Returns” (unpublished1982). (This article is scheduled for publication in Review of Economics and Statistics.)
ContiniBruno“The Second Economy of Italy,”Ch. 12 in The Underground Economy in the United States and Abroaded. byVitoTanzi (Lexington, Massachusetts1982) pp. 199–208.
De GraziaRaffaele“Clandestine Employment: A Problem of Our Times,”International Labour Review Vol. 119 (September/October1980) pp. 549–63.
DilnotAndrew and C.N.Morris“What Do We Know About the Black Economy in the United Kingdom?”Ch. 10 in The Underground Economy in the United States and Abroaded. byVitoTanzi (Lexington, Massachusetts1982) pp. 163–79.
FeigeEdgar L.“How Big Is the Irregular Economy?”Challenge Vol. 22 (November/December1979) pp. 5–13.
GarciaGillian and SimonPak“The Ratio of Currency to Demand Deposits in the United States,”Journal of Finance Vol. 34 (June1979) pp. 703–15.
GuttmanPeter M.“Subterranean Economy,”Financial Analysts Journal Vol. 33 (November/December1977) pp. 26–27 34.
HanssonIngemar“The Underground Economy in a High Tax Country: The Case of Sweden,”Ch. 14 in The Underground Economy in the United States and Abroaded. byVitoTanzi (Lexington, Massachusetts1982) pp. 233–43.
HenryJames S. (1975) “Cash and Crime: A Modest Proposal” (unpublishedDecember1975).
HenryJames S. (1976) “Calling in the Big Bills,”Washington Monthly Vol. 8 (May1976) pp. 27–33.
Internal Revenue ServiceEstimates of Income Unreported on Individual Income Tax Returns Publication No. 1104(9–79) U.S. Department of the Treasury (WashingtonSeptember1979).
IsachsenArne JonJan ToreKlovland and SteinarStrøm“The Hidden Economy in Norway,”Ch. 13 in The Underground Economy in the United States and Abroaded. byVitoTanzi (Lexington, Massachusetts1982) pp. 209–31.
KleinLawrence R. and Richard F.Kosobud“Some Econometrics of Growth: Great Ratios of Economics,”Quarterly Journal of Economics Vol. 75 (May1961) pp. 173–98.
MacafeeKerrick“A Glimpse of the Hidden Economy in the National Accounts,”Economic Trends No. 316 (February1980) pp. 81–87.
MolefskyBarry“America’s Underground Economy,”Ch. 3 in The Underground Economy in the United States and Abroaded. byVitoTanzi (Lexington, Massachusetts1982) pp. 47–67.
PettenatiPaolo“Illegal and Unrecorded Employment in Italy,”Monte dei Paschi di Siena Economic Notes Vol. 8 (No. 11979) pp. 14–30.
SimonCarl P. and Ann D.Witte (1980) “The Underground Economy: Estimates of Size, Structure, and Trends,”in Government Regulation: Achieving Social and Economic Balance Vol. 5Special Study on Economic Change, Joint Economic Committee (96th Congress, 2nd SessionWashingtonDecember81980) pp. 70–120.
SimonCarl P. and Ann D.Witte (1982) Beating the System: The Underground Economy (Boston1982).
SmithAdrian“A Review of the Informal Economy in the European Community” Commission of the European Communities Economic Papers No. 3 (BrusselsJuly1981).
TanziVito (1979) “Underground Economy Income Tax Evasion and the Demand for Currency in the United States 1929–76” (unpublishedInternational Monetary FundNovember51979).
TanziVito (1980 a) “Underground Economy Built on Illicit Pursuits Is Growing Concern of Economic Policymakers,”IMF Survey Vol. 9 (February41980) pp. 34–37.
TanziVito (1980 b) “The Underground Economy in the United States: Estimates and Implications,”Banca Nazionale del Lavoro Quarterly Review No. 135 (December1980) pp. 427–53.
Mr. Tanzi, Director of the Fiscal Affairs Department, holds a doctorate in economics from Harvard University. He was formerly a professor and Chairman of the Economics Department at American University.
International organizations have also started paying attention to the underground economy—obviously an important phenomenon. See, for example, De Grazia (1980), Tanzi (1980 a), Smith (1981), and Blades (1982).
In the actual calculations, the velocity of currency and of demand deposits are determined separately.
For a more detailed discussion of this point, see the author’s paper, “A Second (and More Skeptical) Look at the Underground Economy in the United States,” Ch. 6 in Tanzi (1982), pp. 103–18.
That method was first suggested and utilized in an unpublished paper (Tanzi (1979)); a substantially revised version was published in Tanzi (1980 b). Of related interest are also Cagan (1958) and Henry (1975). Henry’s paper was written while he was a graduate student at Harvard University; appreciation is due to him for bringing the paper to the author’s attention.
Obviously, some underground activities are not the result of taxes; therefore, estimates will not capture these activities. Convention, however, excludes the income generated from illegal activities from national accounts; it is likely that this income would vanish if it could be discovered and taxed. It should also be realized that the opportunity cost of the resources used for these activities is likely to be very low.
The equations have been corrected with a first-order Cochrane-Orcutt correction for serial correlation. The coefficients in parentheses are t-values. Two asterisks indicate significance at the 1 percent level; one asterisk indicates significance at the 5 percent level.
See Tanzi (1980 b) for the detailed procedure.
This is clearly a key and controversial assumption. See Tanzi (1980 b) for more details.
These estimates of tax evasion relate only to tax evasion that is associated with currency use and underground activities. But, obviously, many forms of tax evasion (claiming nonexistent exemptions, exaggerating deductions, nonreporting of interest income received, etc.) have nothing to do with currency usage or underground economic activities. Thus, these estimates do not measure total tax evasion.
For a recent paper on the relation between high marginal tax rates and tax evasion, see Clotfelter (1982).
For example, a drug smuggler who makes a large gain from his criminal activity would see his “income” dramatically reduced if (a) he were caught and jailed or forced to get a legal job, (b) if he gave up his activity because of improved law enforcement, or (c) if drug consumption were legalized. Therefore, these criminal incomes are totally arbitrary ones. In the author’s view, it makes no sense to include them in a definition of underground economy that emphasizes either potentially recoverable tax revenue or the use of resources for activities that should be measured by the national accounts.