The international face of the hidden economy

Considerable empirical research has been undertaken in a very wide range of countries, and employing various techniques, in an attempt to obtain measures of the magnitude of the underground economy. These measures generally provide somewhat varied evidence. For example, Frey and Weck-Hanneman (1984) report that for 17 countries from the Organisation for Economic Cooperation and Development (OECD) in 1978, the size of the HE (as a percentage of GNP) varied from 4.1 percent for Japan, 8.0 percent for the United Kingdom, 8.3 percent for the United States, to 13.2 percent in the case of Sweden. Canada was assessed at the sample mean of 8.8 percent, a figure that should be compared with the 5 percent to 7 percent of GDP that Mirus and Smith (1994) estimated for Canada in 1976, and the 15 percent that they estimated for 1990. Spiro (1993) estimated the Canadian underground economy at between 8 percent and 11 percent of GDP in 1993, and other Canadian (and international) evidence may be found in the various papers in Lippert and Walker (1997). By way of summarizing the various empirical studies for Canada, Mirus and Smith conclude that "a number of studies suggest significant growth and an order of magnitude of 12 to 15 percent of GDP for Canada's underground economy when we include illegal activities but exclude barter-based transactions which our definition says should be counted" (1997: 8).

Evidence from the United States in 1970 yields a range, for the ratio of the HE to GDP, from 2.6 percent (Tanzi 1983) to 11 percent (Schneider and Pommerehne 1985), while other studies summarized by Aigner et al. (1988) report American figures in 1978 that range from 4 percent (Park 1979) to 33 percent (Feige 1982) of GNP. In contrast, Bhattacharyya (1990) estimates the HE for the United Kingdom to be 3.8 percent of GNP in 1960, peaking at 11.1 percent in 1976, and averaging about 8 percent in 1984; while a British Inland Revenue analysis reported by Chote (1995) suggests that the HE may be 6 percent to 8 percent of GDP. Schneider (1997) provides some recent comparative information for a wide range of countries, some of which is summarized in figure 1, together with the earlier (and comparable) results of Frey and Weck-Hannemann (1984).

Figure 1

There are also several surveys of the literature on measuring the Hidden Economy, including those of Blades (1982), Boeschoten and Fase (1984), Frey and Pommerehne (1982, 1984), Gaertner and Wenig (1985), Kirchgaessner (1984) and Weck (1983). These studies can be grouped loosely in the following way.

1. Tax Evasion: It may be possible to use surveys relating to taxation compliance or tax-audit data to obtain estimates of underground activity (e.g. Isachsen et al. 1982; Morgensen 1985; Giles 1997e). Though this approach generally under-estimates the size of the HE, it does provide a profile of the underground labour force.
2. National Income and Expenditure: A positive "initial discrepancy" between the expenditure and income estimates of GNP may reflect hidden activity (e.g. Macafee 1980; Park 1979). However, this approach is rather crude as such a discrepancy can be just the cumulation of various measurement errors.
3. Labour Force Participation Rates: Contini (1981), Fuà (1976) and others have estimated the size of the HE from changes in the labour force participation rate. A decline in this rate over time or a low rate relative to those in comparable economies may reflect a movement of the workforce from the measured economy into hidden activities. One weakness of this approach is that many participants in the HE also work in the measured economy, so an under-estimation of unrecorded output is likely. Tedds, in her structural modeling of the Canadian HE (1998), takes account of multiple job-holdings.
4. Currency Demand: Changes in the size of the HE can be judged from movements in the demand for currency, i.e. notes and coins in circulation (see, e.g., Tanzi 1980 and Spiro 1993). The "transactions approach" (e.g. Feige 1979) infers the size of overall economic activity from the total quantity of money. The difference between this inferred activity and observed economic activity measures the extent of the HE. There are weaknesses with this approach too, including the need for accurate measures of the total volume of transactions and an assumption of a constant transactions ratio. Other currency-demand approaches to measuring the HE include adding tax rate variables to the demand equations (e.g. Cagan 1958 for the United States; Macecish 1962 for Canada) or allowing for differential velocities of circulation in the measured and hidden sectors (e.g. Bhattacharyya 1990 for the United Kingdom).
5. Latent Variable Models: The approaches above focus on just one cause of underground economic activity and one indicator. Frey and Weck-Hannemann (1984) and Aigner et al. (1988) use "latent variable" structural modeling to measure the size of the Hidden Economy. The (unobservable) latent variable here is the extent of underground activity, perhaps expressed as a percentage of measured real GDP. The MIMIC (Multiple Indicators, Multiple Causes) model of Zellner (1970), Goldberger (1972), Jöreskog and Goldberger (1975), and Jöreskog and Sörbom (1993) allows for several "indicator" variables and several "causal" variables in forming structural relationships to "explain" the latent variable(s). This approach using the latent variable / MIMIC model is the basis for our own analysis of the HE in New Zealand, the full technical details of which are given by Giles 1997a.

Given the variety of estimation methods that have been used to measure the size of the HE, it is not surprising that the results show wide variations across countries and over time. However, when we concentrate on one or two of the more comprehensive methods, and fair comparisons are made with comparable data across different countries for a specific time-period, some interesting patterns emerge. As can be seen in figure 1, the average HE in countries belonging to the OECD seems to be about 15 percent of GDP; The HE in Canada seems to be of typical size, at least on the basis of the evidence to date. Further, in percentage terms, the size of the HE has grown over past 20 years. Schneider's (1997) figures suggest that it has tripled since the 1960s. Figure 2 illustrates the variation among measures of the size of the Canadian HE, depending on the method and data used.

Figure 2.

Why has relative size of the HE apparently grown over time in virtually every country where such studies have been undertaken? (Clearly, there are several answers to this question and a useful discussion may be found in Caragata 1998a: 71-76.) One fact is plain: in almost all of these countries the overall tax burden as measured in terms of the effective tax rate (the ratio of tax revenue to GDP, say) has grown. This is very clear when one inspects the OECD's data constructed on a comparable basis. (The latest year for which this comparison can be made is currently 1994.) Figure 3 illustrates this point for Canada and New Zealand, which both had very similar effective tax rates in 1994 even though, as described below, their statutory tax rate systems are vastly different. Canada's effective tax rate was 36.1 percent (an increase of 15.3 percent over the period shown); New Zealand's was 37.0 percent (an increase of 35.0 percent); and the OECD average was 38.4 percent (an increase of 30.2 percent).

Figure 3.

In some cases, the tax system has also grown more complex, making it more difficult for some tax-payers to comply. In New Zealand, however, just the opposite is true. Some of the tax-related highlights of the fiscal reforms in that country included the following:

1. October 1986: Introduction of the GST at a rate of 10 percent, with no exceptions, and written into the marked prices of retail goods. Simultaneously, wholesale taxes (which had been up to 20 percent) were abolished; the five-step (20 percent to 60 percent, plus a 10 percent surcharge) personal statutory tax scale was simplified to a three-step scale: 15 percent, 30 percent, 48 percent.
2. 1987 Budget: A single, flat, personal statutory tax rate was proposed but not implemented.
3. 1988 Budget: The company tax rate was reduced from 45 percent to 28 percent, and the personal tax scale was simplified to two steps, with marginal rates of 24 percent and 33 percent.
4. 1989 Budget: The land tax rate was reduced from 2 percent to 1 percent, the GST rate was raised to 12.5 percent from 1 July 1989, and the company tax was raised to 33 percent from 1 April 1989.
5. 1990 Budget: Excise taxes on automobiles and land taxes were abolished, and the GST threshold was raised from NZ$24,000 to NZ$30,000.
6. 1991 Budget: There were various major international tax and company tax reforms.
7. 1996 Budget: Reductions were made to the personal income tax rates, making them 21.5 percent on the first NZ$34,200 and 33 percent on income above that.
8. 1997 Budget: From 1 July 1998, personal income taxes are to be reduced to 19.5 percent on the first NZ$38,000, and 33 percent on income above that.

Another important factor affecting the size of the HE appears to be the tax-mix. As can be seen in figure 4, the ratio of revenue from personal income taxes to GDP has changed in different ways in different countries between 1980 and 1994. In Canada, it increased by 24.1 percent to reach 13.4 percent in the mid-1990s. In New Zealand, by contrast, this ratio fell by 18.2 percent to a value of 16.6 percent in 1994. Some of the reasons for this will be made clear below. Overall, for OECD countries, the effective personal income tax rate was 12.4 percent in 1994, up just 2.5 percent since 1980. The case of New Zealand is interesting, especially given the economic reforms in that country since early 1984. Figures 5 and 6, which exclude taxes imposed by local authorities provide a useful overview (if these taxes are included, then the tax burden in New Zealand is at its highest point in 100 years). We see that there has been a decline in personal taxes since the 1980s, and an increase in indirect taxes since the mid-1970s.

Figure 4.

Greater regulation of labour and product markets is also a factor that is generally regarded as driving the HE, though again this influence has been on the decline in New Zealand (though not significantly so in Canada) since the mid 1980s. On the other hand, the growing trends in the international labour market towards more self-employment and more multiple (part-time) jobs together with the growing rates of bankruptcy and of fraud and related criminal activities that have been documented for many countries are consistent with an expanding informal sector in the economy.

New Zealand's hidden economy and tax gap

Since 1995 we have been developing econometric models of the aggregate size of the HE in New Zealand and the changes in such activity since the late 1960s. To do this, we have developed some new modeling techniques, including combining the estimation of a currency-demand equation that allows for both hidden and measured activity in a novel way, with the estimation of a MIMIC model.

Figure 5.

This work comprises the first empirical investigation of the HE in that country and is more comprehensive than most similar studies for other countries. In particular, it employs novel data for regulatory effects and detailed taxation data. There appears to be no other research into this problem that combines both of the above types of models and none that takes explicit account of the non-stationarity of the time-series data in an appropriate manner. More recently, we have also corroborated our macro-level time-series estimates by analyzing cross-sectional data based on the tax-audit records for individual firms (see Giles 1997f). The central part of our modeling involves the estimation of a structural MIMIC model in which the HE is a latent variable. By using a range of measurable indicator variables and causal variables (see table 1), we are able to generate a predicted annual time-path for the HE from 1968 to 1994.

The responsiveness of the hidden
economy to taxes

In section 2, we noted some important patterns and trends in both the aggregate tax burden and the tax mix in New Zealand and other countries in recent years. Our own work (much of which is summarized in Caragata 1998a and 1998b) and that of several other authors from other countries supports the view that the chief causes of underground activity and tax-evasion include high and rising tax burdens, inflation, rising real disposable incomes, and declines in ethical standards. The link between the tax burden and the relative size of the HE in New Zealand has been explored by means of extensive simulations by Giles and Caragata (1996). Their results show that there is clear positive causality from Tax Revenue / GDP to HE / GDP and that, on average over their sample, just over one-half of the HE is "hard-core"--i.e. unresponsive to changes in taxation policy. Even in the limit when Tax Revenue / GDP = 0, the size of the HE is still 4.7 percent of GDP.

They find that the split between hard-core and soft-core hidden economic activity varies over the business cycle: hard-core tax evasion increases, relatively, during cyclical troughs. For example, they estimated 65 percent of evasion to be hard core in 1968 but only 39 percent in 1994. In the latter case, this meant that the hard core evasion amounted to 5.1 percent of GDP and implied a tax loss of NZ$1.2 billion. By way of support, the micro-evidence from the tax-audit records analyzed by Giles (1997e) indicates an extrapolated figure for a hard-core HE equivalent to 5 percent of GDP. The other NZ$2 billion of the tax gap in 1994 was, therefore, apparently tax-sensitive. Giles and Caragata (1996) also show that successive reductions in the effective tax rate will reduce the ratio HE/GDP, though not uniformly. The elasticity between these two ratios increases with the tax burden, a unit elasticity apparently arising at an effective tax rate of 47 percent. Table 2, part (a), provides some details for the average responses over the period from 1969 to 1994.

Figure 6

More specifically, in 1994 a 10 percent (not a 10 percentage-point) cut in the effective tax rate implies a 7.6 percent cut in the HE/GDP ratio. So, it is estimated that reducing the tax burden from 34.5 percent to 31.1 percent of GDP in New Zealand that year would have reduced the HE ratio from 11.3 percent to 10.4 percent of measured GDP. Recalling figure 6, it is also interesting to ask what would be the corresponding effects of changes in the tax mix? This is addressed (with or without changes to the overall tax burden) in further simulation work reported by Caragata and Giles (1996). They show that an increase in indirect taxes relative to direct taxes reduces the size of the HE/GDP ratio. Again, some illustrative average figures are given in table 2, part (b). There, the overall tax burden has been held constant by interchanging the shares of personal and indirect taxes, as shown in figure 6, year by year. Comparing tax burden and tax-mix effects, Caragata and Giles show that reductions in the former can have more impact than even quite substantial changes in the tax-mix. For instance, on the basis of the 1994 figures, a 10 percent uniform reduction in the tax burden is predicted to have the same effect as the rather dramatic policy change of interchanging the personal and indirect tax proportions of total tax. Perhaps the most striking implications of these simulations are that the impact of tax changes on the HE starts to decelerate at an effective tax rate of around 21 percent of GDP; and that combining tax cuts with a change towards relatively more indirect taxes enhances the reduction of the (HE/GDP) ratio.

Some lessons for the Canadian situation

Although there is no definitive way to measure the size of the HE, we are now developing more sophisticated and reliable procedures for doing so, and the modeling that underlies the preceding discussion of underground activity in New Zealand is now being applied with Canadian data by the author and others (Tedds 1998). The tentative results to date clearly support the international evidence that the relative size of the HE has been growing over the past two or three decades. It is also clear from this evidence that the extent of the tax burden is a major driving force for the HE, and that the nature of the tax-mix also matters.

As far as the actual magnitude of the HE is concerned, the evidence for Canada is varied. Tax compliance in Canada, however, seems to be around the OECD average, suggesting underground activity of the order of about 15 percent of GDP in recent years. Of course, this is only a convenient way of stating this measure for comparative purposes--it does not imply that the official figures for Canada's GDP should be adjusted upwards by 15 percent to compensate for the HE, because many of the associated activities fall outside the definition of GDP. However, on the basis of this figure of 15 percent, there may be a tax gap of approximately CDN$35 billion in Canada.

Judging by the evidence from New Zealand, and using the response rates calculated by Giles and Caragata (1996), reducing the Canadian effective tax rate from 36.1 percent to 30 percent (which, interestingly, is that of Australia) would probably reduce the HE ratio from about 15 percent of GDP to about 13 percent. Ignoring the stimulation to economic growth, and other dynamic effects (which undoubtedly are significant, but are as yet unexplored), the tax gap would then fall from 5.4 percent of GDP to 3.9 percent of GDP; or from CDN$35.5 billion to about CDN$25.5 billion. If the role of government could be reduced even further, so that the Canadian effective tax rate were brought down to 21 percent--the optimal rate calculated for New Zealand by Caragata and Giles (1996)--the HE ratio might fall to 10 percent of GDP and the tax gap could fall to 2.1 percent of GDP or CDN$13.8 billion.

These figures are tentative, of course, but work that is currently in progress will enable us to provide more definitive measures shortly. In the meantime, the detailed empirical work with the New Zealand data provides us with some very important messages regarding the connection between the size of government, high effective tax rates, and the magnitude of the informal economy. Regardless of their precise numerical details, these are lessons that no policy-maker can afford to ignore.

References

Aigner, D.J., F. Schneider, and D. Ghosh (1988). Me and My Shadow: Estimating the Size of the Hidden Economy From Time Series Data. In W.A. Barnett et al. (eds.), Dynamic Econometrics Modeling: Proceedings of the Third International Symposium in Economic Theory and Econometrics (Cambridge: Cambridge University Press): 297-334.

Bhattacharyya, D.K. (1990). An Econometric Method of Estimating the "Hidden Economy," United Kingdom (1960-1984): Estimates and Tests. Economic Journal 100: 703-17.

Blades, D. (1982). The Hidden Economy and the National Accounts. Paris: OECD.

Boeschoten, W.C., and M.M.G. Fase (1984). The Volume of Payments and the Informal Economy in the Netherlands, 1965-1982. Dordrecht: Nijhoff.

Cagan, P. (1958). The Demand for Currency Relative to the Total Money Supply. Journal of Political Economy 66: 303-29.

Caragata, P.J. (1998a). Why Are Your Taxes Too High? Wellington: Seascape Press.

------ (1998b). The Economic and Compliance Consequences of Taxation: A Report on the Health of the Tax System in New Zealand. Wellington: Inland Revenue Department.

Caragata, P.J., and D.E.A. Giles (1996). Simulating the Relationship between the Hidden Economy and the Tax Mix in New Zealand. Working Papers on the Health of the Tax System, No. 22. Wellington: Inland Revenue Department.

Chote, R. (1995). Black Economy Believed to Exceed $100bn. Financial Times (10-11 June).

Contini, B. (1981). The Second Economy of Italy. Taxing and Spending 3: 17-24.

Feige, E.L. (1979). How Big is the Irregular Economy? Challenge 22: 5-13.

------ (1982). A New Perspective on Macroeconomic Phenomena: the Theory and Measurement of the Unobserved Economy in the United States: Causes, Consequences and Implications. In M. Walker (ed.), International Burden of Government (Vancouver, BC: The Fraser Institute): 112-36.

Frey, B.S., and W.W. Pommerehne (1982). Measuring the Hidden Economy: Though There be Madness, Yet is There Method in it? In V. Tanzi (ed.), The Underground Economy in the United States and Abroad (Lexington: Heath): 3-27.

------ (1984). The Hidden Economy: State and Prospects for Measurement. Review of Income and Wealth 30: 1-23.

Frey, B.S., and H. Weck-Hannemann (1984). The Hidden Economy as an "Unobserved" Variable. European Economic Review 26: 33-53.

Fuà, G. (1976). Occupazione e Capacità Produttiva: La Realtà Italiana. Bologna: Il Mulino.

Gaertner, W., and A. Wenig, eds. (1985). The Economics of the Shadow Economy. Heidelberg: Springer-Verlag.

Giles, D.E.A.(1997a). The Hidden Economy and the Tax-Gap in New Zealand: A Latent Variable Analysis. Working Paper 97-08. Department of Economics, University of Victoria.

------ (1997b). Causality between the Measured and Underground Economies in New Zealand. Applied Economics Letters 4: 63-67.

------ (1997c). Testing for Asymmetry in the Measured and Underground Business Cycles in New Zealand. Economic Record 72: 225-32.

------ (1997d). The Hidden Economy and Tax-Evasion Prosecutions in New Zealand. Applied Economics Letters 4: 281-85.

------ (1997e). Modelling the Tax Compliance/Non-Compliance Profiles of Audited New Zealand Firms: Evidence from the ORACLE Database. Mimeograph. Department of Economics, University of Victoria.

------ (1998). The Rise and Fall of the New Zealand Underground Economy: Are the Responses Symmetric? Forthcoming in Applied Economics Letters.

Giles, D.E.A., and P.J. Caragata (1996). The Learning Path of the Hidden Economy: Tax and Growth Effects in New Zealand. Working Paper No. 21. Wellington: Inland Revenue Department.

Goldberger, A.S. (1972). Structural Equation Methods in the Social Sciences. Amsterdam: North-Holland.

Internal Revenue Service (1996). Individual Income Tax Gap Estimates for 1985, 1988 and 1992. Publication 1415 (Rev. 4/96), Doc. 96-13553. Washington, DC: Research Division, Internal Revenue Service.

Isachsen, A.J., J. Kloveland, and S. Strøm (1982). The Hidden Economy in Norway. In V. Tanzi (ed.), The Underground Economy in the United States and Abroad (Lexington: Heath): 209-31.

Jöreskog, K., and A.S. Goldberger (1975). Estimation of a Model with Multiple Indicators and Multiple Causes of a Single Latent Variable. Journal of the American Statistical Association 70: 631-39.

Jöreskog, K., and D. Sörbom (1993). LISREL 8: Structural Equation Modeling with the SIMPLIS Command Language. Chicago: Scientific Software International.

Kirchgaessner, G. (1984). Verfahren zur Erfassung des in der Schattenwirtschaft Erabeiteten Sozial Produkts. Allegemeines Statistisches Archiv 68: 378-405.

Lippert, O., and M. Walker, eds. (1997). The Underground Economy: Global Evidence of its Size and Impact. Vancouver, BC: The Fraser Institute.

Macafee, K. (1980). A Glimpse of the Hidden Economy in the National Accounts. Economic Trends 316: 81-87.

Macecish, G. (1962). Demand for Money and Taxation in Canada. Southern Economic Journal 29: 33-38.

Mirus, R., and R.S. Smith (1994). Canada's Underground Economy Revisited: Update and Critique. Canadian Public Policy 20: 235-52.

------ (1997). Canada's Underground Economy: Measurement and Implications. In O. Lippert and M. Walker (eds.), The Underground Economy: Global Evidence of its Size and Impact (Vancouver, BC: The Fraser Institute): 3-10.

Morgensen, G.V. (1985). Sort Arbejde i Danmark. Copenhagen: Nyt Nordisk Forlag.

Park, T. (1979). Reconciliation between Personal Income and Taxable Income, 1947-1977. Mimeograph. Washington, DC: Bureau of Economic Analysis.

Schneider, F. (1997). Empirical Results for the Size of the Shadow Economy of Western European Countries over Time. Discussion Paper 9710. Institut für Volkswirtschaftslehre, Linz University.

Schneider, F., and W.W. Pommerehne (1985). The Decline of Productivity Growth and the Rise of the Shadow Economy in the US. Mimeograph. University of Århus.

Spiro, P.S. (1993). Evidence of a post-GST Increase in the Underground Economy. Canadian Tax Journal 41: 247-58.

Tanzi, V. (1980). The Underground Economy in the United States: Estimates and Implications. Banca Nazionale del Lavoro 135: 427-53.

------ (1983). The Underground Economy in the United States: Annual Estimates, 1930-1980. I.M.F. Staff Papers 30: 283-305.

Tedds, L.M. (1998). Modelling the Canadian Underground Economy. Extended Essay. Department of Economics, University of Victoria.

Weck, H. (1983). Schattenwirtschaft: Eine Möglichkeit zur Einschränkung der Öffentlichen Verwaltung? Berne: Peter Lang Verlag.

Zellner, A. (1970). Estimation of Regression Relationships Containing Unobservable Variables. International Economic Review 11: 441-54.