Fiscal Illusion and Consumer Sovereignty: an Exploratory Study

Werner W. Pommerehne, University of Zurich
ABSTRACT - Data on the Australian federal government's performance as perceived by the voters and government revenue and expenditure variables are used to estimate a model of interaction between the voters' and the government's behavior. Special emphasis is given to the role of fiscal illusion on the voters' side in evaluating government performance, and to its manipulation by the government toward securing reelection. The model is developed and the empirical results are interpreted under the perspective of consumer sovereignty.
[ to cite ]:
Werner W. Pommerehne (1981) ,"Fiscal Illusion and Consumer Sovereignty: an Exploratory Study", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 539-544.

Advances in Consumer Research Volume 8, 1981      Pages 539-544


Werner W. Pommerehne, University of Zurich

[The author would like to thank Friedrich Schneider for his most helpful comments, and Sandra Stuber for editing the paper for English.]


Data on the Australian federal government's performance as perceived by the voters and government revenue and expenditure variables are used to estimate a model of interaction between the voters' and the government's behavior. Special emphasis is given to the role of fiscal illusion on the voters' side in evaluating government performance, and to its manipulation by the government toward securing reelection. The model is developed and the empirical results are interpreted under the perspective of consumer sovereignty.


Consumer sovereignty refers to a process of choice in which the choice is free (most major characteristic), informed (so that the act of choice is meaningful), personal (made by and not for people), and responsible (choosers must be aware of the consequences of their choice). It is often said that these preconditions for consumer sovereignty are not fulfilled in the marketplace. Consumer advocates argue that buyers are faced with high information costs, offensive and often misleading advertising, and shoddy and unsafe goods. They have sharply attacked these obstacles to consumer choice, also criticizing the whole market economy for, among other things, the monopolistic practices applied by many producers and suppliers.

However, as was pointed out by Roland McKean (1973) in his address to the Southern Economic Association, the government sector has for the most part been ignored in these analyses and attacks, though both goods and services which are provided by market and by the government are relevant to the consumer's well-being. As he further says (1973, p. 285), "I conjecture that consumers lose more from monopoly power and non-profit incentives in the provision of most government services than from monopoly power in activities of similar magnitudes in the private economy." Under the very best of circumstances, which are rarely attainable, government sector output may approach the market in the extent to which the suppliers adapt to the consumers' desires. In reality, however, the supply side push may dominate the forces of the demand side pull in terms of the structure and pattern of public sector output, this for several reasons (see e.g. Diamond 1970, Breton 1974, Wagner 1978, Niskanen 1979): Though in a democracy there are at least two parties competing at the electoral level for the support of the voters, following an election there is only one majority for the duration of the legislative period and thus the winning party has a legal monopoly. Moreover, politicians and parties cannot be held legally liable for their promises, all the more so as the particular commodities promised are usually not very tangible entities but only vague notions. In addition, there is seldom any other recourse possible than changing one's vote at the next election. There is rarely any alternative to the government-provided goods and services; private options are often legally prohibited. Last but not least, as long as the public services are financed through taxes (instead of through user charges), there is no free choice. Richard Wagner (1978, p. 95) concludes from this: "Whereas the natural state of the private sector seems to be competition, the natural state of the public sector in a democratic society appears to be monopolistic." Thus under a perspective of consumer sovereignty one would expect that consumer research, when extended to the government sphere, would have a large (and growing) area for analysis in the near future.

Though a monopolistic government has discretionary power which it can use for achieving its own ideological goals, it cannot do whatever it wishes. In the same way as a monopolist on the market is faced with the constraint of a decreasing demand curve, a monopolistic government is subject to constraints, of which the most important one in a democracy is the need to be reelected. One way for it to improve its reelection chances is to exploit fiscal illusions held by the voters. Fiscal illusion as used here is taken to mean the systematic misperception by voters of the public revenue burden they bear and the benefits they receive through government spending. It can be seen as a kind of rational ignorance in consumer choice, going back to the different natures of the various public revenue extraction items and the public services, and thus to their different information costs. It also relates to the only minor extent to which the consumers can influence the output of the public sector as compared to that of the market sector, and the week incentives for informing oneself on the public policy under discussion and its financing that result therefrom.

In the following fiscal illusion will be analyzed in the framework of a model of interaction between the voters and the government [The general framework used here was developed by Frey and Lau (1968). Its usefulness for empirical research has been demonstrated by a series of applications, see Pommerehne, Schneider and Lafay (1980) for a recent survey.] with Australia being taken as the concrete example, mainly because its political framework at the federal level is highly suitable for our analysis. [The federal government controls over 70% of all public revenues and expenditures, and also has the power to quickly make discretionary changes on the revenue side and, to a lesser extent, in public expenditures. For more details see Schneider and Pommerehne (1980a).]


It is assumed that, in evaluating the government's performance, voters maximize their own utility: i.e., the extent of their support for the governing party - as measured by current popularity or by voting for the government at election time - depends on how satisfied they are with the government. Because of the high costs and low benefits involved, the consumer/voter in general has little or even no incentive for becoming fully informed about the government's performance. This is true with regard to informing oneself not only about one's present and future burden of government receipts and the benefits received through public spending programs, but also about the past, which may be discounted by the voter.

Besides this, one can expect that the information on the various revenue items which is absorbed on a day-to-day basis may be biased in a systematic way because of the different costs involved in becoming informed. These information costs are dependent on the revenue extraction system's degree of fragmentation. A three hundred year old tradition in public finance suggests that the more complex the mix of public revenue sources is, the more the individual will tend to underestimate the true burden of the cost of government. A series of small public receipts may be felt less than a single large tax, particularly if it is extended over time since small losses are more easily forgotten than large ones. Moreover, because of the varying degrees of visibility, of timing of extraction, and of hidden shifts in the revenue burden, it can be concluded that voters/taxpayers will perceive certain public revenue items (such as indirect taxes and public debt incurred) less than others (such as e.g. a pay-as-you-earn income tax). All this opens up opportunities for the government to install extraction institutions that will ceteris paribus decrease the perceived cost of government and thus favorably influence the voter's evaluation of the government's performance (Pommerehne and Schneider 1978).

Voters may also be subject to systematic misperception of the benefits of public spending. These also involve positive information costs, which - due to the nature of the public goods and services - are often much higher than in the case of market goods. Many public goods have "credence" qualities (Darby and Karni 1973) which are characteristic of goods and services that are utilized in combination with other goods composed of uncertain properties. Because the quality of many public goods is difficult to judge even after provision, their evaluation may require a greater degree of costly expertise. However, it seems to be very difficult to derive a hypothesis as to what kinds of individually perceived public services will be systematically more or less strongly felt. This problem arises because of the very great difficulty in discriminating between actual and perceived benefits. Whereas the cost of government activities to a taxpayer via taxes and other forms of fiscal extraction is, to a large extent, identifiable - so that if the perceived and actual costs can be reasonably determined, the difference can be attributed to fiscal illusion - this does not hold for the spending side. For this reason the following analysis deals mostly with taxes and other forms of government finance, and does not try to distinguish between preferences and favorably biased perception when government spending is considered.

It is also assumed that voters rarely draw a close connection between public services and their actual financing sources. This assumption may be valid as the principle of non-affectation is used at the federal level. The link between federal revenue and expenditure is further obscured by the government's ability to incur debt.

As a measure of how Australian voters/taxpayers perceive and evaluate the government's overall performance, a monthly Gallup Poll was used showing the percentage of the citizenry "willing to vote for the Australian government at a federal election" at that point. This data series for current government popularity (POP) gives us the dependent variable. Turning to the independent variables, first the revenue system's fragmentation is measured by nE/i=1 (REVi)2, where REVi equals the i-th revenue amount for all federal revenue items. This measure of complexity moves closer to 1 the more the government relies on only one source of income; it is 1/n if it uses all n revenue sources equally (each accounting for 1/n of the total). In order to capture the influence of the different items on the public revenue side, the various fiscal extraction devices are grouped into two classes of direct taxes (a pay-as-you-earn income tax and assessed income tax, and a tax on dividends and interest); one class of indirect taxes (sales tax, company taxes, license fees); one of customs and various duties; and one of government debt incurred. All these variables are measured as shares of total government revenue. The various spending items are also grouped into classes: transfers (social security and welfare payments to individuals, payments to disabled people, and retraining programs); health programs, education, general and scientific research, and cultural and recreation measures; investments in transportation and communication, water and electricity supply, other public utilities, and investment grants to the states; public administration, law and order, public safety, and legislative services; and defense expenditure. These variables are all measured as shares of total expenditure. Finally, the ratio of total federal revenue to GNP is introduced to take into account perception of the increase in the overall federal revenue burden that took place during the 1970s.

Government's overall performance, as perceived by the voters, relates however not only to the provision of public goods and their financing, but also to its ability to steer the economy. This aspect, i.e. that the federal government is held responsible for the state of the economy, is captured by using a monthly Gallup survey taken since January 1970 which asks "Are you satisfied (or not) with the current economic performance of the government?" as an additional independent variable. [Perceived government economic performance is again influenced by popular indicators of the current and past states of the economy, such as the rates of unemployment and of inflation, which can be easily culled from day-to-day sources; for a detailed analysis of this relationship see Schneider and Pommerehne (1980b).]

For government popularity a negative sign and a significant coefficient are expected for the ratio of total federal revenue to GNP, and for the measure of the complexity of the whole revenue system (i.e., the lower S(REVi)2 is, the more the government relies on all revenue items alike and thus the higher is ceteris paribus the current popularity standing). Negative signs are also expected for those individual revenue shares which are strongly perceived. Similarly on the spending side, a positive sign is expected for those individual expenditure items which are favorably perceived by a large part of the voters. A positive sign is also expected for the perceived government economic performance (i.e. the better the perceived ability of the government to steer the economy, the more voters tend ceteris paribus to support the governing party). All data are monthly and a three-month lag is taken for all independent variables as it is assumed that the citizen needs some time to notice a change in taxing and spending (using weighted lags up to one year, no major significant changes occurred in the following estimation results). To test how good the voters' memory actually is, it is assumed for simplicity's sake that it can be measured by a lagged endogenous variable and that the same size of discount rate holds for all independent variables (i.e. a Koyck transformation is used). The results for the GLS estimation of the government popularity function for the period 1971.1 through 1977.9 is given in equation 1. [More exactly, both the equation for explaining perceived economic performance (fn. 4) and that for government's popularity are simultaneously estimated by GLS technique. This also explains why the variable "Government economic performance'' in equation 1 is not lagged.]


d.f. = 62, R-2 = 0.97, h = 1.27.  The figures in parentheses below the parameter estimates indicate the t-values (one asterisk indicates statistical significance at the 95%, two asterisks at the 99% confidence level, using a two-tailed test); d.f. shows the degree of freedom; R2 is the corrected coefficient of determination; and h indicates the test statistics for autocorrelation.

The results show first that citizens do indeed seem to highly discount changes in the revenue and spending structures: The coefficient of the lagged endogenous variable indicates that over 95% of what happens is forgotten within the space of one year. Secondly, the government's economic performance as perceived by the voters does influence the government's overall performance (popularity) in a positive and significant way.

As expected, total revenue burden and the complexity of government's revenue system have both a negative and strongly significant influence on current government popularity. The marginal impact of these variables is also quantitatively important: An increase in, for example, total revenues as a share of GNP by 1 percentage point ceteris paribus reduces current popularity by 1.21 percentage points; in contrast, the marginal decrease in government economic performance of 1 percentage point is only about one-third of this (0.38 percentage point). Going to the individual revenue items, one only finds a highly significant influence on government's current popularity in the case of the two direct taxes. This result is fully in line with the classic argument of public finance literature that direct taxes are most strongly felt, whereas indirect taxes and most other revenue items (especially government debt) are much harder to detect. Moreover, a stronger dampening impact on government support of direct (income) as compared to more indirect (sales) taxes has also been found recently for other countries, including the United States (Eismeier 1979). Looking at the spending side, one finds a significant positive marginal effect on government popularity only in the cases of transfer payments (mostly to private households) and of expenditure on health and education programs. The impact of the latter, however, is only one-third of the first. The relatively high marginal influence of transfer payments is not implausible as it may be argued that they strongly reduce uncertainty concerning future direct benefits and thus may be valued highly by voters. Even if not all voters/taxpayers end up being public transfer beneficiaries, the great majority of them may still expect to. Comparing the influences of both sides of the government budget, the total marginal influence of the revenue items is almost five times as large (in absolute terms) as that of the spending items. Again, a similar result shoving a much stronger influence on voters' support for the government for the revenue side as compared to the spending side has also been found recently for the United States (Niskanen 1979).

In order to do an additional check of these empirical results, an ex ante forecast for the period 1977.10 through 1978.12 was done based on GLS estimates for 1971.1 through 1977.9. The prediction is clearly superior to naive forecasts (Theil's inequality coefficient is 0.05; the average root mean squared error, 0.62%; and the average mean error of deviation 0.63%), indicating that the most important variables determining evaluation of the government's overall performance have been taken into account.

To summarize: The empirical findings show that voters/taxpayers seem to be unaware of the full cost of government activities when the revenue system is fragmented and especially when those types of taxes and other means of revenue extraction are used which are harder to detect than others. There is also evidence that some spending items, such as additional transfer payments, seem to be more popular and more strongly felt than increases in other expenditures. These effects on both sides of the budget and the fact that the voters highly discount past government activities provides opportunities for the government to use its fiscal instruments in a systematic way to retain the popularity it needs to allow it to secure reelection.


Let us assume the existence of a government that wants to maximize its utility, here understood mainly as the pursuit of its ideological goals. Even if it is assumed that the government cannot be voted out of office in the middle of a legislative period, it is still subject to various constraints in trying to achieve its ideological goals, of which the most important is the reelection constraint. Thus the government faces a dynamic maximization problem of when to undertake what fiscal policy action in order to maximize its utility.

It is assumed that the government is not able to resolve this problem, and that it will behave in a satisfying manner. It takes the results of popularity surveys as the best current available indicator of its reelection chances. If the current popularity level of the government is high, and/or if there is plenty of time left until the next election, the government will use its various fiscal instruments to pursue its ideological goals. In comparison with a right-wing government, a left-wing government will generally increase public sector activity including new and/or expanded spending programs. A look at the Australian governments in the 1970s shows that Labor (left-wing) governments explicitly stated preferences for more spending on education and improvements in the welfare and health care systems, and decreased outlays for national defense. The Country-Liberal (right-wing) governments stated preferences for a much smaller level of growth in current and future government activities. There were also major differences in preferences as to how to finance public expenditures. Whereas a Country-Liberal government tends to favor tax financing, a Labor government relies more on incurring additional debt.

If the reelection chances are indicated to be poor, the government will concentrate on securing reelection rather than on pursuing its ideological goals. For this purpose, the government - regardless of who is in power will try to create favorable fiscal illusions on the part of the voters by means of a systematic revenue and spending policy, counting in the voters' short memory to aid it in this. Before an election it will decrease the more direct and strongly felt taxes in particular (such as personal income tax), and increase harder to detect revenues (especially public debt) in order to finance additional popular spending or to cover the deficit caused by the lowering of strongly felt revenues.

When formulating the use of fiscal instruments, it is assumed that in a state of low popularity the government will react all the more strongly the greater is the ratio POP*/ POPt (the critical level of current popularity POP* having the value of 51%). The second important factor is reflected in the discretionary variable "Time since last election" (TSLE), the obverse of time left until the next election, which takes the values 1,2,...,30,1,1, starting with the beginning of each legislative period. [This formulation implies that the government's reaction six months before an election will be stronger then when the election is still far off.] The government has also to take legal obligations and the behavior of the public administration into consideration in using its revenue and spending instruments. It seems realistic to assume that the public administration does not simply follow the wishes of the government but rather tries to maximize its own utility. It tends to resist major changes because this may threaten its own position and prefers to make only small and incremental changes. It would therefore seen to be useful to take the past spending and revenue structures into consideration as these are the starting points for changes therein. In addition to the legal and administrative constraints, the government has to take changes in the balance of payments into account, as well as the size of the budget deficit (a positive sign in the case of deficit, and a negative sign for budget surplus), whose maximum equals the maximum incurrable additional debt, which is set through the legal framework.

It is now possible to formulate the following equation for the i'th fiscal instrument (INSTit):


for i = 1, 2, ...., 26.  [The method used here to measure ideological preferences regarding the fiscal instruments is rather simple: The intercept is broken up into two dummy variables in order to capture the different use of instruments by Country-Liberal and Labor governments with a shift parameter. It is assumed that the ceteris peribus conditions are fulfilled so that ideological preferences can by measured in this way.]

The 26 instruments refer to 10 federal revenue and 16 spending items, and are calculated as shares of total revenue and total expenditure respectively in order to capture the changes in the structure and eliminate the typical time trend effects. The explanatory variables are lagged by six months as it is assumed that the Australian government needs at least half a year to react to political and economic changes (when a longer or more complex lag structure is used, the estimation and the forecast results do not improve significantly).

The theoretically expected signs of the administrative and economic constraints are a1 > 0 (for both the revenue and spending sides), a2 < 0 and a3 > 0 for revenue items, and a2 > 0 and a3 < 0 for expenditure items. In the case of a4 and a5, a decrease is expected in the share of those revenue items which are strongly felt by the voters, and an increase in popular spending items. For the two ideological parameters a6 and a7 the signs already discussed are expected, for example that a Labor government will have a greater preference for increases in education than Country-Liberal governments (a6education < a7education) The results of the simultaneously estimated revenue instruments are given in Table 1.

Table 1 shows that the two types of government have significant differences in their ideological preferences regarding which instruments are to be used. When there is leeway for action, as for example after an election, the right-wing government (Country-Liberal) reduces in particular the assessed income tax, the esteem tax on dividends and interest, and, to the greatest extent, government debt. A left-wing (Labor) government acts in the opposite manner. It increases indirect revenue item (including debt), and, among the direct taxes, especially those which will be reduced by a right-wing government. It is interesting to observe that both governments deviate from their ideological preferences regarding the revenue instruments when trying to secure reelection. Goaded by a low popularity level and/ or an upcoming election, both governments significantly reduce direct taxes, the earned income tax the most. They also reduce excise taxes, which are shown separately on purchase receipts. The harder to detect revenues are deliberately not reduced; in some cases, government debt for example, they are even increased ceteris paribus.

Another interesting finding is that compared to a left-wing government, right-wing governments show a greater preference for a simple and comprehensive revenue structure that relies on relatively few and quantitatively important revenues items. However, when trying to secure reelection, both types of government follow a policy of fragmentation of the total revenue, thus creating a more favorable perception of the government on the voters' part.

An ex ante forecast was also done for 1977.10 through 1978. 12 to provide an additional check of the empirical results. With one exception (government debt) the predictions for the 10 revenue items are far superior to those achieved with naive forecasting methods. The beat predictions are those for direct taxes (Theil's inequality measures being ca. 0.20; the average root mean squared errors ca. 1.75%; and the average mean errors of deviation ca. 1.20%), i.e. chose instruments which are discretionarily changed before an election (and there was a general election in March 1978).



For reasons of space the empirical results for the spending side are not presented here but only briefly discussed. Looking at the ideological differences between both types of government, Country-Liberal governments favor additional expenditure for capital formation, defense, and law and order. A Labor government prefers to decrease the last two and favors instead additional transfer programs (including foreign aid), education, and health care. Again, if a government is afraid that it will not be reelected, all transfer payments to individuals and households (but not foreign aid) are significantly further increased in addition to the most favorably perceived expenditure on education and health. This strategy of increasing especially transfer payments before an election seems to be quite common, and can be observed to operate similarly for instance for the United States (Tufts 1978, chapter 2). The ex ante forecasts which have been made to test the model's predictive ability in an election year lead to superior results for 11 of 16 spending items (evaluated by Theil's inequality coefficient which is smaller than 1). The instruments for securing reelection again provide the best forecast results (with an average mean error of deviation of less than 1.5%).

If these results are compared to the findings for the revenue side, it can be seen that the use of the spending instruments is of less importance in securing reelection. This is not implausible and may simply reflect the often stated greater rigidity of the expenditure side as compared to the revenue side. [Additional evidence for this argument is gained when a regression is carried out with the lagged endogenous variables of the revenue and spending items (in billions of Australian dollars), thus capturing only the legal and administrative influences. The mean of the explained variance is then 82.5% for the revenue items, but 93.1% for the expenditure by categories.]

To summarize: The general hypotheses in this paper are that (1) voters will be rationally underinformed and thus subject to systematic misperceptions of fiscal variables, and that (2) the government in a representative democracy will try to exploit such fiscal illusions in order to strengthen its position, especially when it is trying to secure reelection. The general framework of the analysis is that of a monopoly, though one its need for periodic popular approval. Empirical evidence for systematic misperception of the cost of government is offered by the results for the government's popularity function: The findings show that voters/taxpayers seem to be unaware of the full cost of government activities when a complex revenue system and indirect and more hidden revenue items are used. There is also some evidence that a few spending items are favorably perceived by the majority of voters, though it remains open how much of this is due to preferences and how much due to misperception. This plus the fact that voters discount past government activities all helps to provide opportunities for the government to act as if it had a monopoly position. As the empirical results show, significant differences between the two parties' use of fiscal instruments for ideological purposes appear after an election was taken. However, when trying to secure reelection each government deviates from the pursuit of its ideological goals if necessary and uses the fiscal instruments in a predictable fashion in order to exploit fiscal illusions held by the voters.


Several conclusions and suggestions arise from this analysis, all relating to the provision and diffusion of more and better information on the government sector to politicians and consumers. However, it seems useful to distinguish between information dealing with government sector output, i.e. its product and cost, and with the institutions and processes of real world public decision-making.

To start with the first: Enough has been said in the above to conclude that the true cost of government to the consumer/taxpayer has a large grey area and may be systematically misperceived in many circumstances. Under the perspective of consumer sovereignty, which stresses the right to information leading to meaningful purchasing decisions, better knowledge of the true and of the perceived cost of government to the individual would seem desirable. Some work has begun in this direction, for example with the development of a consumer tax index for Canada (Walker 1976). Moreover, it can be expected that especially the analysis of the perceived tax burden and its determinants may profit from applied perceptual and learning psychology, and from consumer research and its techniques of direct consumer inquiry. The right to information also requires knowledge of the true and the perceived nature of government services, their quality, how long they are provided, and, for example, their spatial distribution. It would seem that this kind of information could be strongly improved using techniques of consumer and marketing research, e.g. multiattribute product analysis and hedonic price approaches (Palda 1980).

However, even if this kind of information is greatly bettered, the problem remains that there may be still only be weak incentives to exploit it. Non-neglectable costs arise to the individual voter in making collective decisions, and the additional cost of doing as on a well-informed basis may far outweigh the gains. As there are usually only a limited number of alternatives to be chosen among, a system of mandatory payments prevails, and the number of voters is large, it may even be irrational for the individual to become sufficiently informed to make an intelligent decision. Under this perspective it would be desirable to improve in particular information on the functioning of the government's institutions and on how the political process really works. This information relates to such questions as to how strong (or weak) the competition between parties in a representative democracy really is, and thus how small (or large) the leeway is that an elected government has to deviate from the voters' wishes. Moreover, it would also provide the basis for research on how the real-life political process and the functioning of government institutions can be influenced so that the wishes of the voters are - quasi automatically -better taken into account by the government.


Breton, Albert (1974), The Economic Theory of Representative Government, Chicago: Aldine.

Darby, Michael R. and Karni, Edi (1973), "Free Competition and the Optimal Amount of Fraud," Journal of Law and Economics, 16, 67-88.

Diamond, Harold S. (1970), "The Budget and Consumer Sovereignty," American Journal of Economics and Sociology, 29, 163o177.

Eismeier, Theodore J. (1979), "Budgets and Ballots: The Political Consequences of Fiscal Choice," in: Douglas W. Rae and Theodore J. Eismeier (eds.), Public Policy and Public Choice, Beverly Hills and London: Sage.

Fray, Bruno S. and Lau, Larry J. (1968), "Towards a Mathematical Model of Government Behavior," Zeitschrift fur National├Ěkonomie, 28, 355-380.

McKean, Roland N. (1973), "Government and the Consumer," Southern Economic Journal, 39, 281-489.

Niskanen, William I. (1979), 'Economic and Fiscal Effects on the Popular Voter for the President," in: Douglas W. Rae and Theodore J. Eismeier (eds.), Public Policy and Public Choice, Beverly Hills and London: Sage.

Palda, Kristian S. (1980), "The Public Sector and Marketing: Should the Twain Meet?" Forthcoming in Proceedings of the Second Annual Special AMA Conference on Marketing Theory.

Pommerehne, Werner W. and Schneider, Friedrich (1978), "Fiscal Illusion, Political Institutions, and Local Public Spending," Kyklos, 31, 381-408.

Pommerehne, Werner, Schneider, Friedrich, and Lafay, Jean-Dominique (1980), "Les interactions entre Tconomie et politique: synthFse des analyses thToriques et empiriques," Forthcoming in Revue Economique.

Schneider, Friedrich and Pommerehne, Werner (1980a), "Politico-Economic Interactions in Australia: Some Empirical Evidence," Economic Record, 56, 113-131.

Schneider, Friedrich and Pommerehne, Werner (1980b), "Illusions in Fiscal Policy: A Case Study," Forthcoming in Statsvetenskpalig Tidskrift.

Tufts, Edward R. (1978), Political Control of the Economy, Princeton: Princeton University Press.

Wagner, Richard A. (1978), "Advertising and the Public Sector Economy: Some Preliminary Ruminations," in:

David G. Tuerck (ed.), The Political Economy of Advertising, Washington, D.C.: American Enterprise Institute.

Walker, Michael ed. (1976), How Much Tax Do You Really Pay? Introducing the Canadian Consumer Tax Index, Vancouver: Fraser Institute.