3 Compensation
It is impossible to address the concept of a taking without being drawn into debates (with deep constitutional implications) over when it is desirable (efficient) to pay compensation, at what level it should be set and who should be liable to pay it. This section discusses these questions, through a review of literature, and without distinguishing between physical and regulatory takings. The nature and implications of such a distinction are considered in the next section.
3.1 Why compensate?
Compensation for takings is to a large degree about incentives. A prime element of this is the problem of “fiscal illusion,”[14] where the true cost of the proposal is hidden. The Italian scholar Amilcare Puviani in 1903 in The Theory of Fiscal Illusion (Buchanan, 1987) addressed fiscal illusion through the question “If the ruling group desires to minimize taxpayer resistance for any given level of revenues collected, how will it set out to organize the fiscal system?”[15] Takings, which carry no explicit fiscal cost or a reduced cost, can become the preferred means of achieving a policy objective, through reducing the apparent cost, even if the total social cost actually exceeds the total social benefit.
A second argument relates to incentives for private parties to encourage use of the takings mechanism rather than engaging in a voluntary transaction at their own cost. Issues around the advantages of making beneficiaries pay compensation and of requiring proportional sharing of the benefits of a taking are discussed in later sections. Finally, there is the longer-term incentives issue, dealing with implications of takings over time for social acceptance of the legitimacy of state action, and for the value of privately held property rights.
There has been an extensive debate in economics literature about the implications for efficiency and national welfare of paying or not paying compensation for takings. The following discussion and Table 1 attempt to summarise the key points of that debate.
Compensation has historically been argued to be due when “a given person has been required to give up property rights beyond his just share of the cost of government” (Stoebuck, 1972). This principle is not usually applied to taxation, possibly because the burden is widely spread and general tax decisions made by the legislature have greater social acceptance than bureaucratic or executive action to take specific property rights from specific individuals. The latter may also offer more scope for abuse or victimisation.
It has been argued that compensation for the use of “eminent domain” and “no taxation without consent” are fundamental rules limiting the power of government and crucial to the consent of the governed (Wilkinson, 2001). This is a constitutional debate.[16] More prosaically, compensation mimics the consideration paid in a voluntary exchange and willingness to pay it may show that the public benefit exceeds the cost. Payment of compensation can also separate the social benefits of the taking from the income distribution effects, reducing any unfair redistributive effects (Blume and Rubinfeld, 1984).
Arguments in specific cases for not compensating include (Wilkinson, 2001) where the costs of identifying and assessing rights to compensation exceeds the benefits of paying it, the benefits affect the same people as the costs, and the taking is restricting the use of market power to the risk adjusted return on capital. It has also been argued, from an economic rather than constitutional perspective, that full compensation creates moral hazard. Investors seek to raise the cost of compensation to discourage a project, knowing that at worst they will be reimbursed, while landowners ignore the social costs created by their inefficient investment if a taking occurs.
One response to this is that under-compensating creates an incentive for takings rather than voluntary exchange (due to fiscal illusion). Another is that not all the cost of an investment (Wilkinson, 2001) or all the values owners attach to their land (Miceli and Segerson, 1998) will necessarily be reflected in market value. The latter problem is not easy to address in practical terms as owners have an incentive to over-state non-market values.[17] It can also be argued that a lack of compensation encourages over-investment if it is perceived that the government is less likely to take valuable land. There is evidence that governments over-compensate for or avoid taking high-value land, and under-compensate for and therefore accumulate excessively, low-value land (Munch, 1976).
Compensation can also be seen as a form of public insurance against a taking (Miceli and Segerson, 1998) particularly if the level of compensation is uncertain. This insurance can be argued to be offset by any tendency of risk averse owners to accept lower offers to avoid litigation costs and by fiscal illusion by governments. Limited counters to these problems include fiscal constraints by government and a possible rule, which would impose all legal costs on the government if it lost (Eposto, 1996).
It has been argued that compensation is not required where the benefits affect the same people as the costs (Wilkinson, 2001). This was covered by Justice Holmes in the concept of “average reciprocity of advantage” where “the share of each party in the benefit …. Is sufficient compensation for the correlative burden (Brauneis, 1996). Regulation that singles out specific property owners without compensation would not meet this equal protection test – but that would not necessarily mean it was inefficient.[18]
The above arguments are primarily subjective and open to differences in interpretation. A number of attempts have been made to design explicit rules to avoid or mitigate the incentive problems associated with compensation. Since these rules are variations on a theme of achieving an efficient outcome, but the variations reveal interesting subtleties about the impact of different objectives and of governmental behaviour, they are summarised together in Table 1 rather than discussed at length in the text.
In terms of the central question underlying this paper “what protections against takings are appropriate, and specifically what is the role of compensation”, the rules in Table 1 focus sharply on the efficiency of compensation in terms of maximising national welfare.
Considering efficiency requires determining what efficiency standard to apply. The literature has argued (Fischel and Shapiro, 1988) for a variety of standards. Pareto-efficiency would require that no one is left worse off and society is better off. Michelman’s standard is for society to be better off but compensation is paid only if it is less expensive than demoralisation costs. The Kaldor-Hicks criterion requires that projects should proceed if society is better off but compensation does not then have to take place
Whichever efficiency standard is chosen, the application of most of the rules cited above requires assessing the social costs and benefits of the taking, and of the effects of not compensating. These assessments carry significant information burdens, while the latter also involves judgements regarding social perceptions of what constitutes a property right and expectations of how rights will change over time.
| Rule | Rule Summary |
|---|---|
| Utilitarian approach – pay compensation if cheaper than the effects of not compensating | Test is fairness. Linked to Hume’s theory of property, with private property emerging from rules needed to allow association of individuals. Arbitrary encroachment on private property has serious disvalue. Bentham’s theory of social utility suggests protection of extant rules on property from capricious redistribution is crucial to gains in productivity. Reject taking unless efficiency gains (E) outweigh either settlement (S) or demoralization (D) costs:[19] E > min (D, S). This does not require Pareto-efficiency. Undertaking inefficient projects, however, if E > min (D, S), would raise demoralization costs and compel compensation. (Michelman, 1967). |
| Compensate when government acting as an enterprise | Compensation when government is acting in an enterprise capacity, but not when it is acting in an arbitral capacity (Sax, 1964). |
| No compensation for insurable risks | Compensation as insurance against adverse effects of government regulation. Private insurance limited[20] as is government insurance.[21] Compensation after the fact avoids these problems, but should only be paid when benefits of doing so exceed the costs. Focus on risk aversion (inverse to wealth?) and where loss is substantial diminution of value. Compensate for large losses to those unable to insure against those losses (Blume and Rubinfeld, 1984). |
| Lump-sum compensation | With lump sum compensation, if projects are appropriately evaluated, investors are risk-neutral and compensation is independent of current land-use, then private investment decisions are socially efficient. Zero compensation is efficient as a special case. With full compensation, investment will exceed socially optimal level, particularly if over investment discourages takings. If fiscal illusion is present, some compensation may be necessary to avoid excessive takings (Blume, Rubinfeld and Shapiro, 1984). |
| Compensation depends on applicable constitutional model | Private insurance is a limited option and does not address demoralisation costs. Pay compensation or not based on expected government behaviour. Zero compensation if no reasonable compensation rule deters takings (compensation will then create moral hazard for investors) or if the government is Pigouvian, acting solely on benefit-cost calculus (so no constraint is necessary). Under majoritarian government, optimal compensation will be more than zero, to discourage excessive takings, but less than full, to minimise moral hazard (Fischel and Shapiro 1988, 1989). |
| Efficiency | For singling out, ex ante efficiency, so landowners are compensated for acting efficiently. For general regulation, ex post efficiency, so compensate if the regulation is inefficient (Miceli and Segerson, 1994). Full compensation at efficient level of investment as, if landowners consider the value of the land is likely to affect takings decisions, under-compensation will encourage over investment and over-compensation will encourage under investment (Miceli and Segerson, 1998). |
| Society’s gains | If the social benefit is known, pay the owner the social value of the taking for the property, or charge the owner that amount. If social benefit not known, apply “two-part tariff”. The owner announces private benefit and a monetary transfer occurs regardless of whether taking proceeds. The transfer is the social gain, if the taking occurs at a price equal to the private benefit, and may be negative. The private benefit is paid if the taking occurs. (Hermalin, 1995). If social benefit is unlikely to be known, this defaults to two-part tariff which would penalise the owner if a taking occurred with a social value less than the private benefit, as can occur when fiscal illusion applies. |
The rules in Table 1 therefore generally appear to assume a level of information that in practice is unlikely to exist, making the feasibility of applying this economic approach in real life questionable. In the absence of accurate assessments of social benefit and cost, and of government decisions based solely on such assessments, the balance of judgement from these rules (as distinct from a constitutional context around constraining government) is unavoidably subjective. This applies to both whether a taking should proceed, and whether and how much compensation should be paid. Economics does not give us a hard and fast operational framework for takings compensation.
Notes
- [14]Similar approaches can of course be taken in the design of regulation, and regulation itself can serve to hide costs by transferring them to third parties and rendering them implicit rather than explicit. It is also interesting to compare the hypothesised political goal of minimising taxpayer resistance with the economic goal of minimising the deadweight costs of taxation, or regulation. Where the two goals are aligned, so perceived costs equal real costs, presumably the outcome will be the socially optimal one, while otherwise it will be suboptimal.
- [15]Mechanisms suggested included distorting the link between the total cost of a programme and the individual share, charging taxes at favourable times such as when an inheritance has just been received, charging fees for nominal services such as licences associated with positive events such as marriage, linking taxes to public opinion such as surcharges on business profits, using scare tactics, or introducing taxes where the ultimate incidence is unclear (Buchanan, 1987).
- [16]Highlighted in the New Zealand context, by references to vested rights and compensation in the Guidelines (Legislation Advisory Committee, 2001) cited in the Cabinet Manual (Cabinet Office).
- [17]Michelman notes that courts usually award either nil or just compensation so that legislated settlement rules may be appropriate, and cites the example of relocation payments for households displaced by federally financed redevelopment activities (Michelman, 1967). The New Zealand parallel is the solatium payment, separate from the calculation of market value, under the Public Works Act where residential property is taken. The New Zealand legislation is otherwise based on market value, an administratively simple approach, and on making those affected no better or worse off than under a voluntary exchange.
- [18]In Michelman’s terms the risks of abandoning efficient projects due to stringent compensation, or unduly concentrating losses due to les stringent compensation are minimised by insisting on compensation “when settlement costs are low, when efficiency gains are dubious and when the harm concentrated on one individual is unusually great”, and relaxing that insistence “when there are visible reciprocities of burden and benefit, or when burdens similar to that for which compensation is denied are concomitantly imposed on many other people” (Michelman, 1967).
- [19]Demoralisation costs are losses from realising there will be no compensation, and future lost production caused by fear of similar treatment. These costs are distinct from random risks such as natural disasters, and arise from cases where low settlement costs would have made compensation easy, burdens are seen as disproportionate, inefficient projects are seen as vehicles for redistribution, there are no reciprocal benefits, or there is little confidence of benefit from future projects (Michelman, 1967). A Pigouvian government acting on benefit-cost calculus would generate more random risks and smaller demoralisation costs (Fischel and Shapiro,1988).
- [20]The main problems are measurement difficulties, the high risk of moral hazard and adverse selection (landowners may be able to affect the taking decision and those at high risk of a taking are more likely to insure), inability to secure insurance company assets against taking, and the fact that demoralisation costs occur when the effect of the taking is realised by land owners, not at the time of the legal taking itself (Fischel and Shapiro, 1989).
- [21]Government would have incentives to take uninsured land, distorting decisions on whether to insure, and setting premiums for low-probability events is difficult (Blume and Rubinfeld, 1984).
