CHAPTER 4
4.1 INTRODUCTION
The previous chapters have indicated that government actions can affect the size of the rent from an asset. Scant attention has been paid to the benefits and costs associated with different types of government intervention. This chapter will examine the various methods used by governments to create and distribute economic rents. For each method the welfare implications will be examined in detail. Where possible, the methods are explained by means of simple numerical or graphical examples.
4.2 A SIMPLE FRAMEWORK
Suppose the government has a fixed supply of some resource. In terms of Figure 4.1, the supply curve is denoted by S and the quantity of the resource is OA. The resource is fixed in the sense that irrespective of the price offered for it, no additional units of the resource are forthcoming in this period. For example, OA could conceivably be the number of first year places offered in a university. Alternatively, it could represent the number of abalone licences offered by some State fishing authority.
Consider now the demand side. Suppose, for the sake of argument, that the demand curve for the resource is D1 in Figure 4.1. This demand curve represents the market's willingness to pay for successive units of the resource. Let us explain the market here in a bit more detail in terms of the tools developed in Chapter 2 for the case of the egg market. For example, the willingness to pay for the Bth unit of the resource is equal to BC. The willingness to pay declines as more units are purchased. Specifically, the willingness to pay for the Dth unit is DE. So far, we have discussed the willingness to pay for additional units. The total willingness to pay for OD units is the sum of all the vertical distances corresponding to area ODEF. Suppose that the price per unit of the resource is, for whatever reason, set at P1. At that point, agents would be willing to purchase up to OD units of the resource. Their outlay would be equal to area ODEP1. The outlay represents a cost to consumers since they have to forgo expenditure on some other item as a result. The total willingness to pay for OD units is ODEF. The purchasers are giving up less than what they are gaining. The difference between the total willingness to pay and the outlay is represented in Figure 4.1 by area P1EF. This area is often referred to by economists as consumer surplus.
If the demand for the resource is D1 and the supply is S then the market price for it will be zero. The abundant supply at every particular price means that the price of the asset will be bid down to zero. At this price, OG units of the resource are demanded and GA units remain unused. In this situation, the resource yields no rents and there is no point for the government to price this resource. In fact, if the government were to decree that the price of the resource is P1, then the quantity demanded would fall to OD, the consumer surplus would be P1FE and the revenue to the government would be P1EDO. Hence, the net loss to the collectivity would be DEG. What this example shows is that in the case of abundant resources there is no need for the government to set a price. It should simply distribute them at zero price.
Consider now the case of a scarce resource. This case is similar to that described in the previous figure save for the fact that the supply curve has been shifted to the left. In such a case, demand outstrips supply at a zero price. This case is illustrated in Figure 4.2. Here, the government cannot simply distribute the resource free of charge and satisfy all those interested in acquiring it. If the government were to do that, then demand would be OB, supply would be OA and AB units would constitute the unfulfilled demand. As the diagram indicates, the market solution would be for the price to adjust until there is equality between demand and supply. This occurs at P1. If the resource were traded freely on the market, then individuals would bid with each other and purchase a total of OA of the resource. In accordance with the discussion in Chapter 2, the rent accruing to the owner of this resource would be OP1CA and the consumer surplus is P1EC. The total benefit to society is equal to the rent plus the consumer surplus and therefore is OECA. It is worth pointing out that the market clearing price maximises the total benefit to the collectivity; any other price would lead to a lower total return from the resource. For example, if the price were P2 then rents would be equal to P2FGO, consumer surplus would be P1FE and total benefit to society would be lower than at P1 by the area FGAC. The welfare cost of setting the price too high is therefore equal to FGAC.
The analysis here indicates what is feasible under an efficient market process. The government as owner of the resource would sell the asset at a price of P1 per unit and the total wealth of the economy would be maximised. The question becomes whether the government should in fact mimic the market. Are there any other choices? Should it distribute the available supply of the resource by giving it away? And if this is the case, should it be done randomly or on a first-come-first-served basis? The government must choose some method of distribution. Some of the factors on which that decision should turn are explored in subsequent sections of this chapter.
4.3 AUCTIONS
In the economics literature, a widely discussed form of distributing the resource is through auctions. This method can be illustrated with the aid of Figure 4.2. Suppose the government is willing to sell off a parcel of land to a developer who makes the highest bid. In general, developers would be willing to pay up to the expected gain from owning and later selling the land. Suppose that the market for land is competitive, so that this parcel of land is not the only one available to prospective home owners.
Suppose the developer expects to be able to sell the land at P1 per lot. In that case, developers would be only willing to pay up to P1CAO, which is the area of the rent. The government would be better off by the amount of the revenue collected from the auction (P1CAO), home owners would be better off by the amount of the consumer surplus (P1CE) and the developer's return would have been transferred to the government in the form of his bid. In this particular case, the total benefit from the land would have been maximised; the auction has mimicked the competitive market outcome.
In the scenario sketched here, the developer is making no return on his efforts at all. This may affront some readers. The operator will certainly have costs associated with his enterprise. They would consist, for example, of the costs of drawing up the proposed development, and time spent in negotiating and formulating his bid, as well as the normal rate of return on his capital. In terms of Figure 4.2, these costs can be depicted as OAHI. The developer's net gain from obtaining the land is now equal to P1CHI, and this would represent his maximum bid. If he obtains the land for exactly P1CHI, then he will be making a normal return on his capital and time. All of this would alter the size of the benefit to the collectivity, however. Since the developer's costs constitute forgone opportunities, they must be deducted from the total benefit arising from the land. Whereas these benefits were equal to ECAO in the absence of costs, they are now equal to ECHI.
It would be wrong to conclude from this that the government ought to sell the land itself to potential home owners rather than to a developer. Just as the developer incurs costs in the development of the land, so would the government face similar sorts of costs. One might infer that there is no way to choose between the two methods of land disposal. This would only be true if the administrative costs of the government were identical to the operating costs of the developer. There are good reasons to believe that the developer will actually have lower costs than the government. This arises from the fact that the self-interested developer has a more clearly identified incentive to reduce his costs than does the government bureaucrat. The private developer receives the fruits of his efforts himself, whereas the bureaucrat is not directly rewarded as a result of cutting costs. This is merely a specific case of the more general insight that "there is no such thing as a free lunch"; both the government and the market are costly. The economists' job is to identify which particular institution can do a given task at lowest cost to society and the presumption here would be in favour of the developer.
In the discussion above, it was implicitly assumed that the developer possessed perfect information about demand conditions. If he is uncertain about the position of the demand curve, then it may very well be the case that he pays too little or too much for the land. In the first case, he makes above normal returns on his efforts, whereas in the second case, he suffers a loss as a result. If he does end up making large profits, then this result is not undesirable on efficiency grounds. He has gained merely at the expense of the government. Instead of the government receiving higher revenues from the land sale, the benefit is simply transferred to the developer. As we pointed out before, economists have little to offer in choosing between alternative distributions of gains. The point of this is, of course, that one cannot look at individuals who are making large profits and automatically infer that these profits were the result of a conspiracy against the public. The profits arise here from luck, or perhaps, better foresight constituting good entrepreneurship.
On the other hand, the developer could have sustained a loss if his expectation of demand conditions were overly optimistic. In such a case, the government gains at the expense of the developer. But again, total benefits to the collectivity are not changed.
The case of uncertainty highlights a crucial aspect of the competitive process. As we pointed out in Chapter 2, entrepreneurs are continually searching out opportunities for gain, and part and parcel of this process is the distinct possibility of mistakes and failure. The entrepreneurs who continually make gross mistakes should fall by the wayside. Their more successful compatriots will "live" to fight another distributional struggle. This continual process of entry and exit, profit and loss, success and failure constitutes the cost of using the decentralised market system.
An important problem can occur if the government decides to interfere with the market process by "bailing out" some or all of the unsuccessful entrepreneurs. Before we can discuss the so-called "moral hazard" problem in the market place, the following example closer to everyday experience may be useful. Consider an individual who has taken out full insurance on his car. In the event of an accident the insurance company promises to pay the replacement value of his car and any damage caused to third parties. Once the individual has paid the premium, however, he has fewer incentives to be a careful driver for it is the insurance company which will have to pay the damages incurred by the individual. Consequently, it may be the case that the probability of having an accident rises as a result of the insurance. Of course, insurance companies are aware of the moral hazard problem and take measures to ensure that they will not go bankrupt themselves. One such measure is to charge a rate above the actuarial rate, with the surcharge covering the increased probability of an accident. An individual who is particularly prone to accidents will find that either his premiums rise or, worse still, no insurance company will be willing to insure his car. Another measure aimed at curtailing the impact of the moral hazard problem is to have a deductible. These measures are designed to encourage the individual to take better care of his assets even with insurance. There is some loss to the individual even with insurance.
If the government provides compensation to individuals who lose in the market place, then it encourages the moral hazard problem to emerge here as well. For example, take the case of deposit insurance provided by the government. (15) Suppose that all deposits below, say, $100,000 are automatically insured by the scheme; irrespective of the bank's decisions the depositor will be covered by the insurance. Since the bank's liabilities are guaranteed by the government, the bank is, in effect, able to take on riskier lending prospects than it would have done in the absence of deposit insurance. Equally clearly, potential depositors will tend to take less care with their choice of banking institution, since all banks are covered by the scheme. As a result of the deposit insurance scheme, banks and depositors are more prone to make poor choices than they would otherwise have been.
When viewed in this light, it is hard to rationalise the failed attempts by the Western Australian government at saving the Rothwell's merchant bank or the Victorian Government's interference with Tricontinental. Merchant banks (16) are mainly aimed at servicing the corporate sector. In such a setting, the arguments in favour of deposit insurance are even less convincing than in the case of a normal savings bank. Depositors of merchant banks are (or should be) fully aware of the fact that the higher return on their funds reflects the greater degree of risk involved. As a result, the failure of one merchant bank is not likely to influence the economy-wide confidence in the banking sector as a whole.
A more appropriate scheme aimed at getting more security for depositors and forcing banks to take more care, is to encourage the banks to self-insure. Banks do this by borrowing and lending on the international financial markets, thus diversifying their risk. Of course, if there is a downturn in the world economy, then even this form of self-insurance will not be adequate to keep banks from failing. From the point of view of banking as an entrepreneurial activity, there is nothing wrong with the occasional bank going into liquidation. Nonetheless, there are certain aspects to banking that make it different from other entrepreneurial activities and provide some justification for government intervention. In particular, an important aspect of the lubricating role of credit in a market economy is that depositors have some security in the banking system as a whole. If individuals had no confidence in the banking system, then it would lead to a very sharp downturn in the level of economic activity.
The point of all this is that there is a role for government intervention in the banking system. Following a global downturn, the government cannot allow the banking system to collapse since that would create chaos. The government could use its monetary powers, say, by acting as a lender of last resort. The commitment by the government is to support the banks against problems which were outside the control of the individual banks. If individual banks run into financial difficulties for other reasons, for example, through excessively risky lending activities, then there is little case for government intervention. To do so would merely send a signal to that individual bank and other banks like it, that they can be largely immune from making mistakes.
Returning to the general topic of auctions as a method of distribution, some additional insight can be gained by reconsidering the case of premature land settlement discussed in Section 3.2.2. There, we discussed the case of a group of Eastern farmers considering occupying the unsettled land in the West. Suppose that the land in the West is owned by the government and that it wishes to distribute the land in the current period t. If the auction method is chosen, then farmers in the East would be willing to pay up to the net present value of the rents from the Western farm. In terms of Figure 3.2, they would be willing to pay the present value of the area PWCPE. The successful bidder would settle in the West at the optimal time, 10 years from now. Hence, it would seem that the auction system would produce the socially optimal outcome.
Fane and Smith (1986, page 214) argue that there are three potential defects associated with the auction method. First, auctions are not costless; there are administrative costs in setting up and conducting auctions. Fane and Smith argue that these costs are typically a small proportion of the expected revenue from the auction. Whether or not these costs are small relative to the particular resource to be auctioned off is not the only relevant issue here. The important point here is that administration costs must be taken into account when assessing the relative merits of the different distribution methods. On a priori grounds, little can be said about these relative administration costs. For particular cases, however, comparisons ought to be fairly straightforward.
The second possible defect of the auction method is that the participants in the auction may attempt either to collude or to bid strategically in order to keep the price down. Collusion can be overcome to an extent by using sealed bids. In such a case, the partners in the collusion cannot ascertain whether their cohorts have stuck to the deal or not. Strategic bidding is also likely to be mitigated by sealed bids. If you cannot directly observe what other individuals are bidding, then all you can base your strategic bid on is what you expect them to have bid. Because of this extra uncertainty, you must bid more than you would have if bids were not sealed. Of course, it is always possible to obtain inside information about rival bids, say, from the auctioneer or the bureaucrat in charge. We elaborate further on the issue of corruption in Section 5.4 below.
Third, Fane and Smith (1986, page 214) discuss the so-called "sovereign-risk problem". This is a specific example of the so-called "dynamic inconsistency problem". (17) It occurs if an initial announcement by the government about the conditions of ownership is not believed by the auction participants. They, as a result, may bid less than if they had fully believed the government's announcement. The government then faces the choice between a rock and a hard place. It either sticks to its word and is content with receiving less for the resource than it could have, or alternatively, it breaks its word after the auction and attempts to raise additional revenue from the successful bidder by other means. If the government is seen to break its word, then the next time an auction is held its announcements are not likely to be believed either. On the other hand, if the government does not break its word often enough then the sovereign-risk problem should disappear. Indeed, some economists have argued that the existence of so-called reputation effects may force the government to act in a dynamically consistent fashion. (18)
4.4 RANDOM DISTRIBUTION
The layman often finds the market solution, as exemplified by the auction system, to be cold-hearted and markedly unfair; after all, the individuals who end up acquiring the resource are those willing to pay the highest price. This sentiment, or feeling of resentment, is magnified if they identify such individuals as being the richer members of society. As a result, some individuals may argue that the resource should be distributed randomly, so that nobody is favoured over anybody else. Economists' responses to these claims are twofold. First, they would argue that the price of an asset is determined by the total demand for the good. Rich and poor alike bid with each other and in so doing, determine the market price. The rich alone do not determine the going rate and as a result, should not be singled out for blame. Second, although the random system may be considered to be fair, it is easy to show that it is an inefficient method of distribution. The method will not mark a return of the lucky country.
Suppose there is some commodity of which only one unit is desired by each consumer. Some consumers value this unit higher than other individuals. This is reflected by the downward sloping demand curve D in Figure 4.3. Consumers are ranked according to their willingness to pay for the commodity in a descending order. For example, the first individual's willingness to pay is OA whereas the last individual's willingness to pay is zero. The available supply of the commodity is OB, which is in this case by construction equal to half of demand forthcoming at a zero price (OB is half OC). The government owns the units of the commodity and wishes to distribute them randomly to all those individuals interested in acquiring one, regardless of their willingness to pay. Since there are twice as many individuals as units of the commodity, the probability for each individual of acquiring one unit of the commodity is one half. Accordingly, each individual's willingness to pay is halved. For example, the first individual is willing to pay OA if he gets one unit of the commodity for certain. His willingness to pay for the chance of obtaining one unit if the probability of success is 0.5, is half of OA. (19) The same reasoning holds for all individuals. The demand curve is therefore EC, rather than AC.
The probabilistic demand curve EC indicates that the total benefits to the agents in the economy derived from the units distributed in a random fashion are OEC. We can compare this with what would have occurred if the auction system had been used. In such a case, the price would have been set at OE, the point at which demand equals the available supply. Consumer surplus in that case would be AEG and the rent would be OBGE leaving total benefits equal to AGBO. The loss in benefits as a result of the government's random distribution method is equal to the difference between AGBO and OEC. This in turn is equal to AEG. (20) Compared to the auction system, the method of random distribution actually leaves the collectivity worse off as a result. (21)
4.5 DISTRIBUTION BY CHARACTERISTICS
Suppose the government takes heed of the lesson against random distribution and decides instead to allocate the units of the resource on the basis of certain characteristics possessed by the individuals. The system of land distribution on Norfolk Island provides a good example of the method at hand. There, descendants of the mutineers from HMS Bounty have first right to buy any land that has become available. Other examples arise from the Welfare State; mothers receive benefits according to their marital status, Australians receive different land entitlements on the basis of their ethnicity, and students at Australian universities pay differential fees according to their citizenship.
Suppose the government allocates units of the resource on the basis of characteristics that are perfectly correlated with the individuals' willingness to pay. Consider the question of who should receive a certain plot of land. Suppose the plot is on a sacred site of the aboriginal community and that they would be willing to pay more than a developer interested in the same land. In terms of Figure 4.4, suppose that there are only OA (= AD) hectares of the land and two interested parties, the aborigines and the European land developer. Suppose that the aborigines' total willingness to pay is OABC and that of the developer is ADEF. If the government were to assign the available OA units of land on the basis of the ethnic characteristics and makes the assumption that aborigines value the site higher than the developer, then the units of land are allocated to the aborigines, who in this case also happen to be the highest valuing users. The market outcome, in the absence of strategic bidding, would be identical, in that the aborigines could outbid the developer by a maximum of GFBC. In this case, the two methods of distribution both result in an efficient allocation of resources.
Now consider the alternative case where the developer values the land higher than the aborigines. In terms of Figure 4.4, the former has a total willingness to pay of OABC and the latter are willing to pay only ADEF. If the government persists in its method of distribution by characteristics, then the land will go to the aborigines. Two scenarios are possible subsequent to the land allocation. First, if the aborigines are free to sell the land to the land developer, then they could conceivably make a profit equal to GFBC. In this case, the aborigines would be better off by that amount and the developer would be no worse off than before he bought the land, leaving a net improvement to the collectivity. All that has happened is a redistribution of income towards the aborigines.
The second scenario, on the other hand, entails a social loss. If the aborigines are prevented from selling the land by governmental decree, then the land will have been allocated inefficiently. The potential gains from trade with the developer (GFBC) remain unexploited in that case. The analysis echoes the theme developed in Chapter 2. The auction method harnesses the information possessed by all individuals. The method of distribution of characteristics fails to solve the information problem.
Some individuals may react strongly to the suggestions here. They may, for instance, argue that assets with religious bearing should not be assessed on the basis of a market concept such as willingness to pay. Although we defer our discussion of these issues until Chapter 6 below, it must be emphasised that they are important and must be taken into account in any assessment of the appropriate means of distributing resources.
It is useful to discuss the method of distribution by characteristics in a setting involving time. Consider once again the case of settlement of the West by Eastern farmers. We have already shown in Section 4.3 that the socially optimal time to settle the farm will be achieved under an auction system. So far, we have been assuming a unitary form of government. In a Federation, however, matters are not quite as simple as this. Suppose that it is a State government that has control over the sale of the land in the West. Its primary interest will not lie with the present value of all output produced in the two regions together. Rather, its interest may, for example, lie in luring industry to the State or in having farmers settle the land as early as possible. Suppose the ownership is conditional upon the farmer's residing on the land. This is clearly an example of the State government using the characteristic of residency as the grounds on which to distribute the land.
In the setting here, if the State government wishes to encourage settlement in the current period, then it would have to subsidise the settlers. In terms of Figure 3.2, individuals who leave the East in order to settle now give up a return of PE per year in order to earn PW in the West and conditional ownership over the land. The amount of the subsidy required in present value terms would have to be high enough to render the net present value of settling immediately equal to zero. On the basis of the analysis in section 3.2.2 we know that the net present value of Western farming in period t+6 is zero. Hence, areas ABC and PWCPE offset each other in present value terms. In order to compensate the farmers, the subsidy must be at least equal to the present value of the area ABDE.
If this avenue is chosen by the State government, then two things happen. First, the State government loses revenue from the land sales and needs to finance the subsidy somehow. Second, and more importantly, the State's action imposes a welfare cost on the entire Federation. The size of the welfare loss due to the State's actions depends on what method the Federal government would have used if it were in charge of land sales in the country. If it had adopted the auction method, then settlement would have occurred at the end of period t+10. The loss to the economy of the State's "body snatching" policy would in that case be equal to the present value of DCE. This area represents the present value of lost output due to settling in period t rather than at the end of period t+10.
If the State government had not persisted in their requirement that the settlers must physically occupy the land in order to retain ownership, then matters would have been completely different. Farmers attracted from the East would be given the subsidy of the present value of ABDE. But as soon as they arrive in their newly adopted State they would have an incentive to sell the land and receive the present value of the land when settled at the optimal time (area PWCPE). Let us suppose it took the farmer one year to sell his land and return to the East. The welfare cost to the economy as a whole would seem to be the lost output sustained during that period (area EFGD). The portion ABGF of the subsidy merely represents a transfer from the State government to the farmer and as such, imposes no welfare cost on society. (22)
4.6 QUEUING
An alternative method of distributing units of the resource is on the basis of first-come-first-served. In the following, the initial assumptions are that time is the only cost of queuing, consumers are perfectly informed about the equilibrium queuing time and the opportunity cost of time is the same for all individuals. Consider Figure 4.5 which depicts the demand (D) and the supply (S) of the resource in question. As in the earlier case of random distribution, all individuals want one unit of the resource and they are ranked by descending willingness to pay. The willingness to pay is not expressed in money terms, but rather in terms of waiting time. For example, the most keen consumer is willing to spend OA units of time in the queue in order to obtain one unit of the resource. Here, the equilibrium price is P1 and OB units of the resource are distributed. The successful individuals do not pay any money for the resource, but their payment is in kind in the form of time spent in the queue. As a result, the government does not receive any revenue from the resource. This has important welfare implications. In the market case, units of the resource would be sold for money and the government would obtain a revenue equal to the area OBCP1 times the opportunity cost of time.
Since this represents a gain to the government and a loss to the consumers, it is merely a transfer and consequently, has no welfare costs. In the case of the queue, however, the area OBCP1 represents waste. The individuals suffer forgone earnings to the tune of OBCP1 times the opportunity cost of time. These earnings are not transferred to any economic agent, however. They are dissipated in the queue.
In the present case, the waste could have been avoided if the resource had been sold in the market. It would not have been necessary for the individual to wait in a queue in order to acquire a unit of the commodity. In the market case, the total benefit would have been OACB and in the queuing case it is only AP1C. This latter area is simply the traditional consumer surplus expressed in terms of time.
One should not draw the conclusion that wherever there is waiting there is inefficiency and waste. Indeed, it is hard to imagine any act of consumption that does not require time in one way or another. Examples of unavoidable waiting are easy to find: the individual who sits in his car while his petrol tank is being topped up, customers going down aisles in a supermarket in search for items or motorists waiting for the traffic light to turn green at an intersection are just some. It is necessary to devote time to these activities. As economists would put it, it represents an input into the consumption process. As such, expenditure on time in these cases does not represent a welfare cost because there is no other way of achieving the same result at lower cost.
Examples of government services that involve queuing aspects are plentiful and include licence bureaux, airports, public golf courses, public health clinics and municipal swimming pools. In each of these cases the government does have an alternative way of distributing the resource and hence, the waste associated with queuing could be reduced.
It is often claimed that the queuing method is used despite its efficiency cost because it redistributes these resources to the poorer members of society. The poor have a relatively low opportunity cost of their time since their wage rate is lower than that of their wealthier fellow citizens. A poor individual waiting all morning at an outpatients clinic would give up less income in terms of forgone wages than would a barrister. Consequently, the cost of queuing is relatively lower for the poor and more of them will find themselves successful in obtaining a unit of the resource. The rich, on the other hand, "cannot afford" to wait for very long and are likely to miss out and end up buying the unit in the private market at a higher price. (23) This argument was criticised by Barzel (1974) who showed that redistribution will only be towards the poor rather than away from them for certain kinds of goods. In the case of opera tickets, for example, the poor would not benefit greatly from free distribution as they are not likely to want to go to the opera. In the case of hospital care, however, free distribution achieves the distributional objective; the poor will queue in the outpatients section of the public hospital while the rich will attend private clinics.
Up to this point in the analysis, the examination has turned on the issue of how the wealth of the nation can be maximised. Some people would argue that the developer's wealth, or for that matter, the consumer surplus, should not figure much at all in the balance sheet. The government's objective should be to maximise the return or rents associated with the citizenry's property. Rather than concern themselves with the overall wealth of the nation, the mood of these people is captured in the following catch-phrase: "If the developers want to exploit or buy our property, then they ought to be squeezed for every single cent they are worth." In terms of the rubric of economics, the public wants to capture all of the rents accruing from the resource. This issue will be discussed further in the final sections of this chapter.
4.7 EXACTIONS
Exactions are a popular method by which a community can extract the rents from a resource. Exactions take the form of payments in kind to the community. For example, Fischel (1987, page 103n) cites the case of a Los Angeles developer who financed a new museum for the city in exchange for the right to develop a city block into a commercial plaza. In order to examine the welfare effects of exactions, consider a entrepreneur who wishes to create and develop a pulp mill. The discharge from the mill will impose damages on the environment. The higher the discharge from the mill, the greater the cost of the pollution. These costs imposed on the community are in the form of forgone opportunities: it may not be possible to sit in one's backyard because the smell is too bad or toxins discharged into waterways may make swimming and fishing dangerous. Assume that the dollar value of these damages is $140 million and that each of the 1 million citizens sustains a damage of $140. The community would have to be compensated to the tune of $140 million for the citizenry to feel equally well off as they would have without the mill. Suppose that the gain to the entrepreneur if the pulp mill goes ahead is $600 million.
In this case, the mill should be constructed on the basis of economic efficiency: the gains to the collectivity net of damages are $460 million. If the property right to clean water and air is held by the community, then the project will not go ahead unless the community is compensated for the damages. It should be possible to find an institutional arrangement that will exhaust the potential gains from trade. The discussion here parallels the discussion found in Chapter 3 where it was demonstrated that private bargaining can eliminate the externalities associated with an agent's actions. The $140 million worth of compensation can be transferred to the community in several different ways. The money could be paid directly into the government's coffers, a payment could be made to each individual or the payment could be in kind.
Starting with the case of payment in kind, the project proposed by the entrepreneur must be at least equivalent to a cash payment of $140 million, otherwise the community would not be fully compensated. In general, a payment in kind will involve more expenditure than $140 million by the developer. In order to see why this is so, consider the following additions to the example. If the community had received $140 million in cash, then the individuals in the community would have been able to spend the money in their preferred manner. Suppose that, in the aggregate, the community would have spent $10 million on a community centre and $130 million on private goods such as cars, houses and drills. Suppose that the entrepreneur is vaguely aware of the community's desire for a meeting place and proposes to build a community centre of mega-proportions at the cost of $140 million.
One might be inclined to think of these two alternatives as being worth the same to the community, since in both cases the community's "balance sheet" is up by the same amount. Economists have typically argued that this need not be the case. The citizenry's preferred mix is $10 million for a centre and the rest for private goods. A unanimous decision on how the developer could compensate each individual would result in each individual receiving his preferred bundle. It would have been possible for them to choose to spend the entire $140 million on the centre, but they did not choose this option. This means that the privately selected option leads to a higher level of well-being. (24)
In order to attain this higher level of well-being through a community centre alone, the centre must be even bigger and better. Suppose that a centre costing $200 million is equivalent in welfare terms to a cash transfer of $140 million. The project would go ahead, the pulp mill would be constructed, but the community would have wasted $60 million. The community could have been compensated for the damages through a $140 million cash payment, leaving $460 million to the developer. Upon receiving a gift in kind, costing $200 million, the community is just as well off as under the cash payment, but the developer now only clears $400 million and is therefore worse off by $60 million. This would seem to suggest that gifts in kind in general carry heavy welfare costs, and governments should be discouraged from accepting them. But once this is accepted there is still a need to explain the puzzle of why governments do in fact accept payments in kind. This puzzle is explored briefly in Section 5.3 below.
4.8 RESOURCE RENT TAXATION
A widely discussed method of extracting the rents from a resource is through taxation. In order to see why this form of taxation has some desirable features, consider the example, from Fane and Smith (1986, page 213), of a government which auctions off the right to explore and, if so desired, develop an offshore oil project. Suppose also that future rents are taxed at a rate ά, and that rights are deemed to be permanent once granted. In the absence of uncertainty about future oil prices and costs, the stream of rents from the oil field are known. The present value of the stream of before-tax rents is equal to, say, V0. Since all rents are to be taxed at a rate of ά, the present value of after tax rents is equal to (1-ά)V0. Provided the bidding at the auction is competitive, the successful bidder will pay exactly (1-ά)V0 to the government. Over time the successful bidder will pay taxes on the rents derived from the oil field, the present value of which is άV0. The receipts to the government in present value terms are equal to the auction receipts [(1-ά)V0] plus the present value of tax receipts (άV0), that is, V0. In this case, the entire net present value of the rents is captured by the government. All that the rent tax has done is to alter the timing of the government's receipts. The scheme is also neutral, in that individuals are not encouraged by the presence of the tax to undertake lower-valued activities. The exploration takes place, the development of the field is not affected and the government captures the rents from the resource. This is all predicated on the use of an auction to sell off the rights. As we found above this method appears to be less prone to economic waste than any of the other methods available.
Implicit in our discussion of the rent taxation is that the government receives a fraction ά of rents if these rents are positive, but must compensate that same fraction of any negative rents that may occur through time. At times, the government would have to subsidise, rather than tax, the developer with cash payments. Before the reader closes the book, indignant at this suggestion, consider the problems which would emerge if the government attempts to compensate the firm through tax credits. The neutrality of the tax may be destroyed. To see why this is so, consider the case of a resource rent tax (RRT) as analysed by Fane and Smith (1986, pages 215-219). Under an RRT, tax credits due to negative rents can be carried forward at a rate of interest to be used to offset future tax liabilities. If the rate of interest at which they can carry forward these tax credits (the so-called "threshold rate") coincides with that available to agents in the private sector, then the RRT is neutral. This can be shown with the aid of a simple example of a resource which is only expected to last for three periods and has a salvage value of zero.
The second column indicates the rents the company will make over the life of the asset. If the rate of tax on rents is 40 per cent, then the tax payments under the pure rent tax are as given in column three. The after-tax rents under this system are given in the fourth column. Notice that in the first period the government is compensating the company for 40 per cent of its loss. Using a discount rate of 10 per cent, the net present value of the after-tax rents is equal to $4.03 (= -3/1.1 + 9/1.13).
Since the company makes no positive rents in the first two periods, it pays no tax in these periods under the RRT. Rather, it receives a tax credit in the first period equal to its loss times the rate of tax, that is, $2. This tax credit is carried forward at the interest rate used for discounting (10 per cent), so that in period 2 it has grown to $2.20. But in period 2 rents are exactly zero, so that no tax is levied nor are any new credits accumulated. The tax credit from the first period is carried forward to period 3, at the same rate of interest as before. In period 3, the credit obtained in period 1 is worth $2.42. Column seven gives the rent figures net of the RRT. In period 3, rent net of tax is $15, the tax liability is $6 and the accumulated tax credit is $2.42, leaving an after-RRT rent of $11.42.
The present value of the after-RRT rents, calculated using the discount rate of 10 per cent, is equal to $4.03 (= -5/1.1 + 11.42/1.13), which is identical to that under the pure rent tax. This shows that the pure rent tax is identical to the RRT with full loss offsets in the form of tax credits. It is also possible to show that these taxes are both neutral, in that they do not affect the company's decision whether or not to explore and that the government captures all rents from the resource. Under both the pure rent tax and the RRT, present value of tax receipts is equal to $2.69 (= -2/1.1 + 6/1.13 = (6 - 2.42)/1.13). The company would have been willing to pay for the right to explore an amount up to $4.03. The present value of before-tax rents given in column two is equal to $6.72. This illustrates that the entire rent is captured by the government in the form of auction receipts ($4.03) and tax payments ($2.69).
The equivalence between the pure rent tax and the RRT is lost if less than full loss offsets are allowed as tax credits or if the threshold rate differs from the rate available in the private capital market. The following example illustrates this for the case of less than full loss offsets. In Table 4.2, the same project as in the previous table is analysed. The pure rent tax and the RRT are both based on a 40 per cent tax rate, but the RRT allows for only 80 per cent loss offsets.
Since only 80 per cent of losses are turned into tax credits, in period 1 only $1.60 is carried forward to period 3. Using the discount rate of 10 per cent, this tax credit is equal to $1.94 in period 3, when it can be used to partially offset the $6 tax liability. Column 7 shows the after-tax rent payments under the RRT with 80 per cent loss offsets. Using the discount rate of 10 per cent, the present value of this income stream is $3.67. The present value of the after-tax rents under the neutral pure rent tax is equal to $4.03, demonstrating the fact that the two taxes are now no longer equivalent.
It is also possible to show that the RRT with less than full loss offsets is no longer neutral either. Consider Table 4.3 where the rent streams of two projects are given. Column 2 replicates the project used above in Table 4.2. In column 6, rents from another project are given. In this second project, no losses are incurred and hence, no tax credits are accumulated. As before, the present value of the first project before tax at a discount rate of 10 per cent is equal to $6.72.
The before-tax present value of the second project is $6.61. Hence, on the basis of before-tax returns, the first project is more profitable than the second and ought to be chosen.
Suppose now that a RRT is adopted with 80 per cent loss offsets. As calculated above the after-tax present value of the first project is now $3.67. The after-tax present value of the second project is $3.97. Hence, as a result of the RRT with less than full loss offsets, the project that was inferior before tax becomes the superior project after tax. The tax is non-neutral in the sense that the ranking of projects is altered by it. Of course, if 100 per cent loss offsets were allowed, then the after-tax present value of project 1 would have been $4.03 and the relative ranking of the two projects would have been preserved; the RRT would have been neutral.
In the case where rent streams are uncertain, matters become more complicated. In the case of certainty, the government can acquire the present value of the rents from the project in the form of auction receipts and future tax receipts. If the rent tax is set equal to zero, then all the rents are received in the form of auction revenues. If, on the other hand, the rent tax rate is set higher, then revenues from the auction will be lower and part of the rents will be received over time in the form of rent tax payments.
Under uncertainty and with risk neutral firms, the RRT with less than full loss offsets is distortionary for yet another reason. Fane and Smith (1986, page 218) show that, effectively, the RRT becomes a tax on risk taking. This can be shown with the aid of the following example. Consider a risky project that entails fixed exploration expenditure of $E. If the project is successful, then the before-tax rents are $S. Assume that the probability of success is π so that the probability of failure is (1-π). Before tax the expected net rent from the project is equal to the expected rent minus the fixed exploration costs, that is, V0 = πS-E. Provided V0 is positive, the project is worth undertaking. Under a pure rent tax at a rate of ά, the expected net rent from the project is equal to V1 = (1-ά)V0. This implies that the company receives an amount of $άE from the government and expects to pay in rent taxes $πάS. Provided the project is viable before tax, it will also be viable after tax as long as the tax rate ά is less than 1.
Now consider what would happen under an RRT with no loss offsets. If the project is successful, then net rents are equal to (S-E) of which ά(S-E) is taxed leaving (1-ά)(S-E). This is equal to π(1-ά)(S-E) in expected value terms. If the project is unsuccessful, then the loss is equal to E, or (1-π)E in expected value terms. Combining the two terms yields the expected after-tax rent from the project, V2 = (1-ά)V0 - ά(1-π)E. This expression can be rewritten as V2 = V1 - ά(1-π)E. This shows that the value of the project under the RRT is different from that under the pure rent tax. If the project is viable under the pure rent tax, it may not be so under the RRT because the distortion term ά(1-π)E may be so large as to render the project uneconomical. In that sense, the RRT is non-neutral. Since the distortion term is directly dependent upon the riskiness of the project, as captured by the probability of failure (1-π), the RRT can be seen as a tax on risk taking.
4.9 REGULATION
A common method of distributing rents used by the government is regulation. Regulation can take many forms and guises. Individuals are often required to possess a licence before they can carry out certain actions legally. For example, fishermen are required to hold licences to fish certain species, rate-payers need a licence to own a dog and drivers require a licence to operate a taxicab. Regulations which result in a reduction in the supply of a resource yield rents. It is easy to show the effect of regulation on rents with the aid of Figure 4.6. In this case, the supply of the resource (S) is horizontal, implying that it can be produced at constant unit cost, where costs include an imputed return for entrepreneurial capital. The demand (D) for the resource is downward sloping implying a decreasing willingness to pay for additional units. In the absence of regulation, the equilibrium price would be P1 and the quantity produced and sold would be Q1. Here, the producers of the resource are earning a normal return on their outlay; the revenue received from the consumers just covers the cost of producing the units.
Suppose now that the government introduces a licensing system which has the effect that only Q2 of this resource can be sold legally. Due to the restriction in the supply, the price of the product will be driven up to P2. Producers who have the licence will receive in total OP2AQ2 in terms of revenue, which exceeds their cost of production by P1BAP2. This area represents the rents arising from the restriction in the supply.
If the government had sold the licences in the market by means of an auction system, then these rents would have ended up in its coffers; the rents are simply transferred from the producers to the government. (If these licences are viewed as permanent, then individuals would be prepared to pay up to the present value of these rents.) The restriction in supply implies that the factors of production previously used in the production of the resource can be used to produce goods elsewhere in the economy. A measure of this increased output is given by BCQ1Q2. Consumers, however, value the units between Q1 and Q2 as ACQ1Q2, so the loss in willingness to pay exceeds the cost of the resources released by ABC. This area represents an uncompensated loss to the collectivity and is conventionally interpreted to be the welfare cost of the regulatory action.
If the government had chosen to give away the licences, then all that changes is the distribution of income. The individual who first receives a licence earns a rent as a result. If the licences are permanent and resaleable, then a market for them will emerge and the holders of existing licences can expect to receive the present value of the future stream of rents from it. Individuals who buy these licences will not make above normal returns any more; the whole stream of rents has been captured by the original licence holder.
Regulations that lead to rents and subsequent changes in asset values through capitalisation of these rents are plentiful. For example, liquor licence laws that restrict the number of pubs that may operate in an area confer rents to the existing licensees. The market value of any existing pub depends heavily upon the amount of rent generated by the restriction in supply. If one purchases an existing licence, however, there are no more rents to be earned by the new owner. The question of who is the beneficiary of the restriction is of importance in the issue of deregulation. One often hears that the publican has in effect "a licence to print money" and that the industry is sorely in need of deregulation. Matters are not so simple, however. The move to deregulate reduces or even eliminates the value of the licence and, as such, holders of second-hand licences suffer a capital loss. (25) Before deregulation, these individuals were earning a normal rate of return. They paid the capitalised value of the rents for the licence in the expectation that these rents would be permanent. If the government subsequently changes the regulation, then these expectations are falsified and consequently, they will have paid too much for the licence. The question of whether or not these individuals should be compensated is addressed in Chapter 6 below.
ENDNOTES
15. The interested reader is referred to Kareken (1986) for a survey of recent contributions in the deposit insurance literature.
16. See Carew (1985, ch. 6) for an insightful discussion of the origin and role of merchant banks in the Australian financial markets.
17. The problem of dynamic inconsistency was discovered by Strotz (1955-6). Dynamic inconsistency occurs, for example, if the government's announcement regarding future policy is not believed because agents know today that in the future the government will wish to act differently from what it announced in the present. An illustration, from Fischer (1980, page 94), may clarify the problem. At the beginning of a university course, the optimal policy is to announce to the students that there will be a final exam at the end. On the morning of the exam, the optimal policy is to cancel the exam. This saves the students the effort of writing the exam and the instructor the chore of grading. Hence, the optimal policy is inconsistent. A consistent policy would be to announce that there will not be a final exam. This, on the other hand, is sub-optimal because the students are not likely to work as hard if there is no final exam. The dynamic inconsistency problem can be overcome by using a consistent but sub-optimal rule. In the exam case, this would be to have a final exam no matter what. Constitutional rules, such as those discussed in Chapter 6, may fruitfully be seen as examples in this regard.
18. See, for example, Barro and Gordon (1983) and Kreps and Wilson (1982).
19. This is under the presumption that all individuals are risk neutral. Risk neutrality implies that agents are equally happy with a fifty-fifty bet of $0 and $10 on the one hand and a certain return of $5.
20. Area EGH is equal to BCH so that ECO is equal to OBGE. Since benefits under the auction system are AEG plus OBGE, the extra benefits due to the auction system are equal to AEG.
21. As we show in Chapter 5, this conclusion must be modified if rent-seeking behaviour emerges. For example, agents could attempt to lobby the government in order to make it more likely for them to receive a unit of the good. If this lobbying takes up real resources, then additional waste can arise.
22. Again this conclusion must be altered if rent seeking over the subsidy occurs, as we show in Chapter 5.
23. This argument was formulated by Nichols, Smolensky and Tideman (1971) in the context of merit goods.
24. This reasoning is based on the revealed preference notion. Strictly speaking, the argument in the text can only show that the private option is at least as good as the gift in kind. Most economists would assume that the private option is strictly superior to the gift in kind in terms of welfare.
25. Just as they would make a capital gain if a sudden wave of alcoholism, unforeseen at the time of the purchase of the license, hits the community, leading to an expanded demand.
No comments:
Post a Comment