Is the best practice really good enough?
Some of the problems arising in the course of performance analyses of the kind described above may be succinctly and somewhat provocatively summarized in the following five propositions, each of which will be followed by an explanation.
The objectives attributed to the operation of public enterprises (see also the ‘family approach’) do not reflect the specific purpose(s) of public enterprises; they are rather general economic policy objectives
In order to be able to analyse the performance of particular public enterprises such as railroads or post offices, one obviously has to define policy objectives of these public enterprises. In one particular case (Pestieau, 1987), this problem is solved by attributing the purported objectives of general public policy to the specific public enterprises. These objectives (in the case referred to here) are said to be efficiency, equity, financial balance and macroeconomic objectives. For the sake of this argument, it is not important which particular set of general policy objectives is attributed to the public enterprise. The principle of attribution, however, raises doubts. If all the different institutions responsible to a government are considered a family, Pestieau feels justified in suggesting that all members of the family have the same common objectives, and engage in the same activities (which is a non sequitur); that is, all government agencies are held essentially to subscribe to the same set of public policy objectives. This is a very curious assumption.
All the many institutions that comprise the state are, indeed, very different; they can achieve different purposes with different ease. One should therefore expect a division of labour among the different governmental institutions. Public enterprises are unique instruments of economic policy, since they allow governments to participate in markets and thus achieve specific results not attainable through non-market government behaviour.
Whereas the typical instruments of economic policy may affect economic activity in general terms, public enterprises can be directed to specific purposes, specific geographic areas or even specific geographic points, specific demographic or ethnic or other social groups or otherwise specific political or social problems that, for one reason or another, are deemed worthy of political attention. Given this bewildering diversity, it is (almost) never possible to assign or attribute objectives to public enterprises in general, and certainly not a priori. For each particular case, the objectives of the public enterprise in question have to be empirically established by way of conducting a three-step procedure: (i) an analysis of the structure of the public enterprise, (ii) an analysis of its revealed performance, and (iii) an analysis of the interplay between the enterprise’s own approach to account for its performance and the reaction of its principals to that account.The structure (or form) of the institution needs to be analysed, because the constitution of the firm reveals the intentions and purposes of its founders. The performance assessments need to be analysed in terms of all the criteria the corporation itself and its principals deem relevant for describing it. The process of justifying this performance in terms of how the company reports on its performance and how the principals, openly or tacitly, react will reveal and single out those criteria of performance assessment that can then be used by economists who want to undertake a performance study.
Efficiency is (almost) never in itself a policy objective of a public enterprise. It is unlikely (but not impossible) that we may find a public enterprise with essentially the same production function as an efficiently operated private enterprise
Efficiency is not an objective of private entrepreneurship. Why then should we expect efficiency to be a goal for public entrepreneurs? And what is ‘efficiency’ supposed to mean in such a context? Efficiency is not a goal, it is a condition for goal attainment.
While private enterprises can generally be described as being managed in order to maximize the net present value of the firm, this is not meaningful for public enterprises. In democratic societies, ownership rights in public enterprises change hands frequently (and regularly) with the change in government, and owners are better described as tenants. Since public enterprises are fairly flexible tools, they tend to be used for all kinds of short-term objectives, of which at least some are not meant to be publicly discussed. For instance, a public railroad company, apart from its obvious mission of passenger and freight transportation, plays a role in regional and local planning and development, in technology policy, in European and international politics, in policies vis-a-vis specific sectors such as agriculture and mining, and so on. The list of possible objectives is inexhaustible. It is typical of a public enterprise that its basic mission, such as railroad or postal service, has to be fulfilled at an acceptable level, where acceptability is politically determined. The remaining leeway can then be filled in with political discretion. Therefore a public enterprise has to be called ‘efficiently operated’ if the remaining room for discretionary politics is efficiently used up. The definition of public enterprise efficiency can then be stated in the familiar Pareto form: no further objective could be attained without compromising on any other (openly or tacitly) pursued objective. The performance of such an enterprise, then, has to be judged in terms of the attainment of those objectives, subject to the satisficing constraint. As is readily apparent, the performance approach followed by Pestieau, Tulkens and others is only concerned with what has here been termed ‘the satisficing constraint’.The performance indicators suggested so far do not reflect the purpose(s) of running a public enterprise
In principle, any government has the choice of running a public enterprise as if it were a private one. This principle has exceptions.
As has been pointed out before, the public owner of a public enterprise is the owner only for a definite period of time; he/she has the right of use, but not of complete ownership. However, the time horizon may well extend beyond the four or five years we tend to associate with the terms of office available to democratically elected governments. Even at the federal level, a party may well expect to be in office for two terms. In local and state government, much longer terms are often the rule. When several public bodies share in the ownership, much longer time horizons may result from the need to agree on a common policy. Under these circumstances, governments have the option to use public enterprises as a revenue source. Then the difference between a public and a comparable private corporation could conceivably only consist in the length of the planning horizon. The longer the planning horizon of the democratic government, the smaller the difference. The smaller the difference, however, between the objectives of a public enterprise and the common objectives of a private one, the more likely we are to find little difficulty in estimating their production functions. Public enterprises conducted for profit as a revenue source are the best candidates for the performance approach suggested by Marchand et al. (1984).Interestingly enough, we rarely observe governments pursuing this strategy, although there are important exceptions. This choice reveals a preference. The specific discretionary policy services rendered by public enterprises obviously are more important than the revenues which alternatively could be gained. The revenues forgone then yield an important hint for the assessment of the performance of a public enterprise. The policies pursued by spending the public enterprise rent (the rent equals the profits forgone) are off-budget and thus escape parliamentary scrutiny and public discussion. Reconstructing the policy function of a public enterprise by means of the revealed preference approach and, on the basis of this function, assessing the performance of the corporation is an important task for political economists.
For estimating the public enterprise rent thus forgone, the public enterprise efficiently operated for a profit becomes an essential benchmark. Here lies another important application of the technique designed by Marchand et al.Privatizations are not irrational
One of the most characteristic phenomena currently involving public enterprises is the almost universal wave of privatizations. Although privatizations assume many different forms, the phenomenon should clearly be seen as one particular form of conducting public business. Interestingly enough, the performance approach is silent when it comes to explaining privatizations. Pierre Pestieau (1987) creates the impression that the current wave of privatizations of public enterprises is somehow a sign of irrational politics. The property rights approach to the theory of the public firm, as sketched out above, does not agree with such a view (see Backhaus, 1989). As pointed out, the public owner of a public enterprise does not control the complete bundle of ownership rights; he/she is mainly restricted to the use of the firm. At the end of its term, the present government has to turn over the corporation to its political opponent. There is only one way to prevent this from happening and to capture a substantial portion of the value of the firm at the same time: privatization. By selling a public enterprise, the present government can capture up to the entire stream of earnings from an efficiently managed firm. There are three aspects to this political move:
1. future revenues are transferred into the present; hence privatizations are equivalent to bond issues (see Wagner, 1977, for a full development of this argument);
2. the present owner is the last one who can use the public enterprise as a policy instrument; often the privatization process itself is extremely involved and loaded with political objectives;
3. the political power of the opponent is curtailed by blocking his/her access to the public enterprise in the future.
As seen in this perspective, there is nothing irrational about privatizing public enterprises in the context of a politicoeconomic view of public entrepreneurship.
Best practice performance assessments amount to measuring without a behaviourally basedpoliticoeconomic theory
Unless there are identical political circumstances, the best practice of one (public) enterprise is not likely to be the best practice of another. Public enterprises are specific instruments of economic policy, and their objective functions can only be constructed after a careful analysis of their property rights structure and of the intentions of their political owners by means of the revealed preference approach. A performance analysis of a public enterprise is not possible without a theory explaining the political performance of these firms. Performance criteria, then, cannot be taken from economic discourse or from the observation of private firms operating in similar industries. Criticizing a public enterprise for not choosing the most advanced technique (best practice) amounts to ridiculing a hunter with bow and arrow for not using a rifle. The hunter will be amused, or worse, since he views with contempt those colleagues who gun down the deer at a distance instead of giving the animal a fair chance. Similarly, the way public enterprises discharge their services is sometimes as important as the service itself. Public enterprises serve multiple purposes, and the multiplicity of purposes certainly serves as a cover-up for a lot of slack. This is why performance analyses are important. They are meaningful, however, only if they are based on a realistic theory of public entrepreneurship. Such a theory is the precondition for deriving performance criteria, on which, in turn, the performance analysis has to be based. The ‘best practice’ does not meet that standard.