The methodology for testing suggested by the theory seems to require that we must find data which would enable us to determine how much output would be produced and at what average or marginal cost under competitive conditions and under the ‘market’ conditions represented by the principal-agent (the government-bureau) interface. There is a methodological problem which should be apparent from the earlier analysis in Section 7.2.1. Such a comparison is impossible if the goods are ‘pure’ public goods, for such goods cannot by their very nature be produced and sold under competitive conditions in the free market. The nearest we might get to finding an appropriate comparison would be by comparing a country with a highly centralised (and therefore monopolised) production of a public good such as law and order with a country in which such services are decentralised to ‘competing’ lower layers of government or by seeking out an example where a country has changed suddenly from a highly centralised to a highly de-centralised form of provision (or vice versa).
Even if it were possible to find an approximation to competitive conditions, the problems of empirical investigation would not end there. The study would depend heavily on information supplied by those who have a vested interest in the results the bureaux – and who may refuse to supply the information or, if obliged to do so, may supply it in an incomplete or misleading form. In order to discover the equivalent of a competitive optimum, it is crucial to be able to identify a ‘demand curve’ for the service in question which again can only be done by direct enquiry from sponsoring agents – an expensive matter – rather than from examination of some data series. Even assuming that reasonably satisfactory results could be obtained, there must be argument about their interpretation. For example, if it had been ascertained that average costs were higher in a centralised as compared with a de-centralised country, this would not necessarily be evidence that the bureau in the centralised economy was less efficient. The production functions in the engineering sense could be identical (i.e. the same physical resource inputs would produce the same output at every level of output in each country), but if wage rates of bureau employees were higher in the country with a centralised government, average costs would be higher. The bureau in that country might reasonably argue that if wage rates are fixed by national collective bargaining, then they are not wholly within their control. Finally, the most frequent ‘attack’ on empirical work of this kind is that it is difficult if not impossible to achieve consensus on the measures of output, and to measure differences in the quality of output.
It is not surprising to find that most of the evidence collected about the relative efficiency of public and private output is concerned, first, with comparative costs of production and therefore with technical efficiency, and, second, with services which are not public goods in the strict sense but which happen to be produced in both the public and the private sector. Examples are waste treatment, garbage disposal, fire protection, provision of health insurance, airlines, bus transport, construction, office cleaning and forestry. Almost without exception, these studies indicate that costs of production are lower in the privately operated services compared with the publicly operated services.
A striking example discovered by the US Government Accounting Office is provided by the maintenance and repair costs of naval oilers and commercial tankers:
‘Navy support ships are, in many ways, similar to merchant marine vessels; for example, fleet replenishment oilers, which refuel navy ships at sea, have characteristics much like merchant tankers. In fact, between 1971 and 1977, commercial tankers under contract conducted more than 90 underway refuelling operations for the Military Sealift Command. There are, however, substantial differences between commercial and naval ships in the amount of time spent at sea and in the cost of maintenance. Although navy support ships are on sea duty only 20 per cent of the time, comparable commercial vessels are under way between 40 and 70 per cent of the time. Because the wear and tear on equipment is greatest when a ship is at sea, it would be expected that, for comparable equipment, commercial vessels would experience maintenance costs much greater than naval ships, on the average. Such is not the case. Overall, the GAO found that the “Navy's maintenance costs per ship average about $2 million a year compared to about $400 000 a year for a commercial ship”. The navy/commercial maintenance costs ratios for similar equipment items range from a low of 3.3:1 to a high of 51.8:1, which clearly indicates that commercial maintenance costs are much lower on comparable equipment than naval costs. In addition, whereas the typical naval support ship spends between 30 to 68 days in repair each year, the commercial counterpart is out of service for only 11 to 31 days annually. This finding is particularly surprising, for a naval ship is of little, if any, strategic value while in port undergoing maintenance. One would expect that, as much as possible, naval repairs would be made during operations at sea. However, the reverse is true in that commercial ships make as many repairs as possible while under way and naval ship repairs are typically made in port. This is especially unanticipated in view of the fact that commercial vessels carry a crew that is only a fraction of the size of the crew on naval ships. Among the 86 ships in the sample of commercial ships, the largest number of staff was 46, whereas the smallest crew size among the 58 naval ships was 92; overall naval crew size (exclusive of troops) ranges as high as 1438 with an average of 555.
The General Accounting Office report concluded that the Navy could reduce maintenance costs dramatically without significantly affecting readiness of the fleet by adapting commercial maintenance practices. Some years ago, it might be noted, naval shipyards ceased construction of new naval ships and all new construction was contracted out to private yards because construction costs were significantly lower at private facilities that were awarded competitively bid contracts.
Source: Bennett and Johnson (1980)
