Common sense would suggest that public services would want to hire workers good at serving the public’s interest. Thus, hospitals would hire nurses good at nursing; police forces hire policemen good at policing; tax offices hire tax officials good at collecting taxes; and so on. Experience suggests something different. Most people have their own horror stories about poor public service. In December 2001 Canice Prendergast, a then young economist working at the U.S. National Bureau of Economic Research in Cambridge, Massachusetts, published a scientific working paper that neatly demonstrated the potentially perverse consequences of well-intentioned initiatives to give consumers voice and introduce some oversight into public services as means of guaranteeing higher levels of service. But the modest title of Prendergast’s paper belied the far-reaching implications of his study: ‘Selection and oversight in the public sector, with the Los Angeles Police Department as an example.’
Which is better, it might be asked; a police officer who is primarily worried about avoiding the conviction of the innocent, or one who worries most about letting the guilty go free? From a utilitarian point of view, the latter is preferable if the overall objective is to fight crime. On the one hand, mechanisms exist to reduce the probability of the innocent being convicted; on the other, a police officer determined to find evidence of guilt will be more efficient at fighting crime than one worried about harming the innocent. But the public would, quite understandably, not see things that way. Under such circumstances, Prendergast argued, ‘bureaucrats will be disproportionately selected from those who are sceptical of the demands of consumers.’
Such ‘bureaucratic bias’ is a generalized phenomenon, particularly where public services are given performance targets. ‘They’ may not know best from the public’s point of view, but they definitely know better from the point of view of the public interest, as it is defined by their government masters, and especially when defined in the form of performance indicators and targets. It is this discrepancy that can lead to inhumane situations in, for example, hospitals and immigration agencies; the services in question are, often with great conscientiousness, trying to maximize or optimize something different from what the public would consider maximal or optimal. Hence also the need for consumer voice and oversight; there has to be an appeal/complaints mechanism and there has also therefore to be a system of adjudication. And it is here that the perverse effects of good intentions are convincingly depicted by Prendergast.
Take, first, internal monitoring or oversight. The bureaucratic bias identified above immediately becomes a problem. Because bureaucrats are biased against consumers, there will be an innate sympathy with the bureaucracy and ‘too little responsiveness to the legitimate concerns of denied consumers.’ In effect, Prendergast argued, ‘the characteristics that make for good bureaucrats also make for poor overseers.’
But external oversight is not necessarily a better solution. External monitors are generally less well-informed than their internal counterparts would be and their attentions tend to be triggered by consumer complaints. External investigators therefore tend to illustrate another sort of bias; they concentrate on errors where consumers’ interests are somehow perceived to be harmed (the person incorrectly arrested, for example), but not on all possible errors (the criminal not arrested, for example).
Prendergast transformed these arguments into theorems and tested them mathematically by giving various factors theoretical values. He proved, first, that, indeed, under certain given circumstances ‘bureaucrats are optimally under-responsive to consumer complaints’ and that this bias is reinforced by internal monitoring, with its emphasis on the preferences of insiders. However, he also argued that internal overseers have a far greater capacity to discern where bureaucrats have made correct decisions, even though these would be invisible to the consumer. (Though, alas, internal overseers typically have little discretion to reward such good behaviour.)
External oversight, on the other hand, has two clear advantages: it can intervene ‘when society sees fit rather than relying on when bureaucrats see it as appropriate’; and it chooses ‘only those agents who have better incentives as oversight decisions are made with the objectives of society in mind and no longer depend on the make-up of the bureaucracy.’ But external parties are ‘more reliant on signals of wrongdoing before they intervene’ and, most perversely, ‘bureaucrats who face external monitoring can have the overarching incentive to avoid investigation.’ They do this, Prendergast argued ‘by simply giving the good to the consumer’; by not arresting the suspected drug dealer, for example, rather than risking a complaint about heavy-handed policing.
Prendergast then went on to provide the perfect illustration of his theorems in the form of the behaviour of the officers of the Los Angeles Police Department. Following lurid media coverage of the allegedly heavy-handed behaviour of anti-gang squads in the mid-1990s, external oversight mechanisms were introduced. The mechanisms for introducing complaints about police behaviour were greatly simplified and streamlined (above all, by introducing an online formula). All complaints were now investigated by an external agency. Penalties for malfeasance by police officers were increased. The immediate result was a huge increase in the number of complaints made. Subsequently, disciplinary actions and firings increased significantly and a number of officers left the force rather than face action. Presumably as a consequence, the police became less aggressive in their behaviour, as measured by statistics on the use-of-force data, officer-involved shootings, attacks on officers and arrest rates. But, another consequence – and herein lies the rub – in the late 1990s gang homicide rates soared by as much as 100 per cent, year-on-year. As Prendergast puts it; ‘since 1998 officers have been responding to increased oversight by actions which, although keeping them out of trouble, also result in higher crime.’ Where lay the public interest?
In conducting his study, Prendergast was primarily concerned with ‘understanding the constraints that limit the efficiency of institutions’. His article was inspired by the fact that ‘we have learned little about what motivates public sector bureaucrats’. By formally addressing ‘the environment in which bureaucrats operate’ he hoped ‘to generate a better understanding of (i) how bureaucrats are selected, and (ii) how their performance is overseen. In any case, he concluded ‘bureaucrats are inherently hard to monitor.’
Prendergast’s findings are surely of pertinence for the European Union’s institutions. To give an example, European Commission officials managing small project programmes in the early 2000s will recall the introduction of the new Financial Regulation, an understandable and entirely justifiable corrective reaction to the lax financial management that had led, indirectly, to the resignation of the Santer Commission. The Regulation’s emphasis was on sound and efficient financial management and the individual responsibility of programme managers (with an ever-vigilant OLAF hovering in the background), leading to lengthy contracts laden with exclusion and liability clauses. The new recruitment priority was the efficient financial agent bent on avoiding fraud and no longer the consumer-friendly desk officer bent on serving the consumer and thus enhancing the integration process. All of this made evident sense with regard to multi-million euro contracts and very little with regard to grants and subsidies as small as a couple of hundred euros disbursed to, for example, local employment agencies or individual students or well-meaning university departments and faculties. Throughout this process nobody was ill-intentioned and everybody had the public interest at heart. But was the public interest best served?