administrator, the registrant may amend his practices in accordance with the administrator's conception of justice and equity.35
If you find nothing wrong with the SEC bargaining with firms trying to sell stock, reread this quotation, substituting the word "police officer" for "administrator," "citizen" for "registrant," and "right to hold a protest meeting" for "effective date of registration."
In the next chapter we shall look at how regulatory agencies and Congress have changed their views as to the value of administrative discretion, just as legislators and courts have changed their views as to the value of police discretion. Common to both is a fear of arbitrary rule and an inclination to substitute rules for discretion. But we have already seen in the case of the police and prisons that it is not easy to know the circumstances under which rules will improve matters or what kinds of rules will achieve what effects. In chapter 18 we will find confirmation for what we have already seen in chapter 16 and elsewhere: In this country, we have a profound bias toward solving problems by adopting rules.
Neither inefficiency nor arbitrariness is easily defined and measured. Inefficiency in the small, that is, the excessive use of resources to achieve the main goal of an agency, is probably commonplace; but inefficiency in the large--the excessive use of resources to achieve all the goals, including the constraints--may not be so common. To evaluate the efficiency of a government agency one first must judge the value of the constraints under which it operates; to improve its efficiency one must decide which constraints one is willing to sacrifice. The best way to think about this is to ask whether we would be willing to have the same product or service delivered by a private firm. That is the subject of chapter 19.
If we decide that the constraints are important then we should be cleareyed about the costs of retaining them. Those costs arise chiefly from the fact that most bureaucrats will be more strongly influenced by constraints than by goals. Constraints apply early in the process: You know from day one what will get you into trouble. Goals apply late in the process (if then): You must wait to see if the goal is achieved, assuming (a big assumption) that you can state the goal or confirm its achievement. Constraints are strongly enforced by attentive interest groups and their allies in Congress, the White House, the courts, and the media; goal attainment is weakly enforced because an agency head can always point to factors beyond one's control that prevented success. Constraints dissipate managerial authority;