Federal Role in Decentralization

 

The role of the federal government in the proposed governmental design would be to perform only those governmental functions and tasks within these functions for which the federal government were clearly more efficient than subordinate governments. Included among these functions would be most of the original functions of the federal government such as defense and foreign affairs. While the initial decentralization would transfer many operational tasks of the newer multilevel governmental functions to subordinate governments, the federal government would retain some operational tasks in most multilevel governmental functions. In addition, for all tasks performed by subordinate governments the federal government would perform the research and development task in the separation strategy for governmental innovation.

To illustrate why the federal government would retain some operational tasks in multilevel governmental functions consider environmental regulation. Currently it is more efficient for corporations to lobby for federal environmental regulations than to lobby for state environmental regulations. At the federal level they can reduce their lobbying costs and achieve a single national standard less restrictive than the standards adopted by the most environmentally activist states. Thus the current political criterion for environmental regulation create economic incentives for placing environmental regulation at the federal level.

The decentralization criteria would shift some but not all environmental regulation to the states. To take a specific example, acid rain is a national and international problem which cannot be effectively decentralized to the states; hence, the regulation of the environment would remain a joint federal and state activity. For example, it is hard to argue that the federal government has economies of scale in regulating the internal environment of houses and offices because the variation in state standards would provide much useful empirical information concerning the costs associated with the various safety standards. This empirical information would lead to better performance through imitation. Thus the federal government would retain control over such environmental problems as acid rain and other air and water pollutant flows between states and national borders. The states would have operational control over environmental standards and incentive systems within their boundaries as long as they meet the pollutant standards at their borders.

Similarly, while the regulation of most aspects of industrial safety would be transferred to the states, some aspects would not be transferred. Nuclear power regulation would not be decentralized because the consequences of a nuclear meltdown can be so catastrophic over such a large area.

However, the decentralization of governmental tasks would increase the role of the federal government in coordinating these tasks performed by subordinate governments. For example, as was pointed out decentralization of more environmental regulation to the states would increase the need for federal coordination. And in its role as a coordinator the federal government would be subject to a professional review as in all its other tasks. For example, under the professional review if the federal government tried in induce all states to raise the drinking age from 18 to 21 by threatening to withhold highway trust funds, many states would immediately file professional review suits for decentralization. At the very least some states would win experimental variances.

Also, for all decentralized governmental tasks the federal government would assume the research and development task  for governmental innovation. Hence the federal government would continue to perform systematic experiments to test drugs and the impact of chemicals on the environment. Also, the federal government would perform more social experimentals to test alternatives theories underlying policy choices. But because of the risk factor not all policy alternatives could be systematically tested prior to implementation by subordinate governments. In these cases the federal government would systematically analyse the consequences of subordinate governments adopting different objectives and policy alternatives.

Whether or not decentralization would increase empirical learning about the decentralized government policies would depend on the choices made by subordinate governments. To be sure, decentralization would give subordinate governments a larger role in estimating the common weal; consequently, decentralization would increase the number of estimates of the common weal which could be selected simultaneously. In itself, however, decentralization would be insufficient to guarantee that the choices made by the states would exhibit enough variation to provide useful empirical information concerning alternatives, because every subordinate government could conceivably choose the same alternative. Variation would be likely to occur only if there were variations in political philosophies and a number of competing theories about how to achieve objectives.

Under the proposed design several factors would lead to greater variation in the estimates of the common weal and of the associated goals and means. First, as will be discussed the redistricting of states each decade would result in greater variation among the states. In addition, as freedom of  location increases, individuals would be more likely to select a state, metropolis and town which would match their tastes, thus accentuating the differences among the subordinate governments. Because there would likely be more political variation among states than there is currently, the subordinate governments would be likely to vary in their estimates of the common weal. And assuming the current controversial nature of the social sciences continues for the foreseeable future, the selection of policies would also likely vary among subordinate governments.

While differences among states, in terms both of political philosophies and social theories, would produce considerable variation in policies, the amount of variation would likely be less than ideal in promoting empirical learning. Elected politicians would be interested in producing results by the next election, whereas the benefits of social experimentation would frequently be long-term. Thus incentives operating on elected politicians would encourage them to select the alternative with the greatest short-term performance. Alternatives which might produce much better long-term results, but which involve short term risks, would not likely be considered. In the proposed decentralization scheme, the responsibility for providing incentives for lower levels of government to empirically test alternative approaches would rest with the federal government. This is because under decentralization, the benefits in increased knowledge would be nationwide, whereas the consequences of accepting risk would be local. To encourage risk taking by lower levels of government, the federal government would provide subsidies to induce lower levels of government to try new alternatives. The federal government might fund part or all of the proposed experimentation, or might even provide a bonus to states for trying a new approach to solving a particular problem. To reduce its costs, the federal government would encourage lower levels of government to apply for experimental variances. For example, it might be more cost effective to subsidize a major metropolitan government in applying for an experimental variance to test a new alternative than to subsidize a state government for the same thing.