|
Navigation About this web General Themes The Questions Economics Statistics Belief formation Russia Political Science Philosophy
|
Centralization and Decentralization by Hans O. Melberg
Introduction Classics I: The debate on Central Planning (and market
failure) Marx criticism of capitalism, of course, is only one side of the coin. On the other side there were people who pointed to the advantages of decentralization. Among them were F.A. Hayek, L. von Mieses and many others. There were many arguments involved. For instance that central planning would be too complex, that central planning was bad because it was impossible to know how to allocate resources without prices, that some kind of information is impossible to pass on, that there was a great incentive to distort information in order to get a low target in the "Plan" and so on; and even if it was technically possible it rested on a naive conception of bureaucrats being concerned with the welfare of all and not - as everybody else - mainly with their own welfare. In short, centralization was bad. These criticisms, in turn, provoked more sophisticated analysis of the problem of central planning. Leontieff devised a mathematical tool (input-output analysis) for handling some of the complexities of central planning. Oscar Lange showed that the planner might use "shadow prices" in their planning based on the prices they observed in the free market abroad. It might be worth the effort to read these authors too in order to know more about the good and bad effects of centralization. And, the collapse of the Soviet Union has not ended the debate. John Roemer, for instance, has written a book in which he argues for more state intervention in some areas. In short, the old and new debate about central planning and more generally about market failure is probably one source of information. Classics II: Bureaucracy Weber may be the classic source on bureaucracy. Another slightly more recent source is the writings of Anthony Downs (e.g. the book "Inside Bureaucracy"). Many others might be mentioned in this same tradition (Imperial economicsts): Mancur Olson ("The Logic of Collective Action"), G. Tullock, J. Buchanan, W. Niskanen. These writers tend to emphasise the bad effects of centralizing bureaucracies and they try to explain why the bureaucratic structures survive. This may be because of collusion between bureaucrats and politicians who both need each other and/or rent seeking behaviour that is not prevented because of unorganized opposition (because prisoner dilemma type incentives makes it unprofitable for the "majority" to organize). Very relevant: Recent academic
articles Another very relevant article is “Queues and Hierarchies” by A.W. Beggs (published in The Review of Economic Studies, 2001, pp. 297-322). He explores one possible reason for the existence of hierarchies: the desire to economize on skilled workers. For instance, “deliveries of babies are usually dealt with by mid-wives.” But, if there is a problem a junior doctor is called; and if he in turn, may call a more senior doctor if faced with a task he cannot handle. This (hierarchical) system ensures that the most skilled workers are “matched” to the most difficult tasks. There is, however, also a cost in the sense that within this system there is a delay before the most qualified person is brought in. The delay entails a cost (sometimes it might be too late to call a senior doctor). The question is then to find the optimal size or organization of the hierarchy given the desire both to economize on skilled workers and to avoid delay. Of course, there may also be many other explanations for hierarchies. The already mentioned Radner has focused on the desire to exploit parallel processing (many people working on the same problem simultaneously), others have put emphasis on the need to monitor workers and some argue that hierarchies may solve coordination problems. In any case, Begg presents a detailed model of one possible mechanism and he also has a good list of references for further study. Probably relevant: Optimal Size I (of firms) Once again there is a whole literature on this. Transaction costs may not be the only factor. For instance, vertical integration might be motivated by externalities or simply a desire to have larger market power/reduce competition (and charge higher prices). Breweries owning pubs could be one example of both and this has been extensively debated in England where the authorities forced breweries to sell pubs (sometime in the 1980s I believe), but they did not force oil producers to sell gas stations. In any case, vertical integration could be viewed as centralization and as such the literature around this might be relevant. The same goes for the theory of conglomerates (desire to spread risk?) and the more modern opposite trend (firms should focus on a core activity in order to build competence). Possibly relevant: Optimal size of … Also within the European Community there is a debate not only on how large the union should be, but also on what level different types of decisions should be made. What decisions should be made at the union level, what should be made at the national level, what should be made at a local level and so on. Torberg Falch has written an article in Norwegain which gives some pointers to this debate (“Stat, Kommune og Velgere: Kommunesektorens politiske økonomi” I Norsk Økonomisk Tidsskrift no. 2, 1999) as well as a Ph.D on the topic. He presents a model by Seabright (from 1996) in which the problem with centralization is that it is more difficult to “discipline” the central authorities than local authorities in elections, but a centralized structure may be better at internalising externalities. The optimal mix then, is a compromise between the two concerns – making sure that the authorities do their best (since it is afraid of not being re-elected; the “disciplining” device) and internalising externalities. Some of his relevant reference include a book or a report by Begg et al who discuss “Making sense of subsidiarity: How much Centralization for Europe” (The Centre for Economic Policy Research, London, 1993). He also mentions a decentralization theorem due to Oates which says that given certain conditions local decisions are at least as effective as decisions in a centralized structure. At first look it sounded quite trivial and to call it a theorem might be slightly over the top. Be that as it may, the list of references is once again valuable for further research. Odd category Personal experience might be another source of information as most of us have encountered good and bad effects of centralization/decentralization. One might even find something of value in sayings like "The person who wears the shoe knows best where it hurts." There might be other relevant sayings, maybe pointing to an opposite tendency, for instance about how impartial friends may have a better idea of "what is best" when the judgement of the person himself is distorted by emotions like anger or love. Lastly, fictional literature, like Kafka's The Process might be of some use as a source of ideas about the problems of centralization and bureaucracies. Official documents One and a half ideas Now, my idea would be to examine what happens when you have longer chains of principal agent relationships. This is maybe not too original (although I have not seen it done formally). Especially it would be interesting to examine the degree to which the contract is inefficient depending on the length of the chain and the degree of asymmetric information. In what way is this relevant to centralization? Sometimes centralization implies a longer chain of principal agent relationships. Information passes through more and longer people before it reaches the decision-maker. However, sometimes it may also imply the opposite. say we eliminate a middle level and move the decisions it used to take up one level. This could be described as centralization, but it shortens the length of the chain. However, one might argue that the "further away" the person in the hierarchy is the larger the informational asymmetry between the principal and the agent. Here is one example: In Norway we pay compensation to reindeer-owners who loose reindeers because of eagles and wolves. As far as I recall it is the middle layer (fylket) and not the state or the county (kommune) that administers the claims for compensation (But I am not sure.) Recently there was a newspaper-report about why many of the claims could not be true - although compensation had been paid - since the eagles in question no longer lived in the area. Local people would probably have known this long before some far away administrator. This could be a good example of how the asymmetry grows larger the "further away" the principal is. The last half-idea would be to read mode about how the risk of making a wrong decision is decreased as the decision moves through many chains. Of course, the opposite could also be true since when there are many people in the chain it is easier to feel that you can avoid responsibility if anything goes wrong (or you may simply believe that by the time something comes to you so many people have looked at it already there is no point in doing it too hard even one more time). Examples include a quotation that was wrong in the national exam in Norwegian in high school in the year 2001 and an election poster with wrong information from Oslo Kommune in 2001 (claiming that some immigrants were eligible to vote when they were only eligible for local - not national elections. Reason: Some bureaucrat simply used the poster from last years local elections) Conclusion |