Decentralization and the search for policy solutions

Journal of Law, Economics, and Organization

Volumn 16, Issue 1, 2000, Pages 102-128

  Kollman, Ken ^a   Miller, John H ^b   Page, Scott E ^a  

^a UNIVERSITY OF MICHIGAN   (United States)

^b CARNEGIE MELLON UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0034390178     PISSN: 87566222     EISSN: None     Source Type: Journal    
DOI: 10.1093/jleo/16.1.102     Document Type: Article

References (32)

1. Page (1996) has shown that this method of formalizing problem difficulty correlates strongly with, but differs from, decomposability or separability.
2. Alternatively, we could use the present discounted values of continued search. Holding all of the parameters constant, the rankings of institutions do not depend on which measure we use. However, when we vary the distribution function F in order to improve the relative search ability of the central organization, our specific functional forms for when decentralization outperforms centralization will depend on the fact that we are comparing final outcomes. The findings would be similar if we were to choose to compare the discounted present value of search, but the analysis would be more complicated.
3. The distribution F captures search ability.
4. We assume, in other words, that there exists an incentive compatible contract where each subunit must demonstrate a policy with a value above v*.
5. However, when problems become highly nonlinear, even econometrics cannot guarantee an optimum. See Harris (1978) and Boeri (1990) for empirical evidence of nonlinearities on public projects.
6. Examples include knapsack, traveling salesperson, and minimal spanning tree problems.
7. For our assumptions about decision making, we borrow heavily from the organizational theory literature and its incrementalist traditions (Dahl and Lindblom, 1953; Simon, 1957; Lindblom, 1959; March and Olsen, 1984; Bendor and Hammond, 1992; Kollman, Miller, and Page 1992; Bendor, 1995). These models assume that organizations confronting difficult problems rely on relatively simple decision rules. They locally adapt.
8. Even if the target changes, after k attributes have been copied, if a new target emerges, all subunits agree with this new target on the first k attributes.
9. Searches, in the structural setting, need not be interpreted as periods per se. They can be interpreted as iterations of a thought process, or repeated attempts to derive the correct econometric model by analyzing data from the experiences of many subunits. Final policy values, moreover, can be interpreted as the value of the proposed policy following this period of contemplation or empirical research.
10. The standard error of the distribution is given in parentheses. Thus, the standard error of the mean estimate 735.5 would be 0.575 (which is the standard error of the distribution divided by the square root of the number of trials, 100).
11. Heterogeneity in small doses might have positive effects. It could add enough noise to the system to shake subunits off low-valued local optima and onto higher-valued local optima. In computational experiments, we found some evidence for this.
12. Bendor, Jonathan. 1995. "A Model of Muddling Through," 89 American Political Science Review. 89:819-840.
13. _, and Thomas Hammond. 1992. "Rethinking Allison's Models," 86 American Political Science Review 301-322.
14. Boeri, T. 1990. Beyond the Rule of Thumb: Methods for Evaluating Public Investment Projects. Boulder, CO: Westview Press.
15. Chang, Myong-Hun, and Joseph Harrington, Jr. 1998. "Organizational Structure and Firm Innovation in a Retail Chain," 3 Computational and Mathematical Organizational Theory 267-288.
16. Dahl, Robert, and Charles E. Lindblom. 1953. Politics, Economics, and Welfare. New York: Harper and Brothers.
17. Darr, Eric, Linda Argote, and Dennis Epple. 1995. "The Acquisition, Transfer, and Depreciation of Knowledge in Service Organizations: Productivity in Franchises," 41 Management Science 1750-1762.
18. Elazar, Daniel. 1987. Exploring Federalism. Tuscaloosa: University of Alabama Press.
19. Harris, R. 1978. "On the Choice of Large Projects," 11 Canadian Journal of Economics 404-423.
20. Holland, John. 1975. Adaptation in Natural and Artificial Systems. Ann Arbor: University of Michigan Press.
21. Kauffman, Stuart. 1993. Origins of Order. New York: Oxford University Press.
22. Kollman, Ken, John H. Miller, and Scott E. Page. 1992. "Adaptive Parties in Spatial Elections," 86 American Political Science Review 929-937.
23. Lindblom, Charles. 1959. "The Science of Muddling Through," 19 American Political Science Review 79-88.
24. March, James, and Johan Olsen. 1984. "Organizational Factors in Political Life," 78 American Political Science Review 734-749.
25. Morgan, Peter. 1983. "Search and Optimal Sample Sizes," 50 Review of Economic Studies 659-675.
26. Osborne, David. 1988. Laboratories of Democracy. Boston: Harvard Business School Press.
27. Page, Scott E. 1996. "Two Measures of Difficulty," 8 Economic Theory. 321-346.
28. Page, Scott E. 1997. "An Appending Efficient Algorithm for Allocating Public Projects," 64 Journal of Public Economics 291-323.
29. Page, Scott E. 1998. "Uncertainty, Difficulty, and Complexity," working paper, Santa Fe Institute.
30. Simon, Herbert. 1957. Administrative Behavior: A Study of Decision-making Processes in Administrative Organization. New York: Macmillan.
31. Stigler, George J. 1961. "The Economics of Information," 69 Journal of Political Economy 213-225.
32. Strumpf, Koleman. 1997. "Does Fiscal Decentralization Increase Policy Innovation," mimeo, University of North Carolina, Chapel Hill.