¹ This question is obviously a gross oversimplification of the problem. See Cnudde, Charles S. and McCrone, Donald J., “Party Competition and Welfare Policies in the American States,” this Review, LXIII (09 1969), 858–866Google Scholar.

² Easton, of course, has defined politics as “the authoritative allocation of values.” See Easton, David, A Systems Analysis of Political Life (New York: John Wiley), p. 21Google Scholar.

³ This measure is a direct statistical analogue of Lasswell's and Kaplan's indulgence-deprivation ratio. See Lasswell, Harold and Kaplan, Abraham, Power and Society (New Haven: Yale University Press, 1950), p. 61Google Scholar.

⁴ Fabricant, Solomon, The Trend of Government Activity in the United States Since 1900 (New York: National Bureau of Economic Research, 1952), ch. 6Google Scholar.

⁵ Fisher, Glenn W., “Determinants of State and Local Government Expenditures: A Preliminary Analysis,” National Tax Journal, XIV (12 1961), 349–355Google Scholar.

⁶ Bahl, Roy W. and Saunders, Robert J., “Determinants of Changes in State and Local Government Expenditures,” National Tax Journal, XVIII (03 1965), 50–57Google Scholar.

⁷ Fisher, op. cit., p. 353.

⁸ Dawson, Richard E. and Robinson, James A., “Inter-Party Competition, Economic Variables and Welfare Policies in the American States,” The Journal of Politics, XXV (1963), 265–289CrossRefGoogle Scholar.

⁹ Key, V. O., Southern Politics (New York: Random House, 1949), pp. 298–311Google Scholar.

¹⁰ Hofferbert, Richard I., “The Relation Between Public Policy and Some Structural and Environmental Variables in the American States,” this Review, LX (03 1966), 73–82Google Scholar.

¹¹ Four statistical terms used in this paper might be unfamiliar to some of the readers. A zero-order correlation coefficient is the simple, bivariate correlation coefficient. It is a summary measure of association. A partial correlation coefficient is a measure of association or the strength of a relationship between two variables—e.g., urbanization and redistribution—controlled for a third or more variables. Thus, it is a measure of the unique portion of the association after the common portion has been controlled. The multiple-partial coefficient is a summary measure of the “explanatory” power of a group of variables. It states the unique portion of the variance attributed to one set of variables—e.g., political variables—after another variable or set of variables—e.g., economic variables—has been controlled. The coefficient of determination associated with any of the three above measures is the square of the coefficient of correlation (simple or multiple) which is equal to the portion of variance explained by the measures.

¹² Dye, Thomas R., Politics, Economics, and the Public: Policy Outcomes in the States (Chicago: Rand McNally, 1966)Google Scholar.

¹³ Ibid., p. 293.

¹⁴ Lineberry, Robert L. and Fowler, Edmund P., “Reformism and Public Policies in American Cities,” this Review, LXI (09 1967), 701–716Google Scholar.

¹⁵ Grumm, John G., “Structural Determinants of Legislative Output”; paper delivered at Conference on the Measurement of Public Policies in the American States, Ann Arbor, 07 28-08 3, 1968Google Scholar.

¹⁶ Sharkansky, Ira and Hofferbert, Richard I., “Dimensions of State Politics, Economics, and Public Policy,” thia Review, LXIII (09 1969), 867–879Google Scholar.

¹⁷ Sharkansky, Ira, “Problems of Theory and Method: Environment, Policy, Output, and Impact”; paper delivered at Conference on the Measurement of Public Policies in the American States, Ann Arbor, 07 28-08 3, 1968Google Scholar.

¹⁸ Walker, Jack L., “The Diffusion of Innovations among the American States,” this Review, LXIII (09 1969), 880–899Google Scholar.

¹⁹ Jacob, Herbert and Lipsky, Michael, “Outputs, Structure and Power: An Assessment of Changes in the Study of State and Local Politics,” Journal of Politics, XXX (05 1968), 510–538CrossRefGoogle Scholar.

²⁰ Dawson and Robinson attempted to get at the concept of redistribution in the article already cited, but their perspectives were restricted by the measure of welfare orientation they chose and a concentration on levels of taxes and expenditures. See Dawson and Robinson, op. cit. We propose a considerably more comprehensive measure addressed specifically to allocations of benefits and burdens.

²¹ Tax Foundation, Inc., Tax Burdens and Benefits of Government Expenditure by Income Classes, 1961 and 1965 (New York: Tax Foundation, Inc., 1967)Google Scholar.

²² This shift raises a procedural problem, since the Census Bureau and the Office of Business Economics define revenue and expenditure categories somewhat differently. To counter this difficulty, we have either grouped the Census Bureau figures into categories at least nominally equivalent to those set up by the Office of Business Economics or we have used allocation bases which appear to be appropriate to the Census Bureau categories. A number of technical differences prevent a direct comparison between the categories so derived and the Office of Business Economics data, but comparability is not a major consideration, since it is only necessary to assume that the allocation bases are appropriate for the revised categorization of the Census Bureau figures.

²³ Sharkansky, op. cit., p. 4.

²⁴ Ibid.

²⁵ This classification is somewhat ambiguous since the federal government levies a tax on employers but allows a credit for state taxes up to 90% of the amount of the federal tax. Since most states have adopted a tax to take advantage of the credit, and in accordance with the Tax Foundation classification, we have considered unemployment compensation to be a state program.

²⁶ Hofferbert, Richard I., “Elite Influence in Policy Formation: A Model for Comparative Inquiry”; paper delivered at 1968 Annual Meeting of the American Political Science Association, Washington, D.C., 09 2–7, 1968, p. 8Google Scholar.

²⁷ U.S. Department of Commerce, Bureau of the Census, U.S. Census of the Population, 1960 (Washington, D.C.: U.S. Government Printing Office, 1964), p. 1–288Google Scholar.

²⁸ Industrialization is measured by one minus the percent of the work force engaged in farming, fishing and forestry work; drawn from ibid., p. 1–249.

²⁹ Urbanization is measured by the percent of the population living ia urban areas; from U.S. Department of Commerce, Bureau of the Census, U.S. Statistical Abstract, 1968 (Washington, D.C.: U.S. Government Printing Office, 1968), p. 367Google Scholar.

³⁰ Education is defined as the median school year completed by persons 25 years of age; drawn from U.S. Census of the Population, 1960, p. 1–248.

³¹ U.S. Statistical Abstract, 1968, p. 286.

³² The Gini index is a summary measure of the inequality of income in a given population. It is derived from a Lorenz curve on which the percentage of total income is arrayed along the y-axis and percentage of consumer units is arrayed along the x-axis. A line drawn at a 45-degree angle across the graph describes perfect equality, since a given percentage of the consumer units will claim an equal percentage of total income at points on this line (e.g., the lowest 10 percent of the consumer units have 10 percent of total income). The Gini index describes, roughly, the area between the 45-degree line and the line representing the actual distribution of income. The larger the area—the higher the Gini index—the more unequal the distribution of income in the population. The Gini index used in this study is from Verway, David, “A Ranking of States by Inequality Using Census and Tax Data,” Review of Economics and Statistics, XLVIII (1966), 314CrossRefGoogle Scholar.

³³ The participation index is defined as the votes cast for the state's Governor as a percent of voting age population. See U.S. Department of Commerce, Bureau of the Census, U.S. Statistical Abstract, 1963 (Washington, D.C.: U.S. Government Printing Office, 1968), p. 367Google Scholar.

³⁴ Measured in terms of average Democratic vote for Governor as in U.S. Statistical Abstract, 1968, p. 367.

³⁵ The measure of interparty competition is a rank order measure integrating the state's competition in the presidential, senatorial and gubernatorial races. For more detail on the measure, see Hofferbert, Richard I., “Classification of American State Party Systems,” Journal of Politics, XXVI (1964), 550–567CrossRefGoogle Scholar.

³⁶ The legislative inducements to participation index is a summed measure noting the extent to which each state has legal measures facilitating participation, e.g., absence of literacy tests and residency requirements, permanent registration, etc. See Milbrath, Lester, “Political Participation in the States,” in Jacob, Herbert and Vines, Kenneth (eds.), Politics in the American States (Boston: Little, Brown, 1965), p. 46Google Scholar.

³⁷ Lane, Robert, Political Life (New York: Free Press, 1959), p. 49Google Scholar.

³⁸ Key, op. cit., pp. 298–311.

³⁹ Schubert, Glendon and Press, Charles, “Measuring Malapportionment,” this Review, LVIII (12 1964), pp. 969–70Google Scholar.

⁴⁰ The measures of party cohesion and interest-group strength were derived from a questionnaire “sent to two or more competent persons in each state, including political scientists …, director of … research, agencies or bureaus, legislative officers and politicians. At least one reply was received from each of the 48 states—in most cases two or three.” See Zeller, Belle, American State Legislatures (New York: Thomas Crowell, 1954), pp. 190–192Google Scholar.

⁴¹ The measure of gubernatorial tenure is an index combining the gubernatorial length of term and the legal possibilities for re-election. The gubernatorial power index combines evaluative indices of budget powers, appointive powers, and veto powers. See Joseph Schlesinger, “The Politics of the Executive,” in Jacob and Vines, op. cit., pp. 220, 222, 226–27, 229.

⁴² Zeller, op. cit., pp. 190–91.

⁴³ Council of State Governments, Book of the States, 1962–63 (Chicago: Council of State Governments, 1962), pp. 178–81Google Scholar.

⁴⁴ Grumm, op. cit., p. 25. The index combined four important qualities of legislative life: (a) compensation of legislators, (b) total length of sessions in 1963–64, (c) expenditures for legislative services and operations, (d) a “legislative services” score.

⁴⁵ Walker, op. cit., pp. 882–883. The innovation index measures the rapidity and extent of adoption within the states of eighty-eight different programs. Such programs ranged from “the establishment of highway departments and the enactment of civil rights bills to the creation of state councils on the performing arts and the passage of sexual psychopath laws.”

⁴⁶ Schattschneider, E. E., The Semi-Sovereign People (New York: Holt, Rinehart and Winston, 1960), pp. 30–33Google Scholar.

⁴⁷ See for example Herring, Pendleton, The Politics of Democracy (New York: W. W. Norton, 1940), p. 362Google Scholar; and Key, V. O., Politics, Parties and Pressure Groups (5th ed.; New York: Thomas Crowell, 1964), pp. 696, 698Google Scholar.

⁴⁸ Fenton, John, People and Parties in Politics (Glenview, Ill.: Scott, Foresman, 1966), pp. 46–49, 50–78Google Scholar.

⁴⁹ For the purposes of this paper, the following states were considered to be in the “South”: Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, Texas, and Virginia.

⁵⁰ We have not used tests of statistical significance in this analysis, since all 48 states have been included.

⁵¹ Analysis of the impact of political variables in the 12-state area of the South is proscribed by limitations on the degrees of freedom. However, regression analysis of the lesser number of socio-economic variables permits some speculation about the South-non-South differences: the multiple co-efficient of determination for the socio-economic variables in the South is .74, higher than for either the 48 states as a whole or for the non-Southern states. Thus, in this analysis we have a situation in which the socio-economic variables account for more of the variance in each region examined than in the 48 states as a whole. Though a precise investigation of this anomaly is not possible within the statistical confines of the present study, we presume that the 48-state analysis contains a number of suppressed relationships which become statistically apparent only in the regional breakdown. The fragmentary data available to us support this presumption. An examination of the zero-order correlations between the socio-economic variables and redistribution in the South reveals that median income, industrialization, urbanization, and education have a strong negative relationship with redistribution. On the other hand, both the Gini index and the percentage of families with less than $3,000 in annual income have a strong positive relationship with redistribution. In short, redistribution varies inversely with what appear to be measures of ability to pay and directly with perceived need for redistribution. Since we would expect federal aid to vary in the same manner, with assistance provided to those states most in need and least able to pay, we assume that South-non-South differences in redistribution can be accounted for largely by differences in the impact of federal aid, and that this differential impact produces relationships obscured in the 48-state data. This conclusion can only be considered tentative and in need of more direct substantiation, but it suggests a fruitful area for future research.