Information for Social Welfare Policy: Toward A Second Revolution
In a typical year, the U.S. Congress considers about 5,000 bills and enacts about 250 pieces of legislation (Bureau of the Census, 1990b:255). These laws mandate new government services, modify or drop existing programs, raise needed revenues, and otherwise meet the constitutional mandate to "provide for the common defense [and] promote the general welfare." Behind each piece of legislation is a complex political process that begins long before a bill is formally introduced into the Senate or House of Representatives and continues long after it is formally signed into law by the President. Major stages in the policy process include:
identifying a problem, such as rising health care costs or declining productivity;
putting the problem onto the political agenda, that is, making the case that action is needed and appropriate on the part of the federal government;
identifying and evaluating alternative policies and programs to deal with the problem;
building agreement on a program and obtaining sufficient support to enact it formally into law;
implementing the program; and
evaluating the program, which often leads to identifying a new problem that may require further governmental action.
Each stage of the process usually involves many actors—decision makers in the executive and legislative branches and their staffs, researchers and policy
analysts from academia and the private sector, interest groups, and the media—and the entire process may span months, years, or even decades. Sometimes the process works in a straight line from problem identification through legislation to program evaluation, but more often it encounters detours and roadblocks along the way.
Each stage of the process also involves the use of information. One kind of information is political. Political information, which is vital for developing strategies and tactics to facilitate the process, concerns which actors support or oppose certain policy options, how open key actors may be to changing their positions and what factors could induce them to change, trends in public opinion, and similar matters. Another kind of information is policy oriented. It concerns the characteristics of policies and programs and plays a role in helping decision makers choose among options at each stage of the process. This report is concerned with the use of the latter type of information for decision making about social welfare policies.
Policy-oriented information not only enters each stage of the political process, but is often embodied in legislation itself. Formula grant programs allocate many billions of dollars of federal appropriations to states and localities, not in predesignated amounts, but on the basis of their characteristics as determined in the decennial census or other data sources. For example, the Job Training Partnership Act allocates funds on the basis of a formula that uses each jurisdiction's share of the unemployed and economically disadvantaged populations, and the Medicaid program determines the federal matching percentage of state expenditures partially on the basis of measures of each state's per capita income.
By policy-oriented information (which we henceforth refer to as "information" or "policy information"), we mean input to the social welfare policy process that has the following characteristics:
The information is "objective": that is, it is based on evidence rather than on opinion or preference. Ideally, the evidence is quantitative, such as the results of a survey, demonstration, or experiment; however, in many instances, the available information may be largely qualitative or based on such a small number of observations that it is difficult to draw conclusions from the results.
The information represents a combination of data and analysis: that is, the information is not simply raw numbers, but the result of an analytical process that assigns meaning to the numbers and assesses their relevance and their limitations. (In this regard, we adopt something parallel to the convention of the Central Intelligence Agency, which defines "intelligence" as "data plus analysis.")
The information relates to the social welfare aspects of policy choices, such as which population groups will be affected by a policy change and why: that is, for the purposes of our study, we are dealing neither with political information—which can also be objective and combine data with analysis—nor
with other kinds of policy-oriented information, for example, information about the environmental effects of policy choices.
Part I of our report considers some of the key issues involved in improving the evaluation of alternative legislative proposals—what we term "policy analysis"—whether the means to the evaluation is microsimulation modeling or some other technique such as cell-based or macroeconomic modeling. In Chapter 2 we present a brief history of the first information revolution, which concludes by calling attention to the resource constraints that limited investment in models, data, and research in the 1980s, with adverse effects on the quality and cost-effectiveness of the tools that support the policy analysis function. Policy analysis represents a link between social science research and the political process. We next review some general characteristics of each end of the link, which are important to understand in order to develop recommendations directed to improving the policy analysis function itself. We then illustrate the capabilities and limitations of microsimulation vis-á-vis other policy analysis techniques through a case study of the Family Support Act of 1988, identifying some of the problem areas.
In Chapter 3 we offer a strategy for investment to bolster the capabilities of policy analysis tools and the quality of policy analysis information to address the issues of the 1990s and beyond. We discuss in detail three key ingredients to policy analysis that we believe require attention regardless of the analysis tool that is used: data, validation and documentation of analysis tools, and communication of the results of analysis to policy makers. (The Appendix to Part I provides technical information related to validation and communication of the uncertainty in model estimates.) We recommend improvements in the underlying databases, a great expansion in the effort devoted to evaluating and documenting the estimates produced by models, and a commitment to finding useful ways to communicate the results of validation studies in an understandable form to decision makers in the executive branch and Congress. These actions are all necessary elements of the second revolution that we believe is needed in the world of policy analysis.