Introduction

One in four children under the age of six in the United States lives in poverty. Although child poverty is high nationwide, there is substantial cross-state variation. In the mid-1990s, for example, the percentage of poor children under age six varied from 11% in Utah to 44% in the District of Columbia (Bennett and Li, 1998). Across the states infant mortality ranged from 5 to 18 deaths per 1,000 live births; high school dropout rates ranged from 3% to 13%; and violent deaths ranged from 19 to 456 per 100,000 teenagers (Annie E. Casey Foundation, 1998).

Such variation suggests that state-level factors, including government policies, influence children’s lives. Extensive comparative research examines variation in social policy across industrialised countries, including the US. The impact of policy variation within the US has received much less attention. Scholars who do consider state variation typically focus on a single policy variable (such as AFDC benefit levels) and therefore fail to explore variation in how states’ packages of programmes – cash and noncash, transfers and services, targeted and universal – affect families and children.

In this chapter we propose a new approach to understanding how state-level policies may affect childhood poverty in the US. We begin by identifying opportunities for public intervention. Using this framework, we identify characteristics of certain government programmes that we would expect to influence families’ resources. Using state-level measures, we employ cluster analysis to identify five groups of states with similar policy packages as of 1994. We then describe the policy packages and examine variation in child poverty rates across the resulting clusters.

The comparative approach we use draws on several lessons from crossnational comparative welfare state scholarship. First, we explicitly characterise and compare policies in and across the US states. Second, we include multidimensional policy variables that go beyond public expenditures. We combine traditional quantitative measures (eg spending per recipient) and qualitative elements (eg programme rules) to capture the magnitude of public investment along with policy ‘architecture’. We consider multiple sets of policies, or ‘policy packages’ by analysing 11 programme areas and three important characteristics of each. Finally, we draw on cross-national scholarship about welfare regimes as a model for studying state variation in the US.