North American Poverty Study

In Canada and the U.S. alike, there’s a well-developed infrastructure for counting the number of people in poverty, but less attention is paid to monitoring the changing conditions of poverty, how those conditions differ across places and “poverty types,” and how those in poverty respond to these differing conditions.

The North American Poverty Study (NAPS), which will be carried out simultaneously in the U.S. and Canada, addresses this problem with an innovative new infrastructure for monitoring poverty. 

The NAPS is founded on two premises: (a) that existing quantitative protocols for measuring poverty tell us relatively little about the day-to-day experience of poverty; and (b) that the U.S. and Canada run a one-size-fits-all policy in large part because relatively little is known about the different types of poverty. We’ll review each of these two problems – and our proposed solution to them – in turn.

We should forewarn: The NAPS is ambitious and unprecedented. We therefore spend some time introducing the rationale for carrying it out and its expected payoff for basic science and policy.

It’s Not Enough to Simply Count the Poor

The current poverty measurement infrastructure in the U.S. and Canada presumes that the main job is simply to count how many people are poor. With the introduction of the Supplemental Poverty Measure (SPM) in the U.S., we know much more about the number of people in poverty, which regions have high poverty rates, and which demographic and racial groups have high poverty rates. Although it’s obviously important to count the number of people who are poor, this approach falls short because it presumes that the SPM threshold correctly establishes where deprivation begins and ends, because it fails to identify the gradations of experience above or below that threshold, and because it fails to provide any information about the experiences of those in poverty as they attempt to forge their lives in the context of deprivation. 

What we lack, then, is an infrastructure for systematically examining how the poor live and how they live their lives in the context of stress, disruption, and deprivation. The NAPS, which is built around the Making Ends Meet protocol (as devised by Kathy Edin and Laura Lein), will provide a new infrastructure for regularly delivering qualitative interview evidence. The interviews are based on modules that elicit (a) a short life history and defining life-course events; (b) the features and rhythm of the subjects’ everyday lives; (c) the availability and closeness of friends and relatives; (d) a detailed breakdown of income from all sources and expenses across detailed categories; (e) the types of formal, informal, and illegal jobs that have been and are typically held; (f) the extent of economic hardship (e.g., difficulties in paying bills) and characteristic reactions to such hardship (e.g., payday loans, turning to friends); (g) the poverty standing of the subjects (using modified government protocols); and (h) the mental health, anxiety, and stress experienced by the subjects. The study will draw representative samples from three poverty populations (extreme poverty, middle poverty, near poverty) along with one middle-class comparison group. Because the interview is long and typically intense, our pretest interviewers have come away with a rich and deep portrait of the subjects and their everyday lives.

The NAPS will become a key resource for basic and policy research on poverty. The following are the most obvious payoffs:

Validity of poverty measurement: The U.S. has mobilized much scientific talent and considerable resources to roll out the new Supplemental Poverty Measure. But the threshold itself is based on little more than convention and hunch. The NAPS will allow us to validate this quantitative monitoring system by examining how low-income people are actually living and whether the SPM and other measurements (e.g., the Official Poverty Measurement) are representing who is and is not experiencing deprivation.  It will likewise allow Canada to develop an official poverty measure that’s grounded faithfully in the empirical evidence on the cutpoints demarcating real and true deprivation.

Gradations of poverty: In recent research, it has been documented that extreme poverty in the United States (i.e., less than $2 per person per day) is surprisingly common, a finding that calls on us to better understand the circumstances of this population. But there’s very little evidence on other types of qualitative differences between various poverty populations. Do the “near poor” indeed face qualitatively different circumstances than the working class or middle class? Are they in the end just like the poor? Are deep poverty and extreme poverty pretty much the same? Do the extreme-poor, middle-poor, and near-poor populations have different lives, experiences, and strategies for making ends meet? Do existing poverty policies address their needs equally well? We simply don’t know.

Hidden poverty populations: Within the qualitative tradition, the tendency has been to focus on particular poverty subpopulations, especially young men, young women with children, and the homeless. By insisting on systematic representative sampling, the NAPS will allow us to overcome this patchwork approach and provide new evidence on subpopulations that remain largely hidden, including childless women, middle-aged adults, the elderly, and doubled-up families.

Retesting key conclusions: Even within commonly studied populations, it is uncommon for qualitative studies to rely on random samples, meaning that systematic biases in the sampling method may affect characterizations of those populations. Likewise, an explicit middle-class comparison group is not always used, a design that lends itself to mistaking certain features of the poverty experience as more idiosyncratic than they truly are. We need an infrastructure that allows us to ensure that the inferences made from classic and highly influential qualitative studies can stand up to rigorous test based on systematic samples.

Early warning system: The NAPS will provide a new infrastructure for undertaking real-time assessments of how the poverty population is dealing with national emergencies (e.g., recessions) and sweeping new policy (e.g., health care reform, school reform). In the last recession, the U.S. needed to make real-time assessments about unemployment insurance, food stamps, disability insurance, and other key programs, but there was little research capacity to make those assessments. 

Develop new theories and interventions: The NAPS infrastructure, once in place, will allow new interventions to be developed from the bottom-up on the basis of qualitative evidence on (a) the everyday lives of the extreme poor, moderately poor, and near-poor, (b) how they react to immediate or long-term need, and (c) how they engage with the informal (or formal) safety net. 

The simple conclusion: During the next several decades, Canada and the U.S. will be making major policy decisions on the safety net and labor market, decisions that will have to be reached without adequate evidence unless we develop an infrastructure for understanding how the poor live, make decisions, and react to policy changes.

One-Size-Fits-All Policy

If the first main problem with U.S. and Canadian poverty measurement is that they privilege counting over understanding, the second main problem is that they operate under a monolithic, uniform conception of what poverty is. With few exceptions, poverty policy is based on a single stylized image of poverty (i.e., urban concentrated poverty), a tendency that will be more easily resisted when the NAPS is available.

The NAPS will take on this problem very directly. At its inception, it will cover 14 sites across the U.S. and 8 sites across Canada, each representing a distinct type of poverty. These sites cover such poverty types as deindustrializing poverty (e.g., Detroit; London, ON), border poverty (e.g., Nogales), suburban poverty (e.g., Los Angeles), ethnically concentrated poverty (e.g., Chicago), classic Black-White poverty (e.g., Washington-Baltimore), depopulating poverty (e.g., New Orleans), high-density poverty (e.g., New York), high-housing-cost poverty (e.g., Toronto; the Bay Area), rural deindustrialized poverty (e.g., Lowell, MA), supply-side poverty (e.g., Racine, WI), new immigrant poverty (e.g., Greenville-Spartanburg-Anderson, SC), rural White poverty (e.g., Harlan County, KY), rural Black poverty (e.g., the Mississippi Delta), and First Nation and spatially isolated poverty (Yellowknife; Pine Ridge, SD). These sites were principally selected to represent the key types of poverty in play in Canada and the United States, but a further objective was to privilege sites that represent these various types in as ideal-typical a form as possible. In most of the sites, a major university is collaborating with us, and a top scholar at that university has agreed to lead the data collection and analysis.   

The results coming out of this 22-site design will allow Canada and the U.S. to test and assess existing and proposed policy against each poverty type. If our policy is sometimes anemic, it’s in part because our image of poverty is anemic: We can better develop policies that are matched to real need and tailored to real diversity once we better understand that real need and real diversity.

It bears noting that local governments and agencies have also been very keen to participate in the NAPS because they too find themselves obliged to run poverty policy in the blind. By virtue of the NAPS, a wealth of evidence about poverty will become available at each site, making data-informed policy possible at the local level for the first time. This new resource additionally will provide rich opportunities for research and training among graduate students and other scholars.

Next Steps

This initiative has required three years of planning and pretesting because it develops an altogether new infrastructure for monitoring poverty. It innovates by (a) developing a common protocol that allows for cross-site and over-time comparisons, (b) relying on representative samples that allow us to generalize to a known population, (c) including a full constellation of poverty populations (i.e., the extreme poor, the middle poor, the near poor) as well as a middle-class comparison group, and (d) making the anonymized NAPS interview transcripts and community metadata available to all scholars (via a secure data center). If the qualitative form has sometimes developed in a way that has made replication difficult, it is because of its own developmental path rather than any intrinsic inconsistency with scientific methods. 

Because the NAPS methodology is so new, there is still some additional planning to complete, although we have already administered pretests in Tucson and the San Francisco Bay Area. There are four remaining phases to the initiative:

Planning phase: The first phase involves finishing the planning process by further refining the qualitative protocols, finalizing the sampling plan, and convening an expert panel to advise on other remaining methodological issues. We also will develop a plan for data security during this planning phase of the project.

Preparation phase: In the second phase of the project, we will lay the groundwork for the data collection by (a) recruiting interviewers for each of the 22 sites, (b) designing and implementing a training program for them, and (c) developing a public-facing website.

Data collection: The data collection phase will entail in-person and household interviews, videotaping, and securing additional metadata.  It also will include coding and processing of all study data.  Throughout the interview and analysis process, strict protocols for ensuring privacy and confidentiality will of course be insisted upon.

Data processing, analysis, and documentation: The fourth phase entails coding and analyzing the data, developing final data products and documentation, and disseminating the study data publicly.

Why We Should Do It

It is no small decision to build a new infrastructure for monitoring poverty. Is it worth it?  Because the social and economic costs of poverty are so high, and because these costs loom large as the U.S. and Canada attempt to compete against countries that pay a lower “poverty tax,” there is ample incentive to find ways to reduce poverty. The future economic success of both countries rests in no small part on whether the burgeoning low-income population is more successfully integrated into the economy.

The U.S. and Canada have been hampered in meeting this challenge because they don’t have a systematic infrastructure for monitoring how the poor get by.  We are left with one-size-fits-all policy because there is no capacity to examine how poverty plays out within reservations, border communities, suburban areas, deindustrializing cities, new immigrant communities, and many other settings. The NAPS solves this problem by setting up 22 poverty observatories that will allow us to develop and assess poverty policy on the basis of systematic evidence.