This chapter is a summary of the final session of the workshop providing brief closing remarks by members of the steering committee and Economic Research Service (ERS) sponsors. Panelists were asked to identify what they saw as unresolved questions and challenges. Panelists were John Cromartie (ERS), James Fitzsimmons (U.S. Census Bureau), Stephan Goetz (Pennsylvania State University), David McGranahan (ERS), Timothy Parker (ERS), Mark Partridge (Ohio State University), David Plane (University of Arizona), and Brigitte Waldorf (Purdue University). The panel discussion was followed by open discussion. McGranahan provided closing remarks.
PANEL REMARKS
Cromartie said he found very disarming Keith Halfacree’s statement that ERS is good at measuring remoteness, population size, and density, but then wondered why it should be labeled “rural.” If it is correct that applying the label is an obscuring last step, he asked how ERS can be more careful to keep in touch with what they are measuring. He also noted that Ken Johnson requested that ERS make available the continuous measure of remoteness, percent commuting, size of the largest city, and so on. Cromartie said that he agrees but doing so is a resource issue.
Fitzsimmons commended ERS for sponsoring this workshop. He said he thinks that the most difficult task facing ERS will be deciding on the purpose of any revision to rural area classifications. He said the workshop
discussions show some uncertainty about the issues. He asked if there is dissatisfaction with any or all of the ERS four classifications or whether ERS has a fundamental commitment to search for the best measure. He suggested that once the goals have been clearly articulated, coming up with a revision might not be difficult. He said that in preparation for the workshop, he looked at files from the late 1940s to figure out what ultimately became standard metropolitan areas. Reading minutes of meetings from 1946, he said he was struck by the fact that the committee had four classifications and they did not know how the four would fit together, whether they should fit together, or whether all should be replaced. The committee came up with an answer, metropolitan statistical areas, that has proven very useful.
He said his basic message was one of optimism. ERS needs to carefully articulate the goal, whether trying to come up with an additional or a single or a smaller number of classifications. Once that decision is made, Fitzsimmons said coming up with the classification would not be difficult. He noted that at the workshop, there has been talk about the ambiguity of some of the thresholds. He suggested that perhaps a group of people could work through the strengths and weaknesses of current and proposed measures.
Fitzsimmons posed one last question. He noted Plane had pointed out the advantages and disadvantages to federal classifications because of the requirement for consistency across the United States. If one were not constrained by this requirement, he asked about the possibility of coming up with a single classification system, and what might it look like.
Goetz remarked that this question is also something that motivated this workshop. What outcome variable should be targeted with this indicator, he asked. He noted that in the past it was poverty, and that has been very important. Is the outcome variable going to be a level measure, or a dynamic measure? For example, will places that have high poverty be targeted, or places that have high growth in poverty? Places where poverty is rising may be very different from those that already have high poverty. Or, he asked, will areas be targeted that are a combination of high poverty and high growth in poverty? Goetz suggested child mobility as an interesting variable to consider, as it is related to the prospects of future consumers and residents of a country. He said that perhaps disproportionately targeting mobility of younger people might help people move up the income ladder in place rather than having to move to cities in order to prosper. In the larger context, everyone is aware that inequality and widening gaps in income distribution are becoming a major concern, he said.
Going back to a point made by Mary Bohman in the opening session, McGranahan reminded the group that poverty rates are no longer higher in more rural counties as was observed in Hines, Brown, and Zimmer
(1975). Using county data from the five-year American Community Survey (2009–2013), he said it is apparent that this relationship no longer holds. Poverty has gone down in more rural counties and up in more urban counties, with little difference across these county groups. This disappearance holds using today’s rural-urban codes or the original one.
McGranahan stated that at least four major changes have occurred. First, farming has become a lucrative and complicated business. Many rural counties now have relatively high education levels as less well-trained farmers have left. Second, the decline in two-parent families has become a major cause of poverty, particularly among children. Single-parent families, more prevalent everywhere, are much less likely to have earnings that yield above-poverty incomes. Third, the recession was much harder on manufacturing than other industries, and manufacturing is relatively rare in more rural counties. Finally, Social Security payments rose considerably in the 1970s, lifting many people over age 65 out of poverty. More rural counties have always tended to have relatively large proportions of elderly. He said that the bottom line is that rural-urban codes no longer identify areas of low economic well-being.
Parker remarked he usually tends to be narrowly focused on the current codes. This workshop has given him many new ideas, but questioned how the small group of people in ERS can undertake the work.
Brown observed ERS did not receive questions about rural classification in the 1970s. To some extent, it is a response to the early efforts to develop a framework to look at geographic variability and many social measures.
Partridge stated that he has approached the possibility of producing rural-urban influence codes and labor market areas in a practical sense. His advice would be to first think of the main characteristics that should be captured with these measures, while culling others of lower priority. He noted that to the extent that economic features are a priority, one needs to have measures that are behavioral in their design such as commuting and the use of functional economic areas. He said ERS will need to get advice on the importance of compatibility over time. They might come up with a better measure, but if it is incomparable with the current measures, it may just cause confusion and a long transition period.
Plane observed the late Ken Bolding once said “knowledge is always gained by the orderly loss of information.” He noted that the workshop did not include a discussion of real estate trends. He stated that he disagreed about the idea of defining rural and then taking urban as the residual because urban areas are undergoing a regime of real change. There has been a trend in migration back to downtown areas, and he commented on the transformation of the suburbs in the decades ahead.
Plane suggested that analysts might think more about the dichotomy
between urban and rural. In fact, the two were getting closer together in the last few decades. However, he suggested that as the United States becomes more like Britain or other places with a more village kind of structure, the countryside and the city may become more distinct.
Waldorf suggested consideration of a context-specific definition of classification. She said it is important for a classification system not to mix up the concept itself with the causes and consequences. This may be related to the whole issue of validation, she said.
She referred to a point made by O’Brien about rural places having low capacity. Policies directed toward rural places have to look at their effectiveness at creating jobs, for example. Looking at classifying rurality through that lens means that causes, concept, and consequences are being mixed up. She suggested that for these kinds of policy questions, it might be better to identify an index of neediness as opposed to an index of rurality.
Waldorf also observed that analysts all agree that rurality is a multidimensional concept. During this workshop there has been talk about size, density, proximity, internal connectedness, and other factors. Waldorf concluded that she likes the concept of rurality and would like to have a definition of rurality and a classification. But she said she is at a loss to suggest any that would satisfy everybody.
Brown stated that in planning the workshop, he was very interested in making sure that the agenda included presenters from the United Kingdom, which he said might help to provide engagement among qualitative and quantitative researchers. He asked how qualitative information about the nature of rural community, community, economy, and society can be merged with quantitative information to develop conceptions of rural structure and rural change.
Shucksmith spoke about co-production of knowledge. He asked if qualitative information simply for evaluating the accuracy of quantitative analysis or does it have intrinsic information. Does it yield intrinsically important information that needs to be considered in a more thoughtful development of knowledge about rural and urban?
COMMENTS AND OBSERVATIONS
Constance Citro (Committee on National Statistics) observed that ERS and the U.S. Department of Agriculture in general need ways to categorize rural areas and the people living in them to carry out their mission. A good classification scheme helps bring order to that mission. She observed classifications get taken up for various programmatic purposes and tend to develop a life of their own, which can make it hard to implement revisions. According to Citro, the basic concept seems right that something
about rurality remains applicable, certainly to a country like the United States with its diverse geography and population. One could imagine some countries becoming totally urban and suburban for which a rurality concept would have little to offer. In the United States, in which some people live differently enough from urban and suburban living arrangements, it makes sense to have a way to characterize those areas.
Citro said if it were really true that there were no differences in the consequences of living in a rural area as opposed to an urban or suburban area for such important socioeconomic dimensions as poverty or inequality, then one might ask why have a rural classification. But, she said, prima facie, rural areas differ from other areas. Consequently, some kind of a rurality classification will continue to make sense. When cross-classified by other social indicators, such as single-parent family poverty, a rurality classification will help illuminate to what extent rural areas are becoming more like or unlike urban and suburban areas, Citro said.
Similar to poverty measurement, she said it could be useful to have several measures of rurality. In the case of poverty, one can use an income measure, such as the official U.S. measure, ask people what they think a poverty level is, or look at deprivation measures. Each of these measures enriches the story in different ways that can be useful for research and policy issues. Similarly, she said, developing several concepts of rurality, including how people themselves feel about it, could be very useful.
Responding to McGranahan’s statement about rural poverty, Partridge said that the Urban Influence Codes were designed in 1970s and they fit the world well then. But the world has changed, and now other indicators may be more informative.
Michael Woods made three observations about statements he heard. He referred to a point by Cromartie that these classifications are produced for statistical purposes only. The second statement was that ERS does not produce classifications for political purposes. The third was Tom Johnson’s remark about the political economy of definitions, and that each definition makes people money and costs money. Woods suggested that one challenge is to square those three statements.
He said using a rural definition, classifying areas as rural, and using those categories for research or for funding programs is making a political decision. They should not be political decisions, he said, because the evidence heard during this workshop is that the term “rural” is meaningless. It is about accessibility, proximity, density, and so on. To bundle those things and call the result rural is a political decision, Woods said.
He said ground-truthing is where qualitative research comes in. One way of asking whether the measures that are produced make sense statistically and quantitatively is to see if people qualitatively feel an area is rural or urban. Woods also suggested a distinction between classifying
populations and classifying territories. He noted that both have a special geographic expression, but there are examples of urban populations potentially living on rural land, and perhaps vice versa. Current methods have identified areas with mixed land use and mixed populations. Woods noted that these are tricky to work with, both for analysis and policy. Qualitative research is very important to understanding what is happening in most mixed areas and whether it makes more sense to align them with urban or with rural.
Woods suggested that if the outcome of the workshop is how rural areas are viewed for the next 40 years and beyond, then the way forward needs to be future-proof. To describe future-proof, Woods referred to the presentations on GIS, sensing, and new technologies. He said those techniques are the future of how to define a measure. There are all kinds of data now available that may go beyond the limited public data previously available. What would happen in 40 years’ time, he asked, if there were no decennial U.S. census? Woods said future-proofing is also about the process of change. What does it mean to be rural, not in a metropolitan society, but in a global society where connectedness means different things; where people may be commuting on a temporary fly-in fly-out basis to work, not necessarily to the largest nearest city; or where food-processing plants in rural areas are effectively recruiting from a continental labor market, not a regional labor market? Woods suggested pursuing a mixed-methods approach for combining quantitative and qualitative evidence to work through the likely dynamics of rural America over the next 30 or 40 years, the time frame for which a useful classification system would be robust.
In terms of ground-truthing, Pender suggested a simple survey with one question: Do you live in a rural area? Responses could be georeferenced and compared with what different classification systems indicate. Qualitative research with focus groups could find why some people may classify a given area as rural and others may classify it as urban.
Bruce Weber pointed to earlier statements about context and whether to define rural in terms of density and proximity. He said his reading of the regional science literature has convinced him that proximity and density are important in terms of development patterns and outcomes. It has also convinced him, he said, that place in the hierarchy is important, and when suburbs of the major metropolitan areas were taken out of the rural urban continuum codes, something was lost in terms of what the Brookings Institution calls the largest segment of the population.
Daniel Lichter observed his frustration about the lack of workshop discussion about the link between changes in race and ethnicity and the U.S. system of resettlement, referring to Census Bureau projections that the United States will be a majority/minority society by 2043 if current
trends continue. The concept of segregation, for example, has mostly been concentrated in urban metropolitan neighborhoods, but the same issues exist in rural America, such as in the Delta, Indian reservations, or Lower Rio Grande Valley. What are the economic transitions from those areas? Lichter stated race is fundamental to the U.S. settlement system beyond big city populations and metropolitan areas. He and his colleagues worked to characterize every place in the United States from big to small. The only place that white people are moving to are areas or places that have an established white population, he said. The suburbs are becoming minority and immigrant. Whites are either moving further out or back into the cities. He said that he knows race is not a concept typically considered in an analysis of rurality but it seems fundamental.
In his final remarks, McGranahan stated that his original thought was to get five or six people together in a room to talk about how ERS might change the codes. He said the idea grew into the two days of discussion at this workshop. He said ERS learned about different variables and contexts for rurality and saw examples of other rural classification systems in a very wide-ranging discussion. He noted ERS needs to look broadly to assess the current definitions before it can decide on next steps. Although the discussions at the workshop did not provide specific ideas for developing a new or revised coding system, he commented, they did provide thoughtful insights and the workshop has been quite valuable. He said he views developing a code as a demographic effort to determine the extent to which place, size of population, and distance constrain and permit economic activity, access to services, resilience to problems, and so on. He thanked all the participants and presenters for their contributions.
This page intentionally left blank.
