300 likes | 437 Views
Place-based Policy and Identifying Spatially Heterogeneous Effects. by Mark Partridge, Ohio State University May 12, 2010 ___________ Prepared for presentation at the Identifying Candidates for Place-Based Policy Workshop Group for the Analysis of Development (GRADE) Lima, Peru
E N D
Place-based Policy and Identifying Spatially Heterogeneous Effects by Mark Partridge, Ohio State University May 12, 2010___________ Prepared for presentation at the Identifying Candidates for Place-Based Policy Workshop Group for the Analysis of Development (GRADE) Lima, Peru Available at: http://aede.osu.edu/programs/Swank/
Motivation • Do we help poor people or poor places? Answer is surely the first, but what about the 2nd? • People-based policies: training, education, relocation assistance, healthcare • Place-based policies are infrastructure, economic development, tax breaks for businesses. • Of course there is cross-fertilization of people and place-based policies. • Place based policies are criticized for slowing necessary adjustments and prolonging the agony. • Any new jobs are taken by new residents and commuters.
Motivation • Skepticism of place-based policies have grown: • New Economic Geography: place-based policies that focus on poor regions inhibit more-productive “cities” from growing. This reduces aggregate economic growth. • World Development Report 2009. • “For these reasons, the standard argument among economists is that people-based policies of supporting worker training and facilitating household mobility are far superior to possibly wasteful place-based policies. For example, Harvard economist Ed Glaeser pointedly made this argument in the wake of Hurricane Katrina. He contends that it would be far superior to award each resident of the city of New Orleans $200,000 than to provide federal support to rebuild it (Pettus, 2006). Glaeser contends that its poor residents would be better off relocating to more vibrant locations.” Partridge and Rickman, 2008 Cambridge Journal Regions Economy and Society.
Motivation: • Place-based policy makes sense when there are barriers to mobility. People can’t move to where the jobs are. Thus, jobs need to be in “place” in combination with people-based supports. • Of course, even people-based policies need to be tailored to each region—e.g., delivering education policies to rural areas is different than in the inner city.
Motivation • Rural Poverty is often ignored even though it is about as severe as in metropolitan America. • Most probably due to its dispersed nature in communities. • Yet, rural poverty is a problem: • “Some of the same signs of despair and breakdown that wore out aging American industrial cities in the 1960's have come to the rural plains. Among teenagers, there is now a higher level of illicit drug use in rural areas than in cities or suburbs, recent surveys indicate. The middle class is dwindling, leaving pockets of hard poverty amid large agribusinesses supported by taxpayers.”Timothy Egan, “Vanishing Point; Amid Dying Towns of Rural Plains, One Makes a Stand,” New York Times (December 1, 2003, Late Edition), p. A1. • “Fundamental structural changes in technology, markets, and organizations are redrawing our nation’s economic map and leaving many rural areas behind.”Robert D. Atkinson of the Progressive Policy Institute. Swank Program in Rural-Urban Policy
The major view in regional science and econ geog. is that rural areas are disadvantaged due to small scale and geographical remoteness: • Lack Agglomeration economies • Low wages, weak job growth • Remoteness from cities exacerbates access to agglomeration. • Examples: • “Oakridge, Oregon was a prosperous timber community of about 4,000 people until its last mill closed in 1990. Many households now struggle in or just above poverty, though they seem determined to remain in their scenic community. Flourishing Eugene could provide employment opportunities, but being 55 miles away limits the ability of Oakridge’s residents to take advantage.” (Eckholm) • “Among Appalachia's problems are that it is “too far from big cities to easily attract businesses.” (Altman)
Overview of National Poverty • U.S. poverty-employment growth link was re-established in the 1990s. • Strong link in 1960s and early 1970s • Weak between 1973-1993 (high poverty in 1993) • The link mostly reestablished itself after 1993. However, the benefits of economic growth are only weakly trickling down.
1959-2003 U.S. Family and Person Poverty Rates 25 20 Poverty Rate 15 10 5 0 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 Year 1959-2003 Family Poverty Rate 1959-2003 Overall Poverty Rate Source: Partridge, M.D. and D.S. Rickman. The Geography of American Poverty: Is there a Role for Place-Based Policies?, Kalamazoo, MI: W.E. Upjohn Institute for Employment Research, 2006
Rural Poverty Overview • Nonmetropolitan America has higher poverty rate than in metropolitan America • [“Rural” and “nonmetropolitan” in my discussion] • We will use the official U.S. poverty measure. • Problems with official poverty rate • What is poverty? • How to measure cost of living? • In Rural America, there is lower housing costs vs higher transport and other costs (no Wal-Mart) • Some public/private services simply don’t exist. • But high poverty rate captures notion of economic degradation.
Average Pop. Weighted County Poverty Rates MSA and Non-MSA: 1989 and 1999 20 Tot al MSA Non- MSA 15 1989 10 1999 1989-1999 % 5 change 0 Total MSA pop MSA pop Central Suburb Total Adjacent Not Total MSA >1 Million <1 Million MSA city MSA city to MSA Adjacent Non-MSA -5 to MSA -10 Source: Partridge, M.D. and D.S. Rickman. The Geography of American Poverty: Is there a Role for Place-Based Policies?, Kalamazoo, MI: W.E. Upjohn Institute for Employment Research, 2006
Source: Partridge, M.D. and D.S. Rickman. The Geography of American Poverty: Is there a Role for Place-Based Policies?, Kalamazoo, MI: W.E. Upjohn Institute for Employment Research, 2006
Spatial Diversity of Rural Poverty • Rural poverty is somewhat clustered. • Low poverty in the Midwest and Northeast. • High poverty in the South and West Coast. • Poverty rates are spatially persistent. • Can be large intra-regional variation in levels. • Large inter-regional variation.
Overview of Past Rural Poverty Research • Many of the same structural problems as for urban poverty • A rural spatial mismatch of jobs • Jobs are often in the city, but not in the country • Thin labor markets weakens rural employment matches • Lacks transportation, childcare, work supports • Structural change out of primary sector & manufacturing • Debate on whether economic opportunities help poor rural families to the same degree as in urban areas (Davis et al., 2003 vs. Partridge and Rickman, 2005, 2006; 2007, 2008a, 2008b). • Social isolation, peer effects, insufficient community capacity or social capital also exist • (Glasmeier and Farrigan, 2003) • Race and ethnic differences: reservations, cotton belt (Leichenko, 2003).
Persistent Poverty Counties/clusters • What about persistent poverty counties/clusters? • Something seems deeply wrong. • But, is this spatial clustering individual self selection to live in poor areas for (say) low-cost housing or access to low-skilled occupations [not a problem]? • Partridge and Rickman (2005) argue that these places are not hopeless poverty traps. • Job growth has even stronger poverty reduction impacts than in other rural counties (less in-migration and commuting). • Education has similar impacts • Single parent shares have stronger adverse poverty impacts in PP counties, while Assoc degree share has stronger favorable impacts • One simple solution is to spur economic growth in persistent poverty clusters. • Of course, how does one go about doing this?
Describe Heterogeneous Effects and Why Place-based policy makes sense even in the U.S. • Summarize Partridge and Rickman’s (2008) in the Cambridge Journal. • More remote areas have higher poverty rates (see figure below and Partridge and Rickman 2008 Journal of Regional Science). • The point is to examine whether the labor supply curve in more remote areas is more inelastic in the long-run. This would suggest place-based policy may work because of labor immobility. • Findings: • Local job growth reduces poverty. • Local job growth has stronger effects in remote counties. • If high poverty is just sorting of the lowest-ability people, then this would not happen. • This suggests that disadvantaged low-income residents are the recipients of the new jobs.
Heterogeneous Effects—Partridge & Rickman 2008 CJRES • More remote areas have a smaller population response to local job growth (relative to metropolitan areas or more proximate rural areas). Thus, fewer migrants take the new jobs. • The response of the long-term resident population share (more than 5 yrs) to job growth was statistically insignificant in remote areas but negative in metropolitan areas—i.e., job growth attracts more migrants in metropolitan areas. • These results support the notion that policies that support job growth in remote areas will reduce poverty.
Figure 2: 1999 Nonmetropolitan Poverty Rates By Distance to Metropolitan Areas Greater than 1.5 Million 1990 Population • Source: Partridge, Mark D. and Dan S. Rickman. “Place-Based Policy and Rural Poverty: Insights from the Urban Spatial Mismatch Literature.” Cambridge Journal of Regions, Economy and Society, 1 (2008, 1): 131-156.
What about Persistent Poverty Clusters? • Source: Partridge, Mark D. and Dan S. Rickman.“Persistent Pockets of Extreme American Poverty and Job Growth: Is there a Place Based Policy Role?” Journal of Agricultural and Resource Economics, 32 (April 2007): 201-224. • Persistent High Poverty Clusters • (USDA, 1959-1989 or 1969-99, 20%+ in every year) • Central Appalachia, Historic Southern Cotton Belt, Rio Grande Valley and Western Reservations. • Each poverty rate cluster seems a little different demographically and in each geographical region.
Figure 1: 1999 Persistent Poverty Counties, 1979-1999 Definition
What about Poverty Clusters—selected results Employment has a bigger effect in Persistent Poverty Counties
Geographical Weighted Regression • We conducted GWR regression models across the poverty clusters.The findings include: • Employment growth’s negative association on poverty does not vary across the clusters. • This suggests job growth’s favorable impact is strong across the sample and self-sorting effects are minimized. • Overall, there is statistically significant spatial variation across at least one variable (not withstanding the job growth results). Among the statistically significant results, I will show you some of the interesting results.
Figure 2: GWR Variation in the Average Surrounding County Regression Coefficient.
Figure 3: GWR Variation in the Female-Headed Family with Children Regression Coefficient.
Figure 4: GWR Variation in the Share of 18-24 Years of Age Regression Coeff.
Conclusions • Recent attacks on Place-based policy suggest that it needs a rigorous evaluation. • A key assumption of those who argue that people-based policies should be exclusively employed is perfect mobility of labor. • These results show that even in the highly mobile U.S., labor supply is more inelastic in the long-term in remote areas and in persistent-poverty clusters. • Thus, place-based policy is potentially effective. Swank Program in Rural-Urban Policy
Conclusions—cont. • Job growth has stronger poverty-reducing effects in remote and persistent poverty areas—consistent with self-sorting NOT driving the geography of poverty. • Spatially heterogeneous effects suggesting the need for place-tailored policies.
Thank you! Papers are posted at The Ohio State University, AED Economics, Swank Program website: http://aede.osu.edu/programs/Swank/ (under presentations) Swank Program in Rural-Urban Policy