PA Regions Rank High in Study of Upward Mobility
Overview of Study
Check out The New York Times' interactive map detailing regional findings from the study.
A new study by a team of top national economists finds that Pennsylvania has enjoyed substantially more upward mobility than many other parts of the United States. Pittsburgh ranks near the front of the pack within the group of the 30 largest regions, and many less populous parts of the commonwealth fare as well or better.
The study by two Harvard and two University of California-Berkeley economists — Raj Chetty, Nathaniel Hendren, Patrick Kline, and Emmanuel Saez — breaks new ground by providing the most detailed information yet on upward mobility in different regions across the county. The study relies on a sample of 6.3 million children born in the United States in 1980 and 1981, divided into 741 regions, or “commuting zones.” Using this enormous database, the study explores how much these people’s income at age 30 depends upon their parents’ income when they were children, aged 15 to 20 years old, along with other variables thought likely to impact intergenerational mobility.
The researchers undertook the study to explore how the level and progressivity of taxes impacted mobility — and found that several tax variables correlated positively with intergenerational mobility, including the level and progressivity of tax expenditures and state earned income tax credit policies. (Pennsylvania does not have a state earned income credit.)
While cautioning that more research is needed on the factors that expand opportunity, the economists identified four variables that correlate with upward mobility: the size and dispersion of the middle class; the quality of K-12 schools; strong families, measured by the share of two-parent households; and civic engagement, including membership in religious and civic groups.
Pennsylvania civic leaders and policymakers should examine the findings of this rich new study closely to guide them in shaping regional and state policy and action going forward.
The Upward Mobility of Pennsylvania’s Regions
Table 1 below profiles the 12 Pennsylvania regions defined by the Harvard and U.C.-Berkeley economists based on three measures of upward mobility and their population. The table also shows mobility and population in eight comparison regions. Appendix Table A1 provides additional information on the characteristics and upward mobility of Pennsylvania’s regions and the eight comparison regions.
- Pittsburgh ranked fifth highest among the 30 biggest metro areas for the share of children born into low-income families who grow up to earn top incomes as adults. Pittsburgh kids born into families in the lowest-income fifth (with parental income below $25,000 a year) had a 10% chance of making it into the top fifth. Philadelphia ranked 19th of 30 on this measure, with an 8% chance.
- Pittsburgh did even better in terms of opportunity at the top end of the scale, ranking first measured by the probability of children from families in the top fifth (incomes over $107,000 a year) ending up in the top fifth by age 30 — 38%. Philadelphia ranked fourth out of 30 on this measure, with 36% of children from top-fifth families making it into the top fifth as adults. Thus, children in the top fifth have about four times as much chance of making it into the top fifth in Pittsburgh as do children in the bottom fifth, while this same ratio is nearly five in Philadelphia, seven in Cincinnati, and seven-and-half in Atlanta.
- In all 12 Pennsylvania regions, children whose parents’ income fell in the bottom fifth had a 7.7% to 12.9% chance of reaching the top fifth. (See Table 1.) These percentages are two to three times higher in Pennsylvania than in the regions with the least upward mobility, such as Atlanta, Charlotte, and a multi-state southern zone spanning most of five states.
- Pennsylvania’s regions also rank high on a third measure of upward mobility: the expected income percentile as adults for all children whose parents had incomes below the 50th percentile, or median. As the last column of Table 1 shows, this percentile equaled or exceeded the 44th percentile for all Pennsylvania regions except Philadelphia. Four of these 11 Pennsylvania regions outranked all of the non-Pennsylvania comparison regions at the bottom of Table 1, and two others were essentially tied with the highest-ranked comparison region (Salt Lake City).
- St. Marys stands out among all the Pennsylvania regions on all three of our upward mobility measures. The chance of a St. Marys child in the bottom fifth rising to the top fifth is 12.9%. The average income percentile of St. Marys children who grow up at the 10th percentile is 44. And St. Marys children whose parents’ income were in the bottom half of the distribution have an average income percentile as adults of 48.4 — very nearly the U.S. median. Appendix Table A1 also shows that St. Marys has a stunningly large middle class — 70% of children in St. Marys grew up in households with incomes between the U.S. 25th percentile and the U.S. 75th percentile. The upward mobility indicators for the Altoona and Williamsport commuting regions are only slightly below those of St. Marys.
- In the 12 Pennsylvania regions, the average income of a child who grew up at the 10th percentile rose to somewhere in the 36th and 44th percentile range. (Table 1 shows the percentile for each region.) This compares with the 30th and 31st percentile in Atlanta and Charlotte.
- Five Pennsylvania regions in the study — St. Marys, Altoona, State College, Williamsport, and Scranton — did even better in terms of mobility than Pittsburgh (measured by mobility from the bottom fifth to top fifth across generations and by the average income of 30-year olds whose parents’ income fell at the 10th percentile).
- Virtually all parts of three southern states, North Carolina, South Carolina, and Georgia, and most parts of two other southern states, Alabama and Mississippi, have low mobility.
According to a New York Times report on the study, “the parts of this country with the highest mobility rates — like Pittsburgh, Seattle and Salt Lake City — have rates roughly as high as those in Denmark and Norway, two countries at the top of the international mobility rankings.” By that standard, most of Pennsylvania has upward mobility comparable to Denmark and Norway — while Atlanta, Memphis, and most of the five southern states listed above have upward mobility below that of any other rich country.
Explaining Upward Mobility: What Works and What Doesn’t
While more research is needed, this study provides powerful new evidence of what works to sustain the American Dream of widespread mobility. As noted earlier, the researchers identified four broad variables that correlate with upward mobility: the size and dispersion of the middle class; the quality of K-12 schools; strong families, measured by the share of two-parent households; and civic engagement.
These four variables help make sense of Pennsylvania’s high ranking in the 1996 to 2011 period of this study. In that period, Pennsylvania’s newly diversified economy — less dependent on the steel industry — and its middle class stabilized after the harrowing job losses of the 1980s. The Ridge and Rendell Administrations increased state education funding, the first targeting rural regions and the second also targeting lower-income urban school districts, with national evaluations showing improved K-12 educational outcomes by 2010. Pennsylvania has a low divorce rate, strong philanthropic organizations in many regions, and a relatively strong labor movement. Civic engagement measured by voter turnout is only in the middle of the pack, in part because Pennsylvania does not have laws such as same-day registration and online voting that increase voter participation
Within Pennsylvania, the particular mechanisms that maintain upward mobility likely differ between the largest metropolitan areas, the most rural regions, and smaller cities. Particularly in rural areas, a strong and dispersed middle class and uniformly high-quality schools help explain mobility. The strength of the middle class reflects, in part, the continued strength of manufacturing in Pennsylvania’s regions, notably in the St. Marys region. As noted earlier, a remarkable 70% of children in St. Marys grew up in households with incomes between the U.S. 25th percentile and the U.S. 75th percentile. The dispersion of the middle class likely figures into the high upward mobility of rural Pennsylvania because low-income children remain part of the broader community. They are not isolated in concentrated pockets of poverty and they attend the same (public) schools as the children of middle-class and upper-income families.
Quality education and a broad middle class remain important factors in many Pennsylvania metropolitan areas, including suburban areas and some cities (for example, Pittsburgh has long been seen as a city with high-quality schools). Labor unions may also figure into the story, impacting both the size of the middle class and civic engagement. Eight of 11 Pennsylvania regions for which metro-area union density figures exist for 2012 rank in the top 87 out of 264 metro areas for which such data are available. This includes Erie at 52, Scranton at 67, Pittsburgh at 72, Harrisburg at 77, and Philadelphia at 80. One barrier to upward mobility in many Pennsylvania metro areas is the concentration of poverty and isolation of some low-income city neighborhoods and schools from the broader middle class.
The four broad variables highlighted by the Harvard and Berkeley economists also shed light on the low mobility in the five southern states. These states fund education at below-average levels, three of them near the bottom of the 50 states. These states have enacted so-called “right-to-work” laws that discourage civic engagement by working people. Georgia and the Carolinas rank 47th to 49th by union density. Interestingly, The New York Times online map of upward mobility shows that the areas with the greatest upward mobility in the United States closely mirrors the states that do not have right-to-work laws, which also tend to be areas in which higher shares of workers belong to unions and organized labor engages most actively with the legislative and political process.
Implications for Pennsylvania Regions: Redouble Your Efforts
Pennsylvania regional and state leaders should take note of the findings in this study as they look to the future. Rather than rest on their laurels, Pennsylvania regions need to deepen civic dialogue about how to broaden opportunity going forward. This is especially important as some of the factors that have sustained upward mobility in the past may not do so as effectively in the future. For example, the manufacturing sector has shrunk substantially since 1996 and even since 2007. Relative wages of the manufacturing sector have fallen, so that even where manufacturing jobs remain, fewer of them lift non-college-educated workers into the middle class. It may also be true that Pennsylvania’s upward mobility since 1996 partly reflects the residue of 1970s private-sector labor union strength, in both the labor market and in the political and policymaking process. Recent reductions in funding for public education, particularly if they persist, also ring alarm bells when it comes to sustaining mobility in Pennsylvania going forward.
The good news is that Pennsylvanians regions from Erie to Reading, Pittsburgh to the Lehigh Valley, Sunbury to Philadelphia are already engaged in civic dialogue about expanding opportunity. Over the next 15 years, given the polarizing pressures of the global economy, Pennsylvania’s regions need to focus like a laser on how to maintain a robust middle class, good schools for all, strong families, and civic engagement, including a place at the policy table for working families.
Implications for the State: We’re on the Wrong Track
What does the current trajectory of state policy mean for Pennsylvania’s relatively high level of upward mobility? Recent years have seen deep cuts in state education funding targeted especially at low-income urban school districts and efforts to suppress civic engagement through passage of a voter ID law now being contested in the courts. Pennsylvania’s Legislature has also considered, although not yet passed, other measures that would increase civic disengagement by divorcing political representation from popular support. These measures include a proposal to base Pennsylvania’s Electoral College votes on popular votes at the Congressional District level, using districts that have been gerrymandered in a highly partisan way. Several proposals to enact a so-called right-to-work law have also been introduced in the Pennsylvania House.
The hard data from the new Harvard-Berkeley study suggest that a continuation of state policies that undermine the quality of education, decrease civic engagement, and shrink the middle class would substantially reduce upward mobility. If Pennsylvanians prize upward mobility, the state needs a new policy direction.
 Raj Chetty et al., The Economic Impacts of Tax Expenditures: Evidence from Spatial Variables Across the U.S., revised July 2013; online at http://scholar.harvard.edu/hendren/publications/economic-impacts-tax-expenditures-evidence-spatial-variation-across-us.
 Parental income is defined as the average of yearly income reported on federal tax forms from 1996-2000. Child income is defined as the average of reported 2010 and 2011 income. All income was adjusted to 2010 dollars.
 Complete and “preferred” mobility measures for each “Commuting Zone” are online at http://www.equality-of-opportunity.org/index.php/data. The New York Times maps online at http://www.nytimes.com/2013/07/22/business/in-climbing-income-ladder-location-matters.html?_r=0&hp=&adxnnl=1&adxnnlx=1374508953-DRVCxk+C+rxmibypMe5+Lw also provide some regional data on two measures of mobility.
 Reporters or members of the public who would like assistance analyzing all of the study’s data for their part of Pennsylvania or constructing a customized comparison group of non-Pennsylvania “Commuting Zones” should contact Stephen Herzenberg at 717-255-7145 or firstname.lastname@example.org. KRC can also help identify local leaders who are knowledgeable about upward mobility or participate in civic efforts to enhance mobility.
 Based on 2009 American Community Survey data, Pennsylvania ties for the fifth lowest divorce rate among men and fourth lowest among women: http://www.divorcestatistics.info/divorce-statistics-and-divorce-rate-in-the-usa.html.
 Pennsylvania has the 16th-highest unionization rate in 2012 (data from the Current Population Survey online at http://www.unionstats.com/).
 Pennsylvania voter turnout was 31st in 2012 (http://elections.gmu.edu/Turnout_2010G.html) and 21st in 2001 (http://bipartisanpolicy.org/sites/default/files/2012%20Regional%20Turnout.pdf).
 As late as 2005, manufacturing employment in Elk County, home of St. Marys, represented 35% of total employment. While down from 50% in 1979, this 35% share is similar to the levels in Pennsylvania as a whole in the 1950s. See Stephen Herzenberg and Mark Price, The State of Rural Pennsylvania, Keystone Research Center, 2007, p. 41, online at http://keystoneresearch.org/publications/research/state-rural-pennsylvania-economic-agenda.
 Interestingly, two small metro areas — in the Altoona “Commuting Zone” — rank highest for union density, Altoona itself ranks 28th and Johnstown ranks 42nd. These and the rankings in the text computed by KRC from data extracted from the March Current Population Survey by Barry T. Hirsch and David A. Macpherson, online at http://www.unionstats.com.
 Based on 2009-10 data, South Carolina and Georgia rank 32nd and 33rd for state funding per pupil, Alabama 43rd, Mississippi 45th, and North Carolina 46th. Pennsylvania ranks 10th. For these rankings, see National Education Association, Rankings and Estimates: Ranking of the States 2011 and Estimates of School Statistics 2012, online at http://www.nea.org/assets/docs/NEA_Rankings_And_Estimates_FINAL_20120209.pdf.
 See the online map at http://www.nrtw.org/rtws.htm.
 Based on voter turnout, these states ranked 24th, 34th, 36th, 37th, and 44th in 2010 but clustered just below the middle in 2010.