• haseeb
Categorized As:
Regional Direction

Building a Multi-Dimensional Index to Benchmark Philadelphia

This Leading Indicator introduces a research endeavor at the Economy League of Greater Philadelphia to analyze peer cities of Philadelphia across several socioeconomic indicators for improved comparative analysis. The interplay of different regional socioeconomic characteristics creates a complex picture of how different metros and cities compare with our own; we intend to form a similarity index measure for comprehensive benchmarking. This data brief highlights some of the key measures that may be included in our final index.


What You Need to Know

  • The Economy League of Greater Philadelphia is developing a multi-dimensional index for benchmarking regions for comparative analysis.

  • The index will identify regions at the county level that resemble Philadelphia across various socioeconomic indicators.

  • We will start with three key measures: crime rate per 100,000 residents, ethnic heterogeneity, and real GDP per capita.

  • As of 2016, Philadelphia had a crime rate of 1,150 crimes per 100,000 people; Philadelphia’s crime rate is 8th highest among the 2,927 counties in that year.

  • Philadelphia is the 18th most ethnically heterogenous county in our analysis.

  • With a real GDP per capita of $64,331, Philadelphia ranks 368 out of 2,927 U.S. counties.


Philadelphia’s Peers on Crime Rate, Ethnic Heterogeneity and GDP per capita

Comparing cities and municipalities across the U.S. is quite complex, considering historic contexts, socioeconomic issues, and vastly different population sizes. A multidimensional benchmarking index allows for a better comparison since it smooths over these differences and produces a proportional score for different areas [1]. For example, while the crime rate is an important indicator of social well-being in a region, it is challenging to compare crime rates across different municipalities because of the variety of crimes and areas of higher population reporting statistically higher types of crimes committed. A multidimensional index can account for these differences by both normalizing and equalizing socioeconomic indicators leading to fairer regional comparisons.  This can potentially aid policymakers and practitioners to identify new solutions implemented in geographic areas that resemble Philadelphia across a comprehensive list of socioeconomic indicators and are facing similar policy issues. (We also conduct this analysis at the county-level both to match available data estimates as well as expand area comparisons beyond the simple city- or metro-levels which can exclude a variety of potential peers.) We begin this analysis by looking at Philadelphia County’s closest peers on crime rate, ethnic heterogeneity, and real GDP per capita. Our goal is to highlight novel insights on benchmarking that can be gained via the complex interplay of just three dimensions. We hope to add more dimensions to this index to keep tabs on major socioeconomic and policy trends and which geographic areas are experiencing similar trends to those in Philadelphia.


Figure 1 shows the overall ranking of Philadelphia County and its peers on crime rate per 100,000 residents, ethnic heterogeneity, and real GDP per capita. The bar graph also shows counties that most closely resemble Philadelphia across these metrics. We created a comprehensive list of metrics across approximately 3,000 U.S. counties using a variety of publicly available datasets, including the American Community Survey (ACS), Bureau of Economic Analysis (BEA), and the Inter-university Consortium for Political and Social Research (ICPSR). The x-axis shows a percentage difference of each metric compared to Philadelphia. For example, Shelby County, TN (top-left) has a crime rate that is 14.3% higher than the crime rate in Philadelphia. We include the ten closest counties to Philadelphia on each metric.



SOURCE: American Community Survey, Bureau of Economic Analysis, ICPSR


According to the most recent 2016 ICPSR data on crime statistics, Philadelphia has a crime rate of 1,150 crimes per 100,000 people; Philadelphia’s crime rate is 8th highest among the 2,927 counties in our dataset. Pershing County, NV and Scott County, MO are the closest counties to Philadelphia in terms of crime rates. These counties do not have a major city and have a far lower population than Philadelphia (less than 50,000 residents) but lie much closer to Philadelphia on crime statistics. On the other hand, Marion County, IN (the core county containing the City of Indianapolis) has a relatively similar population and crime rate to Philadelphia at 1,084 crimes per 100,000 people. This variation shows that a county-level analysis can identify proximity on crime statistics for a wider range of areas than a similar comparison conducted at the metro- or city-level.


Philadelphia also ranks 18th in terms of ethnic heterogeneity (using the American Community Survey’s 2019 5-Year Estimates). We calculate ethnic heterogeneity in each county via the Simpson’s Index which takes into account both the racial and ethnic richness as well as evenness of these populations in each region [2]. Simpson’s Index ranges from 0 to 1 where a “0” indicates complete homogeneity—or that everyone is of the same race or ethnicity—while a “1” indicates perfect heterogeneity or that all races and ethnicities are equally represented. Philadelphia has a Simpson’s Index score of 0.69 which puts it closer to areas like San Francisco County, CA (San Francisco) and Cook County, IL (Chicago) in terms of ethnic diversity. It is important to point out that our measure captures ethnic diversity in each county but does not map out the actual makeup of racial and ethnic makeup in each region. In other words, the ethnic makeup of similarly diverse counties in our measure could look very different in Philadelphia when compared to San Francisco, for example.


In terms of GDP per capita, Philadelphia ranks 368 out of 2,927 counties at $64,331 per capita (via the Bureau of Economic Analysis’ 2019 Estimates). Counties that are closest to Philadelphia in GDP per capita are far less populous than Philadelphia. In this comparison, Union County, NJ has the second highest population of 554,000 residents compared to 1.56 million in Philadelphia, yet its GDP per capita is only 0.24% less than that of Philadelphia. In fact, Union County NJ looks very similar to Philadelphia both in terms of GDP per capita as well as ethnic heterogeneity.


Multidimensional Relationship between Indicators

Figure 2 shows a three-dimensional scatterplot to highlight the complex interplay between the three factors of ethnic heterogeneity, crime rate, and GDP per capita. Each point in the graph shows a county that is closest to Philadelphia on each dimension (hover your curser over each point to see the county’s name). In other words, we plot the top ten counties that are most like Philadelphia in terms of their ethnic heterogeneity, crime rate, and GDP per capita. We also color each point by geographic region. Through this analysis, we find that Philadelphia lies closer to counties that are less populous and less diverse due to its GDP per capita and crime rate. In fact, Philadelphia resembles many counties in the Southern U.S. when we take all three dimensions into account. Similarly populous counties, including San Francisco, CA (San Francisco) and Cook County, IL (Chicago) as well as fellow Northeastern U.S. counties tend to cluster around each other with higher GDPs and measures of racial and ethnic diversity BUT lower crime rates. Our findings show that comparing counties across multiple dimensions can yield a more thorough understanding of how different places relate to each other which can lead to better benchmarking and comparative analysis. 



SOURCE: American Community Survey, Bureau of Economic Analysis, ICPSR


Looking Ahead

In this Leading Indicator, we examined three metrics that may be incorporated into a future similarity index we hope to create for better comparisons and benchmarking. We intend to add additional dimensions across gender, school enrollment, poverty, etc. to our analysis which will contribute towards a comprehensive multidimensional index for U.S. counties. This index will synthesize the various socioeconomic indicators to create a better picture of similarities and differences across regions in the U.S., which can further help identify shared solutions to ongoing policy issues.


Works Cited

[1] R. Shrestha, “What is Multidimensional Poverty Index (MPI)?,” Public Health Notes, Jan. 25, 2022. https://www.publichealthnotes.com/what-is-multidimensional-poverty-index-mpi/ (accessed Nov. 28, 2022).


[2] Zach, “Simpson’s Diversity Index: Definition & Examples,” Statology, Mar. 29, 2021. https://www.statology.org/simpsons-diversity-index/ (accessed Nov. 28, 2022).