Fiscally standardized cities

EDA
clustering
Extensive financial data on over 200 of the largest cities in the United States for over 40 years. Which cities spend the most or the least on government services?
Author

Alex Reinhart

Published

September 7, 2023

Data files
Data year

2020

Motivation

In the United States, city governments provide many services: they run public school districts, administer certain welfare and health programs, build roads and manage airports, provide police and fire protection, inspect buildings, and often run water and utility systems. Cities also get revenues through certain local taxes, various fees and permit costs, sale of property, and through the fees they charge for the utilities they run.

It would be interesting to compare all these expenses and revenues across cities and over time, but also quite difficult. Cities share many of these service responsibilities with other government agencies: in one particular city, some roads may be maintained by the state government, some law enforcement provided by the county sheriff, some schools run by independent school districts with their own tax revenue, and some utilities run by special independent utility districts. These governmental structures vary greatly by state and by individual city. It would be hard to make a fair comparison without taking into account all these differences.

This dataset takes into account all those differences. The Lincoln Institute of Land Policy produces what they call “Fiscally Standardized Cities” (FiSCs), aggregating all services provided to city residents regardless of how they may be divided up by different government agencies and jurisdictions. Using this, we can study city expenses and revenues, and how the proportions of different costs vary over time.

Data

The dataset tracks over 200 American cities between 1977 and 2020. Each row represents one city for one year. Revenue and expenditures are broken down into more than 120 categories.

Values are available for FiSCs and also for the entities that make it up: the city, the county, independent school districts, and any special districts, such as utility districts. There are hence five versions of each variable, with suffixes indicating the entity. For example, taxes gives the FiSC’s tax revenue, while taxes_city, taxes_cnty, taxes_schl, and taxes_spec break it down for the city, county, school districts, and special districts.

The values are organized hierarchically. For example, taxes is the sum of tax_property (property taxes), tax_sales_general (sales taxes), tax_income (income tax), and tax_other (other taxes). And tax_income is itself the sum of tax_income_indiv (individual income tax) and tax_income_corp (corporate income tax) subcategories.

Data preview

fisc_full_dataset_2020_update.csv.gz

Variable descriptions

For each city and year, the following metadata is available:

Variable Description
year Year for these values
city_name Name of the city, such as “AK: Anchorage”, where “AK” is the standard two-letter abbreviation for Alaska
city_population Estimated city population, based on Census data
county_name Name of the county the city is in
county_population Estimated county population, based on Census data
cpi Consumer Price Index for this year, scaled so that 2020 is 1.
relationship_city_school Type of school district. 1: City-wide independent school district that serves the entire city. 2: County-wide independent school district that serves the entire county. 3: One or more independent school districts whose boundaries extend beyond the city. 4: School district run by or dependent on the city. 5: School district run by or dependent on the county.
enrollment Estimated number of public school students living in the city.
districts_in_city Estimated number of school districts in the city.
consolidated_govt Whether the city has a consolidated city-county government (1 = yes, 0 = no). For example, Philadelphia’s city and county government are the same entity; they are not separate governments.
id2_city 12-digit city identifier, from the Annual Survey of State and Local Government Finances
id2_county 12-digit county identifier
city_types Two types: core and legacy. There are 150 core cities, “including the two largest cities in each state, plus all cities with populations of 150,000+ in 1980 and 200,000+ in 2010”. Legacy cities include “95 cities with population declines of at least 20 percent from their peak, poverty rates exceeding the national average, and a peak population of at least 50,000”. Some cities are both (denoted “core

The revenue and expenses variables are described in this detailed table. Further documentation is available on the FiSC Database website, linked in References below.

All monetary data is already adjusted for inflation, and is given in terms of 2020 US dollars per capita. The Consumer Price Index is provided for each year if you prefer to use numbers not adjusted for inflation, scaled so that 2020 is 1; simply divide each value by the CPI to get the value in that year’s nominal dollars. The total population is also provided if you want total values instead of per-capita values.

Questions

  1. Do some exploratory data analysis. Are there any outlying cities? Any interesting trends and relationships? Also, explore the hierarchy of revenues and expenses, and check that values add up in the way the hierarchy suggests they should.
  2. When considering expenditures, there may be different kinds of cities. Perhaps dense cities with efficient public transit spend money in different ways than large, sprawling cities where everyone drives, for example. Extract out important expenditure variables and do a clustering analysis. Are there distinct clusters? How many? Can you interpret what they mean? Be careful about including the hierarchical values in your analysis.

References

Lincoln Institute of Land Policy. Fiscally Standardized Cities database. https://www.lincolninst.edu/research-data/data-toolkits/fiscally-standardized-cities