If you followed my compilation of 2007 Census of Governments data, you know I try to come up with reasonable comparisons by adjusting for the varying level of population and income in different places (by measuring government revenues and expenditures as a percent of personal income), and the differing structure of local government in different places (by aggregating data at the county level). Even so, comparisons aren’t perfect because some places have more government services than others. There are some places that have professional fire departments, while others rely on volunteers. There are some places with free municipal solid waste collection, some with contracted out solid waste collection, and some where people have to hire and pay for their own collection services. Some places have public water, sewer and transit, and others do not.
What every place in the United States has, however, is police and public education. This post uses data from the Governments Division of the U.S. Census Bureau to examine the relative level of spending on these services as a percent of personal income, with some discussion of Medicaid from other data sources mixed in.
If you didn’t read this post on comparative state government revenues and expenditures, and this post on comparative local government revenues and expenditures, back in September, you might want to do so now so that you can better understand what you are seeing in the spreadsheets attached to this post.
Among those attachments, I will first discuss “Education and Police All State and County,” which should print on two pages. The data show that in FY 2007 U.S. public school spending was the equivalent of 4.5% of all the personal income of U.S. residents. It is as if everyone in the U.S spent, on average, 4.5% of their personal income on public schools. By this measure, New York State ranked seventh among states at 5.11% while New Jersey ranked fifth at 5.44%. Measured as a share of personal income, however, some states rank differently than some might expect. Texas is number 20 at 4.68%, although that state’s taxes are close the bottom. Its teachers are not well paid by New York standards, but wages and living costs are low in general there, so teaching is not a bad career. Georgia is number 8 at 5.14%. At the same time, affluent Massachusetts and Maryland are number 48 and 41 at 3.74% and 4.03%. This implies that while better paid compared with teachers in Texas and Georgia, presumably, teachers in those more affluent states are less well off relative to their affluent neighbors.
Public school spending in New York City, at 4.22% of personal income in FY 2007, was somewhat below the national average, and would have ranked 34th if it were a separate state. This is an improvement from the mid-to-late 1990s when it typically would have ranked 49th. In the Downstate Suburbs – Nassau, Suffolk, Westchester, Rockland and Putnam – public school spending was 5.23% of personal income, about the same as in New Jersey, and would have ranked sixth if those counties were a separate state.
In Upstate New York, very high public school spending combines with below average incomes to produce very high spending as a percent of personal income and, as other data I have tabulated has shown, very high spending per capital and very high public school employment as a percent of the population. Public school spending in Upstate Urban counties – Erie, Niagara, Monroe, Onondaga, Oneida, Albany, Schenectady, Saratoga, Rensselaer, Broome, Orange and Dutchess averaged 6.30% of personal income, which would have ranked second behind Alaska if those counties were a separate state. In Upstate Rural counties, the average of 7.74% of personal income would have ranked first. Ahead of Alaska, where individual teachers are flown in to teach individual children in the bush. Ahead of South Dakota, were far fewer people are dispersed over a much larger area.
The county data shows broad variation within these parts of New York State. Among the Downstate Suburbs, public school spending as a share of personal income is lower in the counties where average income is highest, Nassau and Westchester, at 4.93% and 4.29%, and much higher in the other counties. Those counties, perhaps, try to keep pace on a per-student basis despite lower resident incomes. In Suffolk County, second homeowners pick up some of the burden on the East End.
While all Upstate Urban are above the national average in public school spending as a percent of personal income, in some the difference is extreme – Broome at 7.33%, Niagara at 7.37% and Orange at 8.20%. In some of New York’s poorest rural counties, such as Allegany (10.8%), Cattaraugus (10.16%) and Hamilton (10.83%), public school spending is a huge share of personal income, although it is substantially funded by second homeowners and state aid.
The second column in the spreadsheet is local government police spending, which in the U.S. as a whole averaged 0.61% of the personal income of U.S. residents. Many of the top ranked states are the “Sand States” which also have the highest foreclosure levels: Arizona is number 3 at 0.82%, California is number 5 at 0.79%, Florida is number 4 at 0.82%, and Nevada is number 2 at 0.85%. Ohio, where the subprime disaster hit early despite not really having a housing bubble is also in the top ten at number 9 and 0.67%, though well below the others. What that correlation may mean I’ll leave to others. New York State is number 6 in police spending as a percent of personal income at 0.74%.
But only because of very high spending in Downstate New York. In the Downstate Suburbs police spending averaged 0.71% of personal income, which would have ranked 7th if these counties were a separate state. And within the Downstate Suburbs, the really high spending is in Nassau (0.94%) and Suffolk(0.69%) counties. That spending is far above the other Downstate Suburbs, or the average for New Jersey (0.59%).
Spending is also sky high as a percent of personal income in New York City, at 0.92% of personal income, or about 50% higher than the national average. Not because its police salaries are high, but because the city has so many police officers. These figures do not include pension spending, which the Census Bureau tabulates separately. Where New York City really stands out is not in paying those who are working – here the city’s pay is low so its workers will be less motivated and qualified – but in generosity to those who do not work and retire early. This shows the influence of the public employee unions, which also like to negotiate lower pay and benefits for those who will be working and producing public services in the future in exchange for a richer sendoff for those cashing in and moving out. Being able to avoid actually having to live in New York City is another key union goal, now largely achieved.
In Upstate New York, on the other hand, police spending is almost uniformly below the U.S. average as a percentage of personal income, with Allegany (0.69%) and Hamilton (1.66%) counties the only exceptions. One difference between education and police spending is that the former can be paid for with state aid. In some upstate counties, police spending is less than half the U.S. average as a percentage of personal income, in part because crimes is remarkably low by U.S. standards in all but a few urban areas upstate.
The “Education All Year” spreadsheet presents public school spending over time in a line graph, and compares New York City, the rest of New York State, the U.S. average and several other states. County-level data is only available in Census of Governments years, every five years, but data for the rest of the state can be had by taking New York State data and subtracting the City of New York.
As expected, the data shows that the rest of New York State has always been far above the U.S. average in public school spending as a percent of its residents’ personal income, while New York City has been below average. School spending was much higher when the Baby Boom and the large Baby Boom echo (Generation Y) were in school, but has remained sky high in the rest of New York State in the 2000s. Note that the U.S average is also higher than it was even as Gen Y moves out of school and into the workforce. That might not have changed with the recession, as public school spending has fallen but so has personal income.
The data has provided one surprise, compared with that I had previously seen. Perhaps because New York City residents’ personal income was revised downward for 2002 (after 9/11) in later Bureau of Economic Analysis releases, the city’s public school spending as a percentage of its personal income actually rose above the U.S. average that year – the only time in all the years for which data is available dating back to 1972.
Looking at other states, public school spending in Texas moved above the U.S. average as a percentage of personal income in the mid-1980s, as income collapsed in the oil bust. As income recovered in that state, however, Texas kept public school spending relatively high as a percent of it. North Carolina was once about average in public school spending as a percentage of personal income, but as that state became more affluent after 1990 school spending was not increased in proportion to personal income, and fell behind the U.S. average. NC has been devastated by the recession, so that might have changed as income fell.
Among affluent Northeastern states, Massachusetts’ public school spending was above the U.S. average as a percent of personal income in the 1970s and early 1980s while New Jersey was below average. Growing affluence, a dwindling number of young people, and property tax caps drove spending in that state to well below the U.S. average as a percent of personal income in recent decades– in fact its public school spending as a percent of personal income ranked 48th among states in FY 2007. In addition, as I have shown, Massachusetts is number one in my “sold out future ranking,” high public employee pension spending, public employee pension underfunding, high debts, and low infrastructure spending. And people in that state are upset the young people are moving away – and won’t be there to pay taxes for their pensions and debts, work to provide them with services, and buy their houses at inflated prices.
New Jersey’s public school spending was about average as a percent of personal income until the early 1990s, jumped up in the deep (for the Northeast) early 1990s recession as income fell, but has remained far above average ever since. Rather than raise taxes, the state chose to pay for this by not funding its public employee pensions, which is about to lead to disaster and deep cuts in education.
In California, spending on public education had been about average as a percent of its residents personal income through most of the 1970s and probably before that. In the wake of Proposition 13, which cut local property taxes, the state’s relative public school funding plunged along with the relative quality of those schools. Another referendum boosted state funding for education, increasing California’s spending to about the national average, or just slightly below, after 2000. Here, as in New Jersey, taxes were not increased to pay for higher school spending, resulting in a recurring fiscal crisis. To see how taxes as a percent of personal income compare, and have changed over the years, you can read this post and download the associated spreadsheets if you have not already.
Finally, I’m about to do something a little naughty, as the “Education and Medicaid” spreadsheet aggregates data from different sources. This is because the Governments division of the U.S. Census Bureau doesn’t provide data on the Medicaid program, only on the “functions” that program is used to fund – Medical Vendor Payments (to private health care providers), Public Hospitals, and Public Health. And the Centers for Medicare and Medicaid does not provide data on Medicaid spending for counties within states, only aggregating data at the state level, so I use New York State Medicaid data instead.
The problem with combining data sources is that they never measure exactly the same thing. For example, in 2007 New York State’s Medicaid spending according to federal data totaled $40 billion, compared with $48 billion in spending on elementary and secondary schools. But according to New York State data, New York State’s Medicaid spending totaled $43 billion. They two data sources might have covered slightly different time periods, and New York’s Medicaid program includes services/beneficiaries that the federal government does not help to pay for, and thus does not include in its data.
So rather than treat the specific numbers as gospel, I suggest just looking at the two charts in the spreadsheet to get an order of magnitude impression. And comparing New York State, New Jersey and the U.S. average based on federal data (no asterisk) while comparing different parts of New York State with each other (with asterisk).
Chart 1 shows that New York State’s spending on Medicaid and public schools combined is well above the U.S. average as a percent of personal income, while New Jersey is about average despite high public school spending. New Jersey has relatively few poor people and a relatively stingy Medicaid program, so is spending on that program is low. With Medicaid and public schools the two largest costs for state and local government by far, high spending in these categories as a share of its residents personal income is a large part of the reason New York’s taxes are also so much higher than average as a percent of income. Burdens from the past, in debt, pension and retiree costs, explains most of the rest of New York’s excess tax burden.
Shifting to state data, one finds that the Downstate Suburbs are similar to New Jersey, with high public school spending offset by low Medicaid spending due to relatively few poor people and high average incomes. But the Downstate Suburbs, which are extremely affluent on average, pay state taxes to support the less well off in other parts of the state. Medicaid plus public school spending in New York City and Upstate Urban Counties are about the same as a percent of personal income, with New York City somewhat higher. This chart makes the pattern of spending within the state look like a deal – excess Medicaid spending in New York City in exchange for excess public school spending elsewhere in the state, all above and beyond the benefits accruing to public school students and the poor and sick.
There is a big difference, however, between funding for public schools and Medicaid: the federal government pays for at least half of all Medicaid costs, but only a small fraction of public school costs. So the effect of $48 billion in public school spending on New York’s state and local tax burden is much greater than $43 billion in Medicaid spending. To show “the deal” with a rough adjustment for this, Chart 2 shows public school spending plus half of Medicaid spending (instead of all of Medicaid spending) as a percent of personal income.
By this measure, the Downstate Suburbs are above the U.S. average and almost as high as New York City, in terms of the state and local cost of these programs as a share of personal income, and New York City is slightly below the Upstate Urban counties. New Jersey is high as well, about the same as the Downstate Suburbs. But one still sees the pattern of lower public school spending (not including pensions) combined with high Medicaid spending in New York City, with very high public school spending (and less Medicaid spending) in the Downstate Suburbs, Upstate Urban counties and New Jersey.
Of course the recession, which has thrown huge numbers of people who previously had employer-provided health insurance into the Medicaid program, may have changed that pattern. My guess is most of the additional Medicaid recipients are outside New York City, with the possible exception of those coming to New York for taxpayer funded care who are actually from other states.
By either measure, public school plus Medicaid spending is a large share of total personal income in Upstate rural counties, where personal income per capita is low. One might consider it hypocritical that those living in such counties are more likely to vote for Republicans, the alleged small government party. One might consider it understandable that in New York and perhaps elsewhere Republicans are in fact the “small government for other people and more for us” party. Or one might conclude that those in rural New York might be prepared to accept that they are subsidized and need help, but believe the form of the help has not been helpful in the long run. The large transfers of wealth to rural New York implied by the charts only go to those connected with the government, with much of it going to those getting one of the excess public school “jobs.” Meanwhile, those trying to make it in the private sector have either had to leave or have become poorer and poorer over the decades. So depending on how you look at it, rural New York has a sweet deal, or a bad one.