The U.S. Census Bureau is apparently busy crunching numbers for the 2007 Census of Governments, and the first data from that effort — the organization phase — has been released. The data is limited — the number of local governments by category (counties; municipalities such as cities, towns, and villages; school districts; and other special districts) in each state. But as it happens it is relevant to a debate going on in New York State right now — can the state’s high taxes and unaccountable politicians be explained by the large number of overlapping local governments in the portion of the state outside New York City? And can it be improved by consolidating those local governments?
It certainly makes sense in theory. Large organizations have economies of scale that small ones lack, and the more local governments you have, the more mayors, city council members, and school board members, not to mention personnel officers, chief accountants, and superintendents, you have to support. Elected officials with a lot of power attract a lot of attention and, therefore, more electoral accountability, while many special district elections don’t even happen on Election Day, with few voters taking part and virtually no media coverage. But does the theory hold in general? The data, in the attached table, is inconclusive but seems to favor consolidation a little bit.
First let’s take New York City out of equation and look at the rest of the state separately. New York City has just three local governments — the City of New York, (part of) the Port Authority of New York and New Jersey, and a third I’ll give a quarter to anyone who can name it — for over 8 million people, and would really skew the analysis for that reason. (The MTA is a state agency, with New York City Transit tabulated by the Census Bureau as part of the City of New York for historical reasons, while C.U.N.Y. is also a state agency).
There are two ways to compare the number of local governments by state. By the number of governments per 100,000 people, because the more local governments there are relative to population the more that population has to support. And by square mile, because no one wants City Hall to be 100 miles away, so one needs a certain number of local governments even if the population is sparse. For example, the 249.1 local governments per 100,000 people in North Dakota is far higher than the national average of 29.7, and ranked second among the states (behind North Dakota), but the 26.1 local governments per 1,000 square miles is only slightly higher than the national average of 25.3, and ranked 31st. On the other hand the small state of New Jersey has 186.5 local governments per 1,000 square miles, first in the nation and enough to confirm its reputation for fragmented and inefficient localities. Except that at 15.9 local governments per 100,000 people, densely populated New Jersey is little more than half the national average and ranked 36th.
While a less extreme case, the rest of New York State is similar to New Jersey. It may surprise people to find out that its total number local governments, at 30.9 per 100,000 people, is only slightly above the national average of 29.7 (or would have been if the rest of the state were a separate state). With 6.5 school districts per 100,000 people, the rest of NY ranked 21st and was well above the national average of 4.8. But with 10.1 other special districts per 100,000 residents, the rest of NY was actually below the national average of 12.4, and ranked 39th.
On the other hand, the rest of NY’s 75.5 local governments per 1,000 square miles is three times the national average of 25.3. Clearly plenty of people are not driving far to City Hall or the School District headquarters. Many of the other states with many local governments relative to land area are, like New York and New Jersey, in the Northeast: Delaware (2), Connecticut (3 despite having eliminated county government), Rhode Island (4), Massachusetts (6), Pennsylvania (7), Vermont (10) and New Hampshire (12). Illinois, Ohio and Indiana are also in the top 10. The rest of New York ranked 10th in school districts per 1,000 square miles, with 15.3 compared with 4.1 nationally, and 19th in other special districts per 1,000 square miles, with 23.8 compared with 10.6 nationally.
What about the relationship with taxes?
One certainly finds some high tax states with lots of local governments, like Vermont. It ranked 7th in local governments per 100,000 people, 10th in local governments per 1,000 square miles, and sixth in local taxes revenues per $1,000 of its residents’ personal income in 2005. And one finds some low tax states with relatively few local governments. Tennessee ranked 39th in local governments per 100,000 people, 36th in local governments per 1,000 square miles — and next to last in taxes as a share of income.
On the other hand, Louisiana ranked near the bottom in local governments per 100,000 people (42) and per 1,000 square miles (43) and near the top in tax revenues per $1,000 of personal income (4). So did Hawaii (50, 49, 7). Yet there are no examples of states where both the number of local governments per 100,000 people and the number of local governments per 1,000 square miles are high and taxes are low. In that sense, the value of consolidation is proven.
However, for those of you who remember graduate school, the RSQ function of my spreadsheet gives a result of 0.003 for r-squared with local governments per 100,000 residents as an explanation of comparative taxes, and 0.010 for local governments per 1,000 square miles as an explanation. If the number of local governments were solely responsible for the relative level of state and local taxes, that statistic would be 1.0, if it had no effect at all the statistic would be zero. Did I do that right? It’s been a long time (perhaps statistics classes should spend more than a few minutes on descriptive statistics, and non-sampling error, and less time on correlation). This calculation doesn’t make consolidation seem like such a big issue, although in any event a look at the rankings seems to confirm that to the extent that the number of local governments affects taxes, the number per 1,000 square miles — where the rest of NY is very high — is more important than the population ratio.
How about some examples closer to home? New York City has about the fewest local governments anywhere. And sure enough New York City’s public school expenditures on general administration, according to data discussed earlier, is only one third the national average (even without adjustment for the cost of living), and a much smaller fraction of the rest of the state. The city has major economies of scale elsewhere as well. And yet, the city’s taxes per $1,000 of its residents’ personal income would have ranked first if the city was a separate state in FY 2005, and is always in the top three.
One possible reason is diseconomies of scale offsetting the economies of scale. Sprawling governments need lots of layers to control their extensive and diverse operations. Up to some point local governments can save money by being able to support one sizable high school, one substantial police precinct, and one firehouse with more taxpayers. But New York City does not have just one of any of those — it has many, with an intermediate level of command between them and headquarters. It is very unlikely the NYC School Chancellor could spend a day in every school before his or her time was up.
Sprawling governments also have lots of backwaters where corruption can flourish. If you want to work in real estate you can make lots of money in the private sector, and if you want to work in transit you want work with buses and trains — so MTA real estate was ignored and became a corruption sinkhole. If you want to work in education you want to work in a school, and if you want to work in logistics you can make more money in a private firm — so the supply bureau of the old Board of Education was a place where lots of stuff tended to disappear, along with a couple of workers to dared to complain about it. The Staten Island Ferry was also a patronage backwater that key decisionmakers didn’t pay attention to, until it crashed.
Since public employees cannot be fired no matter how little they contribute without extensive documentation, moreover, moreover, public agencies need to be top heavy – with lots of supervision relative to work done — for anything to happen. That eliminates some of the savings that large private sector organizations have, where workers are expected to do their jobs on their own and are fired if the relatively few managers get the impression (without hundreds of manhours of direct observation) that they aren’t. It’s a waste no matter how many governments there are. Because of their labor environment, and the political pressure for more positions for people to be promoted to, it’s hard for large public organizations to “get flat” in their org chart, to use the business buzzword.
And as for contracting and specialized skills, these are easier to support in large organization, but consolidating purchasing and such skills while retaining others in-house is a solution smaller government organizations could pursue, and the Nassau County Executive Tom Suozzi has suggested.
How about electoral accountability? A graduate school law professor, thinking of New Jersey, said the more governments there are, the less attention is paid to them, and therefore the less real democracy there is. He suggested consolidating many local government functions at the county level. Here in New York State, however, we have but one state government. New York City is rare in having a powerful local government with a wide scope of activities, and this focuses attention on the Mayor and City Council, but state governments are important and powerful everywhere, including here. But how accountable is the very powerful and important New York State Legislature? How many real elections do its members face?
I conclude that consolidation, particularly of school districts, and shared or contracted out services, particularly those outside a government’s core mission, can help improve efficiency in the rest of New York State, but other factors are more important. Our biggest problem is a bad state government dominated by relatively few self-interested people and groups. And we only have one of those.