I will protect your pensions. Nothing about your pension is going to change when I am governor. - Chris Christie, "An Open Letter to the Teachers of NJ" October, 2009

Monday, April 30, 2018

Don't Blame Teachers For School Underfunding: A Data Tale From Jersey City

The animosity between NJ Senate President Steve Sweeney and the NJEA, New Jersey's largest teachers union, is already well-known. Add to that the rivalry between Sweeney and Jersey City mayor Steve Fulop, and Sweeney's desire to amend the state's school funding system... well, Sweeney's latest dig at Jersey City's teachers and board of education really shouldn't have surprised anyone:
In a statement issued Friday, state Senate President Stephen Sweeney blasted the Jersey City Board of Education for approving the agreement, which will increase district spending on teacher salaries by 3.31 percent during the current school year and 2.72 percent during 2018-19. The board approved the contract by a 5-1 vote Thursday night. 
Sweeney (D-Gloucester) said the Jersey City school district already receives more state funding than it should – district officials have dismissed this as untrue. Sweeney added that salary increases amid a $71 million shortfall in the district's proposed budget sends the wrong message to other schools.
"What makes it even worse is that the Jersey City Board of Education wrote a blank check that taxpayers in every other school district in New Jersey are going to have to reach into their pockets to pay," Sweeney said. "That's because Jersey City continues to get $151 million a year more in state aid than it would be receiving if the school funding formula was run fairly with the 10-year-old growth caps and Adjustment Aid eliminated." [emphasis mine]
Others have reported Sweeney claims Jersey City is over aided by $174 million; let's stick with the lower figure for now to be conservative (you'll see why in a minute). Sweeney arrives at this figure because Jersey City, and several other districts, benefit from a provision in the School Funding Reform Act (SFRA) called "adjustment aid." This aid was included in the original 2008 law to mitigate against the shock school districts might face when transitioning to the new formula; it keeps districts from falling below the level of aid they received prior to the new law. However, it has also led to some districts currently receiving more state aid than they would get if the provision wasn't included.

Jersey City gets a lot of adjustment aid, which likely helps it keep its local taxes lower than they would be otherwise. To illustrate, I took this chart from the Education Law Center's website†:

There really is little doubt Jersey City should be contributing more local tax revenues toward its schools; whether it can at the moment, given the state's property tax cap, is an open question.* That said, and as ELC** points out in this brief, the district is still not getting all the funding it needs, from either the state or local sources, to provide an adequate education for its students.

Which makes Sweeney's statement even more interesting. Because his clear implication is that Jersey City is giving its teachers a big raise*** on the backs of other school districts, who don't get nearly as much state aid. But he's also claiming property taxes in Jersey City are artificially low, again because of an excess amount of state aid.

Is this possible? Is Jersey City so "over-aided" that can afford big teacher salaries and low property taxes?

Again, I'll leave aside the question of taxes and instead focus on teacher salaries. Because I happen to have data available to take a reasonable stab at answering this question: Are Jersey City's teachers significantly overpaid compared to their colleagues in neighboring school districts? If not, is it really fair of Sweeney to call this recent contract irresponsible?

Let's start by looking at how much JC's teachers make compared to their colleagues in the other school districts in Hudson County (click to enlarge).

At first glance, when we look just at the average Jersey City salary compared to the rest of the county, it appears JC teachers are doing relatively well -- not spectacularly well, but well. Bayonne, Gutenberg, Weehawken and East Newark**** teachers seem to pay a serious wage penalty for not working in JC...

Or do they? One of the problems with simply comparing average (or even median) salaries is that it doesn't account for how teachers are paid in the real world. For example:

Like all public school teachers (and like many, many others in both the public and private sector), Hudson County teachers are paid more when they have more experience; this explains the upward slope of these lines, showing pays raises when teachers gain seniority. Jersey City (the dashed red line) has a slightly earlier bump up in experience than most other Hudson County districts.

However, when JC teachers reach their 30th year, their pay is rather average. In fact, the best-paying district in Hudson County, accounting for experience, appears to be Hudson County Vo-Tech. Which, again, is interesting, given Sweeney's full-throated support for vo-tech schools.****

Now, whether Jersey City is paying relatively more than other districts for its teachers also depends on how experience is distributed. So let's look at that next:

Jersey City does have a somewhat larger concentration of teachers with 15 to 19 years of experience; that might help explain a somewhat higher average salary for all JC teachers than other Hudson County districts.

But teacher pay doesn't just vary with experience. Earning an advanced degree leads to higher pay; living in a labor market that's more expensive, or pays more for teachers relative to other professions, changes pay. Keep in mind: these factors are out of control of both the Jersey City Board of Education and the Jersey City Education Association, the local union that negotiated the contract. It's ridiculous to think either party could buck trends and norms followed across the state.

So how can we determine whether Jersey City teachers are really "overpaid"? I've approached the problem using a regression model: a statistical technique that allows us to "hold things constant." Using seven years of data on every teacher in the state, I've tried to model how experience, full-time/part-time status, labor market, job description, highest degree earned, and other factors affect teacher pay (nerds, I give the details on the regression model below).

The model allows us to predict how much a teacher might earn, given all these factors. The amount above or below prediction (the residual) can't be explained by the variables in the model; we will assume, therefore, that this amount is how much each teacher is "over-" or "under-" paid, relative to other teachers in the state.

So: are Jersey City teachers way overpaid? Put simply: no, not really.
This is expressed as a ratio of actual salary over predicted salary; a ratio of "1" means the salary is exactly what the model predicts, so the teacher isn't "over-" or "under-" paid, given their experience, degree, labor market, etc.

In Jersey City in 2016-17, teachers (as a group) were paid about 3.7 percent more than prediction. That hardly makes them the most "overpaid" teachers in Hudson County: Harrison, Hoboken, Secaucus, and Hudson Vo-Tech teachers were all "overpaid" more Jersey City school staff (again, this doesn't account for administrators, nor for staff without certificates).

Let me stop here and clarify something: I am deliberately putting "under-" and "over-" paid in quotes, because this model cannot account for many other factors that would affect teacher pay. It may be that Jersey City has to pay more to attract the same quality of teacher candidate for a variety of reasons that can't be measured. Maybe teacher candidates didn't want to teach in a district that was under state control for a quarter of a century. Maybe they've heard, as I have, that the state monitors have made staff feel unappreciated. Maybe the traffic sucks.

All I'm trying to do here is provide some sort of empirical analysis to determine whether there's evidence that Jersey City teachers are the beneficiaries of the "over-aiding" of the district. To that end: let's see what the "overpayment"****** of Jersey City teachers costs the district.

I could choose all sorts of denominators to use, but let's keep this simple: how much of the total appropriations of the Jersey City Public Schools can be attributed to the "overpayment" of teachers? About 1.3 percent.

But let's get back to Senator Sweeney's complaint: how much of the "over-aiding" of Jersey City gets gobbled up by the "overpayment" of Jersey City's teachers? About 6 percent -- that's barely a blip.

The idea that Jersey City's teachers substantially benefit from of the "over-aiding" of the district is not supported by a reasonable analysis of the available data.

I'm going to run the risk of pissing off a few friends here, but let me put this on the table:

Senator Sweeney and I have a lot of disagreements. I was, like almost every other teacher in the state, extremely disappointed by his support of Chris Christie's attack on our pensions and health benefits. I think Senator Sweeney is dead wrong about the benefits -- and largely blind to the harms -- of the expansion of charter schools in Camden (call them whatever you want, they're charter schools). I also think Senator Sweeney is dead wrong on taxation.

That said: Steve Sweeney has valid concerns about New Jersey's state school aid formula. He is right to note that the growth caps have got to be addressed. He is right to state that communities like Jersey City ought to be contributing more toward the funding of their schools. He is right to champion the districts in this state that are often overlooked in the debate over school funding, yet whose students are suffering real harm due to inadequate funding.

So I'm willing to take Steve Sweeney at his word. I do believe he is concerned that there are students in New Jersey school districts who are suffering right now because they can't get adequate funding for their schools.


The idea that the students of Bayonne are being denied an adequate education because of the greed of the teachers of Jersey City is just plain wrong.

There is no evidence Jersey City teachers are wildly overpaid. There is no evidence the small bump JCEA members enjoy in their wages is a major part of the "over-aiding" of the district. I understand NJEA gave Sweeney a few bruises. But making arguments that pin the blame for the underfunding of New Jersey schools on Jersey City's teachers is not helpful in the slightest.

Look, schools cost what they cost. If you want certain outcomes, you have to pay for them (we need to have a good long talk about this idea soon...). By the state's own formula, Jersey City's schools are not over-funded.

In addition: if you want good teachers, you need to pay good wages. New Jersey actually underpays its teachers relative to the rest of the labor market. If Jersey City is paying its teachers a bit more, that's a good thing. Why come down on the district for trying to get good people to come into the profession?

Senator Sweeney, instead of slamming Jersey City's teachers for standing up for themselves and demanding decent pay...

Why don't we instead work to get all districts the funding they need to bring the best and the brightest into New Jersey's classrooms?

For the record: I am a proud NJEA member, and I am proud to stand with my fellow public school teachers in Jersey City, and everywhere else in the state.

* I really don't want to wade into this on this post, because, to be honest, I just haven't had time to look at it carefully. But some, like Jeff Bennett, argue Jersey City could increase its revenues without the state raising its property tax cap. Bennett (who, despite our policy differences, I genuinely respect) has told me Jersey City hasn't even raised its tax rates as high as it could under the current cap. I have no reason to doubt Jeff, but I haven't looked into the topic myself. 

** For the record: I have done work as a contractor for ELC in the past.

*** Something worth noting: when you see a report that teachers are getting a "... 3.31 percent during the current school year and 2.72 percent during 2018-19," understand that doesn't mean all of the teachers are getting more money. Public schools operate on salary guides, which provide a raise for every year of service up until a final "step." You need to add money into guide like that just to maintain it. So those at the "top of the guide" might actually be getting no raise, depending on how the guide is structured.

Teacher salary guides is a really complex topic; maybe I'll try to get to it at some point...

**** Actually, the East Newark data for 2016-17 looks off because a lot of the teachers who should be 1.0 full-time equivalents are listed as 0.1 FTEs. I tried as best as I could to clean up this rather obvious mistake.

***** To be clear: I join with Senator Sweeney in supporting vo-tech programs and schools. More Vo-Tech!

I just don't understand why the senator is complaining about Jersey City teachers getting a raise when they make less than the county's vo-tech school. Why isn't he blaming them for underfunding elsewhere? (OK, he shouldn't, but you get my point, right?)

****** Yes, these quotes are stupid. You have a better idea?

The Regression Model:

I have a panel of certificated staff data from 2010 to 2017. 2013 is excluded because some of the teacher characteristics data weren't included. The model I use is:
salary = f(prior_exp_years FTE i.highest_ed_comp i.metajobcode i.lmencode i.data_year i.charter charter#data_year logEnroll)

  • prior_exp_years: Total years of experience, in and out of NJ or the district.
  • FTE: Full-time equivalency.
  • highest_ed_comp: Highest degree earned.
  • metajobcode: Job description, divided into larger categories (i.e., all science teachers bundled)
  • lmencode: Labor market; I used counties. 
  • data_year: The year. 
  • charter: Whether the school is a charter. I know some of you might push back a bit, but the fact is a teacher suffers a wage penalty for working in a charter. Given that reality, it's not rational to expect Jersey City teachers to make charter school wages; in fact, there is a very good case to be made that JCPS teachers are propping up the city's burgeoning charter sector through wage free-riding
  • charter#data_year: Given the volatility of the state's charter sector, interacting it over time seemed reasonable. 
  • logEnroll: OK, so this one had me thinking. We know for a fact that school districts enjoy economies of scale. It may well be those districts then use the savings to recruit more desirable teacher candidates, or make up for recruitment hardships that can't be measured. It may also, however, be that larger districts create larger teachers unions, which leverage more bargaining power. But do districts really have much control over how big they are? Hmm... Ultimately, I kept this in the model because it matters -- but I'm open to debate. In any case, removing it does up the "overpayment" ratio for Jersey City, but only to about 1.06. That's not enough to make a serious dent in the amount JC is "over-aided."
Bad mistake in the original post: I inadvertently put Newark's LFS v. Levy chart up, not Jersey City's. Sorry about that -- correction made.