Suggestions for Home Rule Commission No. 5: November Elections

It will get more parents to vote and it saves money.

The first two Home Rule Commission public discussion meetings are tomorrow (time and place here), and I hope some of you come out to them and talk about the issues I’ve addressed so far or others that are important to you. It’s worthwhile: the concept of a home rule charter is about becoming more self-governing in Dallas ISD, allowing for more flexibility over state regulations and for tweaks to the way we govern the school system. (I’ll be at the morning meeting; I have a funeral in that conflicts with the afternoon session.)

Now, to my next suggestion. In my previous HRC suggestion post, I talked about some state regulatory flexibility that could help kids. Now, let’s go back to governance tweaks that could help kids.

This is an important point: Why is home rule important for our school system? Because it can help kids. That’s the only reason we should do anything in Dallas ISD – to help kids. (Jonathan Chait at New York magazine does a good job explaining why it’s a bad idea to let schools be seen as anything other than a place to help kids — for example, as a job factory for poor communities.) But a lot of people just don’t understand how a new home rule charter could help kids, because most people don’t think about governance. Trust me: There are some changes that we could make to our governance system that likely will help kids. Research tells us so, as does logic. Let’s talk about a research-based change that makes a lot of sense:

Idea No. 5: November Elections

There are nine trustees on the Dallas school board. Three of them are elected every year in May, to three year terms. This seems like a perfectly logical approach. May elections are automatically non-partisan, just like the school board. Three a year means there’s no major upheaval at any one time, allowing for institutional continuity. But the logic falls apart like a Jenga pile after that, as this setup has several major problems. For example:

  • Nobody votes
  • Parents don’t vote
  • It’s unnecessarily expensive

Nobody votes

When I say nobody votes, this is what I mean:

Voting for Dallas ISD between May 2012 – June 2014:

[table id=1 /]

(Two * notes: Districts 6 and 8 had two elections during this time period, so the number of voters is the average of the two, and the registered voter count is collected from the May election. Since District 2 didn’t hold an election, registered voter stats are harder to come by, so in that case, the registered voter count is taken from the most recent election that was contested.)

No matter how you measure it, turnout in Dallas ISD board elections is well under 5 percent. (Tangent alert! What the hell is up with our gerrymandered trustee districts? Look at those total numbers of registered voters: 79k in one district and 34k in another, even though they’re supposed to be balanced. One man, one vote-ish).

Is under 5 percent low? Totes, yes, of course. Here are some randomly selected turnout figures for comparison:

  • May 2011 – Mayor of Dallas – 12.8%
  • May 2011 – City Council District 4 (Caraway, unopposed) – 7.3%
  • Nov 2010 – Governor – 37.1%
  • Nov 2010 – Non Partisan City of Dallas Wet/Dry Proposition Election – 34.2%
  • Nov 2008 – President – 61.3%
  • Nov 2008 – Non Partisan Parkland Bond Election – 53.8%

Sometimes, raw numbers are hard to read. So let’s put this in picto-graph form:


Yes, 5 percent is really low. How low? So low that Dallas ISD has more total employees (20,793) than it does voters (20,119). So low that if voters were all the DISD parents, each one would have about 16 kids. So low that for every vote cast, the district will spends about $250,000 of our money. So low it proves that a) we are very lazy (me included) and b) maybe we should consider a governance change to make voting easier.

Elections are won because a candidate lines up the support of the major interests groups for that election. The fewer people who vote, the easier it is for a narrow group of adults to carry the day on an election that has far-reaching consequences for a broad swath of our citizenry. We have many special interest groups who participate in the electoral process in Dallas. The big ones are: teachers associations, somewhat partisan advocacy groups like TOP, the business community, civil rights groups, and certain well-connected vendors and family dynasties. They raise the most money, a few of them have formal PACs to fund candidates, and they deploy boots on the ground to get their supporters to vote. If you think their primary interest is raising student outcomes, so be it. But their interests are specific, they’re special, and there’s far from a guarantee that they work to push what is in the best interest of kids. Normally, the push and pull of politics means multiple special interests tend to cancel out and the general interest wins at the ballot box, because to win you have to appeal to the biggest block of voters. But the path to victory is a whole lot easier when you only need to get 600 to 2000 people to vote for their man or woman, instead of, say, 20,000 voters to do so.

Some argue that this low turnout election in May brings out only informed voters. Well, it does bring out the voters who have been informed by those special interests, that is true. But what it mostly does is suppress voting. There are plenty of people in Dallas who know who their city council member is but who just didn’t make it to the polls in May. And unless you’re one of those 2 percent of supervoters out there who make every election, think about your own history: Did you actually make it to the voting booth this April during the primary to pick out who we’re going to elect in November? And again in early May for the school board race? And again in late May for the primary runoff? And again in June for the school board runoff? Me neither. The simple truth is that voters are trained to vote in November, and that’s when they vote. Our May election schedule just means a lot of people forget and simply don’t vote, not that the electorate is better informed.

There’s actually research on why these low turnout May elections are a bad thing. For example, this report notes that low turnout elections for municipal governments negatively affect the interest of Latinos and African Americans, as the local government entities work less on their specific needs. For a school system that is 93 percent Latino and African American, our May election model is a very bad thing for those kids.

Also consider this report that notes that a switch to a higher turnout election structure led to an increase in public policies adopted that favor the working poor. Again, in a school system where 89% of our students are eligible for a free or reduced price lunch, the research would indicate that our current electoral model underperforms for them.

Research solely focused on school systems backs this up. There is real evidence that suggests districts that have even year November elections also have higher levels of student achievement. Page 7 of this report notes: “Merely holding elections at the same time as state- or national-level elections is associated with a $.03 per-dollar-spent-per-student boost, or a proficiency rate about 2.4 points higher than a comparable district that has off-cycle elections.” Geek-speaky, but what it means is that it doesn’t matter how poor kids are, how much money is spent on them, or what their race is. Simply moving to November elections leads to a bump in student achievement over time, because the change in perspective of the elected officials leads to an organization-wide shift in focus to improve the educational experience for kids. It seems kind of magic, but it’s very real.

Ignore this evidence of a positive impact on kids by moving to November. Let’s ask, is there any evidence to support the theory that school board members would change the way they vote on issues if they get elected in even year Novembers instead of every May? Glad you asked. The answer is yes. This ground-breaking “election timing effect” study shows not just correlation, but actual causality, between a move to November and the votes taken by school board trustees — in Texas, no less. The study notes that teachers groups tend to be the dominant groups in low turnout May elections, and a move to even year November resulted in policies that were slightly less friendly to their specific interests, since as a special interest group, they are slightly less dominant in November. I don’t think teachers unions are necessarily the dominant special interest group in May elections throughout Dallas ISD (it kind of depends on each trustee district), but that’s not the important piece here. The critical point is that this study proves that trustees respond to changed electoral dynamics by taking different votes. Which makes sense. They are elected officials. They will respond, in whole or in part, to the whims of their electorates.

Since I just kissed his Super Nerd butt in this month’s D Magazine, it gives me great pleasure to point out the following: This research suggests that even though he has been a supporter of home rule, trustee Mike Morath was wrong in this DMN article when he said, “There is no peer-reviewed analysis of governance out there that proves one system is better.” Au contraire, mon frère. See: above. November even year elections are absolutely better for kids, given what happens to school systems where voter turnout is low. Hey Jeffrey Weiss: Maybe you should call him out on that.

Parents Don’t Vote

I feel like I’ve already proven the point that we need to move to November. But I’ve really only covered one problem with the May elections. The second problem is that parents don’t vote. This is only partly solvable with a move to November, unfortunately, which becomes obvious when we look at why I think parents don’t vote.

Voting is secret in this country, which means nobody can find out who you voted for. However, it’s pretty easy to find out whether you voted, because the county keeps track of that. So, I’ve asked the county for a list of the 20,119 people who cast a vote for a Dallas ISD trustee candidate in the last three years, because the voter ID record is linked to a date of birth, so I can find out how old the voters are. I don’t have the data back yet, and I wanted to wait until I did to put this post up. But I also want this post out there before tomorrow’s meetings commence. So for now, let’s use the average of “likely voters” in each district. That number comes out to 64 years young. If this number is anywhere close to the final number I get, you’d be hard-pressed to prove these are parents of current DISD students. (If I’m way off and the average age is much lower, I’ll mea culpa this noise.)

If the elections were held in November, what would the age of the average voter be? That’s easy enough to figure out, too, since we just look for folks who cast a ballot for (as an example) one of the state Supreme Court justices during the last gubernatorial election in November 2010. (As you can see from the above down-ballot turnout numbers listed in the chart above, even down ballot races have high levels of turnout in even year Novembers.) I’m of course waiting on this information, too. But we can use a pretty good placeholder here, which is the average age of those who vote in presidential elections, which is about 44 years old.

The age issue can be fixed by changing election timing, which likely means far more parents vote. But there is a different issue in terms of getting parents to vote that is much harder to fix. If you look closely at the chart above on registered voters by district, you see that districts 4, 7, and 8 are majority Hispanic but have the lowest registered voter counts. This is more fully discussed in the district’s formal 2011 redistricting filing. There is a term called Spanish Surname Registered Voter (SSRV) that refers to the percent of registered voters that have a name that seems Hispanic. It’s a proxy for registered Hispanic voters. In a district with 69 percent Hispanic voting aged population, there is only a 38 percent SSRV. So only half of Hispanic adults are actually registered to vote to begin with in those districts.

This is bad, since 70 percent of the kids in DISD are Hispanic. If 70 percent of voters were Hispanic, it’d be easier to think that parents were voting. But since the SSRV percentage for all of DISD is only about 21 percent (as opposed to Hispanic voting age population, which is about 40 percent), it’s unlikely that we’ll see an electorate that is dominated by parents any time soon. So moving the election date will help get more parents to participate, but if we want to make sure the board really represents the interests of parents, we’ll need more Hispanics voting. This post is already getting a bit long, even for me, so we’ll table that discussion.

I’m hopeful that you and at least one Home Rule Commission member is now saying something like this: “I get it that May elections mean fewer parents vote, and that means more special-interest influence, but why again is that a bad thing?”

The reason comes back to why it is we have a school system in the first place. We only have a school system because we want kids to get a good education. It’s the job of a school board is to make sure that happens. But there are really three major somewhat opposing interest groups in this battle: parents and their kids who consume the system’s product, teachers and other employees and vendors who make the system’s product, and taxpayers who fund the system. Each of these groups is interested in better educational outcomes for kids. But only parents and their kids live and die by those educational outcomes. Teachers, employees, and vendors want kids to have better outcomes, but that’s not their primary core interest. Same thing with taxpayers.

I’m not saying teachers don’t want the absolute best outcomes for kids. Of course they do. But they do work for money, as does pretty much everybody else in the world. I asked a great teacher last week if she would be okay with longer school years or days. “If they pay me more, we can talk about it.” Which is the most sane, reasonable response in the world. But it’s not, “What does the research say is best?”  This is their livelihood. They’ve got to protect themselves. I get it. And it’s the union’s job to do the protecting. What would the response of the teachers union be if Professor Frink invented a robot that taught kids how to read and write and do math and be compassionate and grateful and empathetic and hardworking without the need of a human teacher? That’s not likely to occur, but if it did, I bet we’d see pretty quickly whether there was really perfect alignment between the interests of teachers unions and student outcomes. Again, there is nothing wrong with this. It just is what it is.

I’m not saying the taxpaying public isn’t interested in the best outcomes for kids. But if Professor Frink said those robots were available for $5 a unit, I think those taxpayers would start asking some pretty interesting questions of their elected trustee as to why the school system still spends $60k a year on teachers.

What about parents? They want the best outcomes for kids, period. End of discussion. (Sometimes, too much so, but that’s an okay problem to have.) So, more of them voting means that the taxpayers or the employee groups don’t necessarily monopolize the attention of elected trustees. This likely means the trustees focus more of their attention on improving outcomes for kids rather than, for example, better pay for custodians or a bigger homestead exemption (which are both worthy goals, just less of a priority than improved outcomes for kids). That’s why a move to November elections so that more parents vote will likely be good for our kids.

May Elections are More Expensive

I don’t know that I need to keep giving you reasons to prefer November elections, but I think I’m about to break a blog word-count record, so let’s keep going.

Let’s say you don’t care about research on student outcomes. And you think parents suck and you like the board members just fine. Well, do you like money? Cheese? Scrilla? Coin of the realm?

Here’s the cost on the last three May elections in Dallas ISD:

May 2014 $329,451.00 (later reduced because of no election in District 2)
May 2013 $159,721.65
May 2012 $583,214.58

Elections vary a bit based on the number of poll locations and the number of other jurisdictions with which the costs are shared. But under the three-every-May scenario, electing all Trustees costs about $1,075,000.

If the district were to make the switch to even year Novembers, this would necessitate a switch to four-year terms. Five trustees would be elected in one November, to four-year terms. Two years later, four more trustees would be elected to four-year terms, also in November. Rinse and repeat. (To make the transition to this model, you could have all nine trustees elected the next even year November, and four would be selected randomly to have an initial two-year term. Or something like that.)

Lucky for me, the Dallas Morning News figured out what it costs to get all nine trustees elected in November, except they didn’t split the cost over multiple Novembers to compare it to the annual election of three trustees. But whatevs. They’ve still answered the question as to the cost to elect five and then four trustees across two November ballots, because it’s basically the same as doing all nine at once. That cost: $595,000.

Over a 12-year period, taxpayers are looking at these two cost scenarios:

– Keeping the current May election system: $4.3 million
– Switching to November elections: $1.785 million

If the Home Rule Commission adopts this recommendation, then, it will save taxpayers $2.5 million every dozen years. That covers the cost of the home rule conversation and gives us an extra 25 teachers.

I know that, despite the math-y reasons, some argue that having elections that can swap out one-third of the board every year is preferable to possibly flipping half the board every two. But this doesn’t make sense for a couple of reasons. One, in the scenario where one-third of the board can be flipped every year, then every two years, a supermajority of the board can be flipped. That’s actually more disruptive to the district than the possibility of a slim majority or big plurality changing over every two years. Four-year terms also allow for more organizational continuity because trustees will be in place at least two years any time a new election occurs, and they will much better positioned to mentor newly elected board members on what it is they’re supposed to be doing (not that the new ones ever listen).

Why anybody thinks it is a good idea to have one-third of the board campaigning every single year is beyond me. It means politics is always front and center for at least three of its members. (More than that, really, unless you think these things happen in a vacuum and the rest of the board members don’t concern themselves with their colleagues’ campaigns.)

Can anybody reasonably argue that our current structure has led to great continuity of leadership? Dallas ISD has had nine superintendents in the past 20 years. The every-year-changing board leads to an ever-increasing erosion of support for superintendents, a key reason we’ve had so many. This is bad, because research has shown a solid correlation between longevity of superintendent and student outcomes. But — damnit — I’m talking about helping kids again instead of saving money. How did we end up here?


Keep me up to date on the latest happenings and all that D Magazine has to offer.