Northwestern Professor Kirabo Jackson explains the economics of education investment. The key? Start with early learning and continue.
Transcript: Kirabo Jackson Conversation
We frequently hear the complaint: Our education systems – particularly public education – are broken. Invest in these approaches? That’s just throwing good money after bad.
For example, investing in Head Start may make sense. In other cases, investing in K-12 might be the right approach. But coordinate and sustain ongoing investment in both? Forget it.
Only, it turns, out: That’s not what the data show. And that’s not the conclusion of Kirabo Jackson, a Northwestern University professor, who has analyzed that data, the trends, and the outcomes.
In fact, as Jackson explains: It’s specifically the continued, ongoing investment in kids and their education that delivers exponential results. As you’ll hear, investing in Head Start plus investing in K-12 – that math is simple: 1 + 1 = 3.
A word about Prof. Jackson: He’s the Professor of Human Development and Social Policy and Faculty Fellow in the Institute for Policy Research at Northwestern’s School of Education & Social Policy. An economist by training, his research has been published in articles have appeared in leading economics journals such as the Quarterly Journal of Economics, American Economic Journal, Journal of Labor Economics, The Review of Economics and Statistics, among others.
He’s turned his economic world view to the world of education, and in 2016, Education Week listed him among the top university-based scholars who are doing the most to influence educational policy and practice.
Chris Riback: Professor Jackson is one to watch and listen to. So here’s my conversation with Professor Kirabo Jackson. Kirabo, thanks for joining me. I appreciate your time.
Kirabo Jackson: It’s my pleasure. Thanks for having me.
Chris Riback: So of all the public policy areas to choose to dedicate your career to, you didn’t want something with slightly less controversy than education? I mean, what about infrastructure? No one argues about infrastructure.
Kirabo Jackson: This is true, and infrastructure is certainly important, but my sense in education is really at the heart of understanding society. A better-educated society, I think, is a more functional society, it’s a more productive society, and it’s also a more equal society. So for a variety of reasons education seemed to be what was most, if you want to pull a lever to improve society, I think education is a good one.
Chris Riback: OK. Yes you can improve society through education, but you have stepped into it – so welcome into the quicksand! But on to the more serious front: The point that you’re making is exactly what it seems. So much of your work and so much of your study, and in particular, the study that I want to ask you about today, is the role for education on a path to greater equality and greater opportunity. So let’s get into it. One of the key areas that you study, public school funding, particularly its effects on student outcomes through adulthood. Let’s start broadly. Public schooling does not seem to be, at least to me, in vogue these days. It feels like everywhere you look, the talk is vouchers and charter schools and privatized education. Why does public schooling still matter?
It still improves outcomes if you increase spending, and that was what I was interested to examine.”
— Prof. Kirabo Jackson
Kirabo Jackson: I think the part of the conversation, the tenor of the conversation that we have now in part is driven by a frustration, some of it which I think is reasonable. Frustration regarding the extent to which we are spending a good amount of money in public schools and it does not appear as though it is necessarily being spent in the most effective way. And for sure there are many cases and examples one can give in a large nation such as the United States where money is not spent productively, it’s being wasted. But I think oftentimes when you focus too much on the negative, we sort of lose sight of the positive, and it is entirely possible that money is not spent as effectively as it could be. But it still improves outcomes if you increase spending, and that was kind of what I was interested to examine.
So it absolutely matters. I mean, many, many children, most of our nation’s children are going to public schools and if it’s the case that increasing resources that are available to public schools isn’t going to improve their outcomes, that’s an important thing to demonstrate. The way I like to think about it is this, even if it’s the case that 40 cents on the dollar that goes into the public schooling system is wasted, which is certainly a bad thing, that 60 cents that goes into the public schooling system may have a positive impact. And the question is not whether we’re getting the best bang for your buck, but are we getting enough to justify the expense?
And my research basically is examining that question and actually my conclusion is probably answer is yes. And that is not to say we shouldn’t focus on spending money as effectively as possible, but I see those two things as being somewhat separate questions.
Chris Riback: So let’s walk through it, and yes, the punch line is there, and we will encourage folks to stay with us and we’ll explain how we get to the punch line. But yeah, that’s exactly where it ends up. So the long-run effect of Head Start and public school spending. And by the way, as a side note, looking at the research that you’ve done and teacher labor markets and the effects of teachers on social emotional skills and the effects of single-sex education, there are so many conversations that you and I could have that I would love to have with you on education, but we’ll focus on this one in the long run effective Head Start and public school spending. And you may have just answered it, but what inspired you to do the research here in the first place? Was it a sense that something was off around the public conversation? And then what was your hypothesis going in? What were you hoping to prove or disprove?
Kirabo Jackson: So looking at the Head Start piece, the early childhood education piece. So Rucker Johnson, Michael Walter and I, if you look at the literature that had been done on this, there has been a few studies essentially documenting that early childhood programs sometimes improve outcomes and these outcomes are typically measured as sort of test scores and that they often fade over time. And another finding in the literature, and by fate of our time I mean the children who participate, they look better than those who don’t participate for a few years. But then by about third or fourth grade, it looks as though they look very similar to those who did not participate in Head Start.
And there’s another set of papers that kind of document that the impacts appear to be larger for children from more advantaged backgrounds. So typically, children have from white families tend to benefit more from Head Start than children from black families. So one of the thoughts that sort of came to mind is, well, it’s possible that what’s happening is many children are enrolling Head Start, these early childhood interventions and they’re gaining some skills. But if you’re not sustaining those skills with a sort of nourishing environment, perhaps essentially the benefits are lost over time. So our thinking was that one possible explanation is that many children are enrolling in Head Start, enrolling early childhood or pre-K and they’re not getting the additional enrichment to sustain the gains that they last over time.
And this may be particularly true for individuals from less advantaged backgrounds such as single African-American mothers. So that was where we sort of had the idea that perhaps the benefits of Head Start may differ among those that subsequently attend public K-12 schools between the ages of 5 and 17 that are better funded and better able to sustain any of the initial gains that would have been produced by Head Start.
Chris Riback: How do you design such a study? How do you measure kids’ progress over years? There’s so many different variables. Education in one town may or may not be similar to education in another location. How do you design the study?
Kirabo Jackson: That’s a great question. So one of the things that we do is we actually focused on the longer run outcomes of children and we didn’t focus on test scores. One of the benefits of that is we can sort of sidestep a lot of the questions about, well, what if the test here is aligned to a different curriculum than others? That’s one if you have to sort of contend with. But also ultimately I think the goal of many of these early childhood interventions is not simply to improve test scores, but it’s to improve a lot of the softer skills, executive functioning, self control. A lot of things that are actually very much involved now in some of the more productive Charter school system is making sure kids have those softer skills is something that is emphasized by many of these early childhood programs.
They’re not necessarily well measured by test scores, so by looking at long run outcomes like whether children actually graduate from high school, whether they go to college, and actually even those subsequent earnings is a way to go to get a much more broad sense of whether children are doing better in the long run, using a measure that is more broad. So that’s one, that’s how we sort of addressed that issue and sidestep some of the issues that you raised about how do you measure something in one town versus another town. But then the other question I think that you’re sort of getting at is how do you maybe parse out the influences of safe family background from the influences of Head Start, is that sort of what you’re getting at also?
Chris Riback: Yes. Any of the influences. I mean, that’s the first one that comes to mind. I’m sure you could probably tick off 10 others but yes, exactly.
Kirabo Jackson: So the way that we do that is to try to rely on comparisons that are among individuals that are very very similar in all dimensions other than the fact that some had access to Head Start and some did not. So one thing that we do not do in the study is we don’t just compare individuals who enrolled in Head Start to those who didn’t. We don’t just compare children who participate in Head Start and who didn’t. We take a step back and say, well, we can actually examine the outcomes of children who had access to Head Start because Head Start was available in their town when they were four years old compared to children who did not have Head Start in their town.
And we can even do that within the same family. So within the same family we can ask the question, what happens if one child had access to Head Start because Head Start was available in their town, but their older sibling who was maybe six years old when Head Start was rolled out in their county did not have access and that’s one of the comparisons that we make in our people. Sort of making comparisons among individuals from the same location, sometimes from the same families, but participated in some of which participate in Head Start but some of whom didn’t. And the reasons why there’s differences in participation was actually because there was differences in access, which is beyond the parents’ control.
Essentially if a parent lives in a neighborhood, especially low income parent who lives in a neighborhood, whether Head Start has been introduced in their town or not, has nothing to do with their individual characteristics, so that when we make a comparison like that we’re removing the influence of things like family background, which could potentially make such comparisons difficult in other settings,
Chris Riback: What was the Coleman report?
Kirabo Jackson: The Coleman was a very important report that was commissioned by the department of Ed in the ’60s. It came out, I believe it was 1966 and James Coleman who was a sociologist and his research team they compiled, I think it was the first time anyone had actually linked individual measures of student achievements which is standardized tests to a family characteristics and to school input measures. And he and his team came up with a few broad conclusions that have been very influential for policy over time. So the first of them was that family background was extremely predictive of student achievement as measured by test scores.
And that’s not really surprising, but it was one of the key findings. But the key finding that is often somewhat more controversial is that they show that in their data, once you account for family background, traditional measures of school quality such as school spending or even class size are unrelated to student achievement. So many have looked at that and came to the conclusion that well, it’s all about the families, schools don’t really matter that much. That’s actually not the appropriate takeaway, I should say, even from their analysis. So even though it’s true in the Coleman data that going to a school that has a higher per pupil spending or going to a school that has sort of smaller classes does not change student achievement.
They did find that there were certain schools that had better achievement than others, it just wasn’t associated with whether they were well resourced or not. But even at the end of the day, I would argue that that study is largely based on correlations or it is based entirely on correlations. So that even though it was very influential, it doesn’t answer the policy question of if we were to improve schools, does that change outcomes? Or if we were to move a kid from a disadvantaged school to a better school, does that improve outcomes? And that’s basically the kinds of questions that I’m trying to answer in this research.
Chris Riback: I’ve had the privilege, the benefit of hearing you speak on this previously, and then when I researched more and read more, it was just extraordinary to me the role and impact that the Coleman report played and for how long that has been taken and as you just said, influenced public policy and yet the conclusions that get drawn from it seem to be exactly what you described and now you took a different approach and came out with a different result and potentially different public policy prescriptions that can come out of your work. So that really kind of blew me away and is one of the reasons why I wanted to ask you about the Coleman report.
Kirabo Jackson: Another thing I just sort of mentioned is that the Coleman report was written in 1966 and there was not a strong sense at that time amongst researchers and policy makers that correlation and causation were not necessarily the same thing. Now, of course, people were aware that correlation does not imply causation. But I think that hadn’t been hammered away in the ’60s as the way that it is today. Today almost no piece of research that is trying to make a policy claim would be based on just raw correlations. Is that in essence, if you want to make a call to a claim about what would happen if you improve school quality or if you reduce class size.
Today the standard of evidence would be that you would have to actually manipulate school quality or manipulate school spending or manipulate class size and see what happens. That is to say if you want to answer policy questions, you need to actually analyze actual policies and that’s essentially the way that we would interpret research today. So most of the research today that analyzes actual policies or manipulations of these school inputs find that they’re actually very very effective at improving student outcomes, which is different from what you would get if you just look at raw correlations.
Chris Riback: So let’s look at some of the results. One of them that the variables and the estimates indicate that for poor children increases in Head Start spending and increases in public K-12 spending each individually increased educational attainment and earnings and reduced the likelihood of both poverty and incarceration in adulthood. So explain that because the first step I guess was to look at the effects of investing in Head Start and then, am I interpreting this properly? It separately investing in public K-12. And you came up with conclusions on both of those and then I guess we will look at it in a second what happens if one is or an area is investing consistently throughout those two areas. But that was the first set of findings, if I’m understanding correctly.
If you roll out Head Start, does the effect of rolling out Head Start have a bigger impact in areas that previously had a school finance reform and had higher levels of K-12 spending? The answer there… is that the effects of both together is much larger than the sum of the individual effects. So you basically have a synergistic effect.”
– Prof. Kirabo Jackson
Kirabo Jackson: That is correct. So what we do, just to give you an overview of how we did that was essentially what I just sort of described. We compare the outcomes of children who were of born in the same county. And we rely on the fact that Head Start as a program started in 1965. So for individuals who were say born in 1964, when they were four years old, Head Start may have been available in their county of birth, but for others it would not have been. And for those who are born say in 1958, they would have been sort of six years old at the time and they would not have been exposed. So what we do is we look at the basically outcomes of children who were exposed to different levels of Head Start spending on when they were four years old.
And we essentially document that being exposed to higher levels of Head Start spending in one’s county when you’re four years old, it was associated with improved educational attainment, wages, lower levels of poverty and lower levels of incarceration as an adult. When we do a separate sort of analysis looking at what happens to children when they’re exposed to higher levels of public school K-12, spending using a kind of a similar design. I won’t go into too much detail, but many states basically underwent school finance reforms where in the existing ways in which schools were funded had to be revamped and rewritten. It was deemed to be unconstitutional by state constitutions and states had to rewrite the way in which schools were funded. Funds were allocated to schools to be more equitable.
So districts that were previously low spending districts or districts that were previously low income districts were likely to get large infusions of cash after the passage of one of these court order reforms and others were not. And what we do there is we say, well, we can look at children who were basically between the ages of 5 and 17, school going age when one of these reforms happened in their state. If you were in a district that was previously low spending, school spending levels went up a lot for you. If you were one that was high spending, school spending levels didn’t change very much. And we can basically compare the outcomes of children from the same school districts some of who were in school when this money basically was infused into their public schools as some of whom would have been too old to have benefited from that.
And we document that those children who are between the ages of 5 and 17 when money was invested in the public schools had much better outcomes than those who were say 17, 18 or 19 when this happened. And again, we see improvements in years of educational attainments, wages, incarceration and also being poor later on. And then what we do is we sort of combined these two things and we say, well, there are some areas where they had increases in K-12 spending but there was no Head Start. And there were places that had Head Start, but there was no school finance reform that changed a K-12 spending and then there were some places that had both. And what we basically were able to do because we have sort of these different kinds of districts or locations, we can ask the question, is it the case that getting more money in the K-12 system for your public schools, does that lead to larger increases and outcomes when there’s Head Start present and when there’s not.
Or there’s some places that have Head Start and not. And we basically say if you roll out Head Start, does the effect of rolling out Head Start have a bigger impact in areas that previously had a school finance reform and had higher levels of K-12 spending. And the answer there was basically is that the effects of both together is much larger than the sum of the individual effects. So you basically have a synergistic effect.
Chris Riback: Clarify that because that’s the key point, that’s the payoff point. You’ll forgive me for butchering the math problem, but that was one plus one equals three. That’s how I read it.
Kirabo Jackson: No, that’s exactly right.
Chris Riback: You can tell, I got neither Head Start nor a well funded public education so as far as I know, but that’s what I came away with. Was my interpretation correct?
Kirabo Jackson: That is exactly right. If you basically increase spending on Head Start on its own, you get a modest effect. If you increase spending in the K-12 system, you get a modest effect. But if you do both at the same time, the effect is really really large. And so to use your analogy, if you spend $1 on Head Start, you can get a certain impact. You spend $1 in the K-12 system, you’re getting at a certain impact. But the impact of a dollar of Head Start and a dollar into the K-12 system is bigger than two. That’s exactly right. So the implications are basically that every dollar that we spend in the K-12 system is actually more effective at improving, it becomes more efficient.
You get more bang for your K-12 dollars if you’re also investing in Head Start and vice versa. You get more of a bang for your Head Start dollars if you’re also investing in the K-12 system. So the policy implication is that both of these things need to be done together and we shouldn’t think of them as being separate and that the gains that we observe in the Head Start system will be much much larger if we ensure that the kids who are exposed to Head Start are sort of followed up with subsequent investments in their skills.
Chris Riback: The efficiency argument is so powerful to me because I feel like that’s a nuance, shouldn’t be a nuance, but is an aspect that frequently gets missed. The questions are how much is being spent and what were the results? Usually based on test scores as you identified and that’s not a necessarily great outcome or a great measurement. And the question of efficiency to me feels exceptionally important. So you come out with the study, what was your among public officials and politicians?
Kirabo Jackson: I think as in many things there were some people who were sympathetic to the basic findings and they found it very important. I think people certainly in the early childhood world find the results very compelling. It does help explain a lot of the patterns that we see where the fade out like I described before or the fact that the gains tend to be larger for the more advantaged poor population. So I think people generally found them to be pretty compelling. As you probably are aware, there are some people who are sort of education spending skeptics in general. I think we were able to convince a few of them that maybe money does matter in some cases.
There’s still work to do in that front, but I think one of the things that we’ve tried to do for this paper and other research on school spending is to really emphasize that it’s important to decouple the questions of whether money matters from questions of whether it matters as much as they could. And once you sort of separate those two questions and so certainly how we spend money absolutely matters. And one of the things we’re documenting here is that, yeah, how it’s spent matters. So if you distribute your money between early childhood and K-12, you’re getting a better outcome than if you spent all of it on only one of them.
And that’s kind of one of the results of this or I should say, obviously you wouldn’t spend all your money in K-12, all your money in Head Start. But having a more balanced distribution towards Head Start and K-12 gives you a better average outcome than if you overspent in Head Start and took money away from Head Start and spent on K-12 or vice versa. So our paper is both documenting that spending matters, but also at the same time documenting that how it’s spent also matters for what the outcomes are going to be. So I think it’s been relatively well received because it sort of speaks to, I think, issues that people from both sides of the political spectrum could sort of get on board with.
Chris Riback: Relatively well received is a modest understatement obviously, but anyhow that’s very modest of you to put it. So in reading this and reading other aspects of your research, I feel almost like you could write a book and you can have this title. This is free gratis for you, Everything You Think You Know About Education Funding Is Wrong. Now, you’re an economist by training. Obviously you apply your study and your trade and your discipline towards public policy, particularly as we’ve discussed in the education front. This may be an unfair question to ask you, but why is this such a misunderstood education and investing in and why everything you think you know about education funding is wrong. Why is this such a misunderstood area of public policy?
Kirabo Jackson: That is an excellent question. So I’m actually in the process of writing a meta analysis right now. Actually going back and looking at some of the older research on school spending, the old writings on school spending and looking at some of the new findings on school spending. So one of the things I should mention, I think I sort of alluded to this before, there’s been kind of a shift. It’s referred to often as the credibility revolution in economics in general and sort of hip this question. So there are a lot of studies that have come out in the past five years on per pupil spending and those are very compelling using very credible experimental or quasi experimental designs and showing that school spending does matter.
That includes some of the work I’ve talked about today and others by some other researchers and there’s a whole earlier literature that looked at this. And I went back to it and sort of read through the findings and one of the things that jumped out at me which was kind of remarkable was that, in fact, if you look at the old studies, they actually don’t support the notion that spending doesn’t matter. But it turns out that for whatever reason, there were some very influential scholars who were basically interpreting the data in that way. I could summarize how I came to that conclusion as follows. So if you have a whole bunch of studies out there and you look at, say we have 100 studies and you want to know does school spending matter?
One way to look at it would be to say, well, if school spending matters, then we should see that every single study on the topic should find that school spending has a positive impact. Now clearly that’s not exactly right ’cause for random reasons, for statistical reasons, some studies may find a positive association, some may not, some may find zero. So then you might say, well, what is the threshold? So clearly it’s not every study has to find the positive impact. And what often had been done previously, particularly by the person who was very prominent, was he basically looked and said, well, he wanted more than half roughly of the studies to find a positive impact. But it turns out that’s actually not the right way to think about it.
If you have 100 studies and 20 of them find positive impacts and none of them find negative impacts, that’s actually pretty compelling evidence of a positive impact on average. The reason is that in order to even find a positive impact, it needs to pass a statistical test threshold, which is difficult to do. And most of the time, most studies don’t pass that if there’s no impact. So even 20% of studies finding positive impacts, suggest that it’s a real impact. So it turns out, I want to get too much into the weeds here, but even the older literature if you look at it rigorously through a statistical lens supports the notion that schools spending matters. But for whatever reason, I think scholars have looked at it, interpreted it without thinking about it sort of statistically and say, well, if half of the studies find positive but half find nothing, then there’s nothing systematic.
And I think that’s basically just the wrong way to do it. And part of it is I think people who have a particular political view point are sometimes likely to accept evidence that is consistent with their worldview and that’s sort of what happened.
Chris Riback: I think I understood what you’re saying and more broadly it does feel like where you sit depends on where you stand in education and around interpretations of education spending and impact and public policy. And it does feel like an area where people try to take results and have them get interpreted to support their preexisting view or life view as opposed to going where the data lead. At least that’s how it’s feeling and that’s kind of what I’m taking. I would like to amend one thing I said previously, so when you do this meta analysis, that’s going to be huge. So I take back my free gratis offer for the title or the subtitle. I just want a small percentage, that’s all. I mean we can negotiate details later, but that is. That sounds like everything you think you know about education funding is wrong and that’s incredibly important. Kirabo, thank you. Thank you for your time, and thank you for the work you’re doing.
Kirabo Jackson: Well, thank you so much for the kind words and thanks for having me.