Matt's Blog

A Most Stupendous & Audacious Undertaking

Previous Entry Share Next Entry
Cost/benefit analysis of teacher salaries... good teaching is worth an extra $390,000/classroom/yr
There's been some buzz in the media about our failing educational system lately.   I thought I'd do a little analysis of the benefits of education. 

Suppose that a good vs mediocre K-12 education increases your annual expected earnings (in 2011 dollars) by $15,000/yr.  I think it's probably more than that, but I'm being conservative.  If you work for 45 years, that's an extra $675,000.  Suppose the government gets 25% of this extra salary -- $169,000.  For this to be a net increase in revenue for the government, they'd need to spend less than an additional $13,000 per year per student on the 13 years of government-funded education.  If you have a class with 30 students, that's an additional $390,000 per classroom.

As long as the school system spends less than *an additional* $390,000 per classroom per year to raise student earning potential by $15,000 a year, the government will financially come out ahead on the deal.  

So why isn't teacher accreditation as difficult as medical school, and why don't teachers have starting salaries of $120,000/yr?

By the same logic, programs for gifted students should receive even more funding per person, as those students (on average) have the potential to generate the government even more revenue.  This may sound cold, but I'm trying to look at this as a business opportunity since proving financial benefit will often convince more people than abstract ideals of the pursuit of excellence.  The math doesn't lie -- kids with intelligence and drive should be given the best education money can buy. 


I should say for the record that I'm not a fan of teachers' unions in their current form.  In my direct experience as a high school student, the unions often protected mediocrity at the expense of excellence.  Bad teachers were shuffled around from district to district beause they could not be fired for anything short of sexual misconduct.   Good teachers were not substantially honored or rewarded.

  • 1
(Deleted comment)
They are usually funded at the state level. But local communities can levy extra taxes to fund their schools. Typically the taxes can not go into the schools general fund but must be use for a specific purpose: buildings, computers, arts.... But this varies state by state.

One thing I've heard a number of times (though I can't cite any good references) is that throwing more money at education doesn't consistently improve education results. The problem isn't just making more money available, and that might not even be one of the biggest problems.

Your example of extremely tough teacher accreditation (presumably with ongoing re-accreditation) seems like it would make a big difference, and much higher salaries would be necessary to implement something like that, but I suspect that the cultural and institutional changes required would be far more difficult than finding the money.

OTOH, saying "spending more on education doesn't work" is also a great excuse if you just don't want to spend money on education, so I try to remain a bit skeptical of such claims. I'd like to see some good studies...

The problem with tough teacher accreditation (and re-accreditation) is that a bad teacher is generally better than no teacher. It's only worth making it tougher to be a teacher if you first get to the supply-demand point where supply of people who want to be teachers significantly outstrips the optimal number of teachers. So what you first need to do is increase the pay and prestige of teaching to get enough people wanting to be teachers, that you can then select only the best and brightest. But perhaps that's what you mean by "much higher salaries would be necessary to implement something like that" and "cultural and institutional changes".

Disagree about extra funding for gifted ed. Intelligence is in no way a condition for success or productivity. I actually think much more money should be thrown at the lower end of the spectrum. It would yield far better results socially and economically.

Studies have shown that hard work and plain old graft has much more effect on a person's ability to do well than intelligence.

For his actual calculations, what we'd need to know is what the marginal effect on future productivity of improving a teacher for gifted education is, versus the marginal effect on future productivity of improving a teacher for standard, or special education. Matt seems to presuppose that the former is substantially larger than the latter, but it's certainly not immediately obvious, and I think I've heard of research (that you seem to refer to) that suggests that in fact gifted students benefit less from improved instruction than ordinary or special ed students.

I agree with the last point, but if you don't encourage gifted students they end up coasting through high school without having to try. Then they get to University and everything changes.

I may be bitter because despite being one of the top students at my school, all it taught me was how to be successful without applying myself. It took me several years at university to unteach those habits.

In my view, the issue you're discussing is probably as much a socialization issue as anything. By way of background, I teach at college level and I have taught a very wide range of courses at graduate level, basic university 101 level, and remedial entrance level. I also do a lot of military education. Personally, I went to a private gifted elementary school and then was in the gifted program for the rest of my public school career. I also attended an exclusive honors college. My disabled sister, who has an IQ of about 70, went through a number of different educational programs ranging from special ed to those for the Educably Mentally Retarded. She cannot balance a checkbook, but can kick anyone's ass in checkers.

While I do believe in some form of differed ability curricula, most smart people I know, at every age, would do best with a good kick in the pants and some social skills. My sister's educational needs were far, far greater than mine, and frankly those at the middle and bottom of the ability spectrum are very poorly served educationally throughout their entire lives and as a result need more services. As an educator, I would MUCH rather have money thrown at the bottom end than the top. Economically it makes far more sense.

As posted in the other comments, I agree about spending money on kicks in the pants and social skills.

You do bring up a good point about spending more money on people on the bottom end of the spectrum to reduce a later need for social services. OK, I'm up for throwing extra money at both ends.

I agree with you about intelligence not being the sole condition for success. I know plenty of intelligent people who lack follow-through and get by because they're very productive during rare moments of focus or because they have enough knowledge to bluff their way through things. Part of what I'd do in the gifted program is work hard on enhancing study skills, willpower, focus, and drive. Lots of very academically intelligent people lack people skills as well, and that's worth focusing on. It's kind of strange that in first grade, you learn math from the teacher, but your social skills from other first-graders.

I'm also advocating throwing huge amounts of money at the lower end of the spectrum too. IIRC per-student funding in public schools is under $10,000/yr, and I'd be proposing almost doubling that.

I'm still wondering what your overall goal is with this. It sounds like you're simply interested in making people just better cogs in the system, which doesn't make a lot of sense from a holistic economic or societal perspective. What do you mean by "productive" and why are you wanting to measure it in income and tax generation? Higher income also does not suggest higher taxes paid either--for all sorts of reasons. Are you interested in making stronger economic systems with fewer resource drains? If so there are far more effective educational optimizing strategies than focusing on gifted ed.

If Supermarkets Were Like Public Schools


"Suppose that groceries were supplied in the same way as K-12 education. Residents of each county would pay taxes on their properties. Nearly half of those tax revenues would then be spent by government officials to build and operate supermarkets. Each family would be assigned to a particular supermarket according to its home address. And each family would get its weekly allotment of groceries—"for free"—from its neighborhood public supermarket.

No family would be permitted to get groceries from a public supermarket outside of its district. Fortunately, though, thanks to a Supreme Court decision, families would be free to shop at private supermarkets that charge directly for the groceries they offer. Private-supermarket families, however, would receive no reductions in their property taxes."

See also "Suppose there were food insurance"

A few modifications I would propose to your calculations.

First, for your calculations, it seems that what matters is not the increased earnings of one individual, but the net increased productivity of society. To see this, imagine a world where there are 100 jobs, 50 of them paying $10,000 and 50 paying $100,000 a year. If this setup isn't changed by giving someone a better education, but you can make sure that they get one of the high-paying jobs by increasing their education, then the government will get no extra tax revenue from the increased education, even though the individual will get substantially increased income. (One other person will suffer the negative externality of being pushed out of the high-income job.) Presumably, the salary of individuals is often determined by competition of this sort, so that increased pay for an individual doesn't entirely accrue to increased overall earnings, and your initial calculation would provide an overestimate of increase in tax revenues.

The only way that individual education can increase government revenue is if it increases the total value of future salaries, and doesn't just transfer those salaries from someone else to the educated individual.

But of course, that's exactly what we do expect - educated individuals innovate and increase productivity not just for themselves but for others. (Lots of that increased productivity results in other people temporarily being out of work, but of course we're all better off now because there were a few decades of skilled farmers, weavers, and others whose jobs were displaced by a small number of people overseeing machines.) This value is very hard to calculate though. This is especially true for things like media and communications, where we don't actually seem to see increased productivity in any traditional sense that comes from the massive increase of computing power over the past few decades. (For comparison: the massive improvement in quality of life resulting from the scientific revolution wasn't triggered by the printing press or the development of modern physics and astronomy, but rather was triggered by the eventual application of these things to textiles and agriculture, which was 200-300 years later. Does that mean the printing press and Newtonian physics had no economic value?)

I really don't think that education or even intelligence is necessarily the key factor in landing a high paying job. Also, in some fields, like computer science, getting a PhD will actually decrease your earning capacity over the course of your career. So what do we mean by education here? I know a crap load of PhDs who would have been far better off becoming mechanics or hairdressers. These generalizations sound nifty, but in actuality they translate into very little for those who are interested in policy outcomes.

Besides, most people know that actual *wealth* isn't earned through wages anyway. Maybe we just need more courses in investing.

I agree - my main point is that the salary of the job is not the relevant issue - what really matters is how much value one creates for the world through one's activities. Whether an academic research computer scientist creates more or less value for the world than Bill Gates, or the average programmer at Google, or the inventor of Farmville, is a very difficult question. (You can probably make fairly good arguments that any one of these four creates the most value, and also fairly good arguments that any one of these four creates the least value. A lot will turn on how you measure value.)

But yes, a lot of this depends also on what we mean by education.

And of course, the reason I see these questions that way is largely why I'm working in philosophy - I'm really interested in these questions of what we do and should mean by "value", "education", "information", and so on.

"Value" according to whom? The inventor of Farmville has created no value for *me* whatsoever through that act alone, but as we know it's not that simple. I also think that these glib emphases on "productivity" can also go horribly wrong. We just cannot reduce people to simple economic actors (even if we did really know how economies work, and they, too seem to be situational). Many of these things just cannot be measured.

It isn't just that the Social Darwinist implications of nasu_dengaku's suggestion are "cold", I just don't think the suggestion holds up to many important questions.

As far as PhDs go, I think that while the PhD is the right choice for some people, far too many people choose PhDs because it's just the next step in a familiar system. For that matter, I think a lot of people choose college because it's the next step in a familiar system, and they'd get more out of college if they took a year or two off first.

Having interviewed many CS PhDs for various positions, I can see why their starting salaries are often lower -- they've spent so much time focused on a particular problem that sometimes their intellectual toolbox and worldview are slanted. This isn't always the case but I have seen it occur many times.

There are a lot of places that won't look at your resume if you don't have a college degree. I personally think this is rather silly as I know many highly qualified people without them.

Clearly I'm enjoying this conversation. I don't mean to be busting your chops ;)

I think there are a number of reasons people get PhDs, and most of them are bad. Some want the prestige of being an academic (ho ho ho ho ho), some have been sold the idea that it will be of more net worth to them salary wise. Some use grad school to avoid the real world--they want to keep thinking grand thoughts. The reason that CS PhDs have a lower salary point is because CS academia is not as dynamic as industry, for one, so the time you spend doing your PhD you may actually be losing valuable skills. For another, you just don't need a PhD to get a job in CS, so if you do enter industry with a PhD, you have lost earning time.

I still don't think college is a waste of time, not at all, but it depends on what the goal is.

Another correction - even though you adjust for inflation by saying the increase in salary is $15,000/year in 2011 dollars, you don't include a discount rate. Discounting the future is important not just because of inflation, but for a variety of other reasons. For one, there's a lot of uncertainty about the future, which increases as you go farther. For another, if the government wants this to be a net revenue increase, then it needs to consider the opportunity cost of spending that $390,000 on something else right now, or the interest cost of borrowing that money to be paid back over the 45 years starting with the student's entry into the workforce.

So if you imagine $390,000 total being paid back over 45 years, starting 6 years in the future, we can then calculate (for a given interest rate) what total amount can be borrowed - that's the amount we can increase spending in a classroom. (I would guess that the end result is going to be on the order of half the amount you predicted, which is of course still a rather large amount.)

(Deleted comment)
Wait... I thought you're supposed to put limbs in a woodchipper?

(Deleted comment)
So why isn't teacher accreditation as difficult as medical school, and why don't teachers have starting salaries of $120,000/yr?

I believe it's because it's a job associated with women. The statistics are pretty clear: the more women in a profession, the lower the pay, and when a profession begins to get more women, the pay rate goes down.

I'm of the opinion that it is very much a direct result of sexism.

We should have a conversation about this and other economic issues sometime. Maybe we should try to get dinner sometime in the next few weeks?

Sounds good. You know my email address, right?

Edited at 2011-05-10 05:50 pm (UTC)

I want very much to respond to your post in-depth, but I can't because no matter how concise I attempt to make my arguments, I end up writing full sermons on the subjects of society, culture, and education.

To state my piece briefly: the problem with America and education isn't one of money, it's one of culture. Until we live in a culture that universally understands that knowledge is power, education is the appropriate discipline for attaining knowledge, and those who assist us with our discipline in acquisition of knowledge deserve our respect, no amount of money thrown at our educational system will actually make it work.

I agree it's a problem of culture. America's been top dog for so long that people don't take striving for achievement seriously and instead simply feel entitled. Countries with something to prove, on the other hand (eg India) are pushing hard to improve themselves, and value education far more. I think we need to have other countries kick our ass more visibly for change to occur.

Your point reminds me of another strange thing I noticed -- in rich neighborhoods, the price of a tutor can be astronomical. I heard of a math grad student making $500 an hour in Manhattan. If parents want to spend extra money on education, why are they pouring it all into tutors instead of spending more on tuition and teachers' salaries?

I think the reason is that it's very hard to find the best teachers and pay them a lot more.

I wonder if we'll move towards a new equilibrium where people who can afford it get more and more of their education from private tutors (as aristocratic children used to) because it's easier to determine the quality of an individual tutor.

Interesting. This does make me wonder how well students at the best private schools in NY compare (in, say, SAT scores) to students from similar socioeconomic backgrounds at public schools.

Private school tuition is high enough that the schools should presumably be able to pay teachers the salaries I'm talking about. Whether they actually do is another matter.

Also, welcome to my blog!

  • 1

Log in

No account? Create an account