The Swedish school system is in a severe crisis. After decades of decline in PISA scores Sweden scored 28 out of 34 countries in the PISA result of 2012, this although the School spending is among the highest in the world. With last years exceptional levels of refugee immigration in Sweden, the school system is facing immense challenges with many new students coming into the system, teacher shortages, as well as need to support many refugees with poor schooling background.
A highly educated workforce is at the same time becoming more important than ever. An advanced economy, driven by globalization and automation makes the prospects of the unskilled workers bleak. As usual Warren Buffet puts it well:
Recently, however, the economic rewards flowing to people with specialized talents have grown dramatically faster than those going to equally decent men and women possessing more commonplace skills.This widening gap is an inevitable consequence of an advanced market-based economy. It is simply a consequence of an economic engine that constantly requires more high-order talents while reducing the need for commodity-like tasks.The remedy usually proposed for this mismatch is education. Indeed, a top-notch school system available to all is hugely important. But even with the finest educational system in the world, a significant portion of the population will continue, in a nation of great abundance, to earn no more than a bare subsistence.
Today I will focus on the top notch school system and come back to the inequality problem some other time. With a soon 1 year old child I am quite worried of what lies ahead in education and as a consequence, in life opportunities.
The problem with Swedish School system
"The quality of an education system cannot get better than it's teachers"
A highly cited McKinsey report from 2007 puts forward the following three conclusions for a well functioning school system:
- Getting the right people to become teachers
- Developing them into effective instructors
- Ensuring that the system is able to deliver the best possible instructions for every child
Several studies shows that class size does not matter, it is the quality of teaching that matter and the major driver of quality of teaching is of course the teacher.
Problem 1 - Getting the right people to become teachers
The graph below illustrates how the top 20% of teacher's impact on average student performance relative to a bottom 20%. The impact of teacher quality is huge
In Sweden the status of the teaching profession is very low. Anyone in Sweden can become a teacher and although we have a teacher shortage the available seats at the University does not fill up. The admission scores of standard tests allows for teachers being admitted with score of bottom 0,2% of test takers, a score well below what a monkey would produce. However, last year Sweden the news media bolstered of an improvement with 5% of the admitted to teaching studies being high performing students - very poor indeed. Of all high performing students from high school I know, no-one choose to become a teacher.
Compare this to the admission standards in Europe's highest performing school system, Finland:
Problem 2 - Teaching instructions
The Swedish school system used to rely on traditional teaching methods. In recent decades, modern “individualist” or “progressive” pedagogic ideas took hold. The idea is that pupils should not be forced to learn using external incentives such as grades, and children should take responsibility for their own learning, driven by internal motivation. Rote memorization and repetition are viewed as old-fashioned relics. Teacher-led lectures have increasingly been replaced by group work and “research projects.”
Solution - how would I solve this if I were running the thing
So how do we solve this? We have a teaching population of poor quality and at the same time a shortage of teachers, even if the Universities fill all seats with teaching students in the coming year. At the same time, like it or not, we need to remove many poor quality teachers and replace them with high quality teachers. Further, we want to secure behavior that maximizes teacher performance.
- Open up teaching for all high achieving academics
Drastic times call for drastic measures. We have had 25-30 years of no competition entering University to become a teacher and the only way to solve this is to move the competition to the teaching profession itself. There is no other way to fast and significantly turn the system around with many pupils coming into the school system and poor quality a teacher stock that needs replacing.
It will take too long to turn the system around even if all students admitted to the 5 year "teacher programme" next semester are of "Finnish Quality". It is 5 years lead-time to get the first quality batch out there and the replacement of existing poor teacher population will still not happen given projected future under-capacity.
Therefore, the only way forward is to lower the barriers to entry for top talent in the teaching professioin. I propose the following requirements for teaching license:
- All individuals with a master degree, who finished say top 10% of their class
- They need to pass some kind of pedagogical test
- Benchmark high performance selection systems, e.g. management consulting firms with 5 interviews looking at in field testing.
- Have absolute performance hurdles as well as relative performance hurdles (e.g. Only top 5-10% of test takers) to be granted teaching license
2. Independent Administrated National Tests
Darren Gee at Peyto is one of the CEO I am most impressed today. Among many other good practices, they contract for an independent testing of their drilling performance each year where the evaluate projects reserves, and future production. The money spent and a price deck gives them the projected cashflow and return on capital (high hurdles given uncertainty of commodity prices). The score card and related incentives contribute to the organizations laser like focus on what matters in business, return on capital.
Sweden should quickly establish independent testing and evaluation, which are base for student grading, for three reasons:
- Huge grade inflation
- We have had a privatization reform that lead to competition among schools in attracting students. One of the major selection criteria is of course the grades students leave the school with. This have led to incredible perverse incentives, where the School's performance (and often survival) are dependent on the student grades, which the School are setting themselves. In Sweden we therefore have had a quite long period with improving grades for students where all independent testing (e.g. PISA) shows that student performance is rapidly declining
- Independent tests are a great way to measure, and incentivize, teacher performance
- How to you measure teacher performance? Are you a good teacher if your students score top 20% in the Nation? Well it depends, if they scored top 99% last year you are a lousy teacher and if they scored bottom 10% you are a great one. Clearly, the best measurement of performance for a teacher is the change in relative performance of the student. How the students measure up from the test last year to this year is how much value you add as a teacher relative other teachers.
- By standardizing the tests and evaluation we can make the processes more efficient
We need to make sure our testing is fool proof from cheating and corruption before looking at the next policy proposal:
3. Teacher incentives - Jack Welch philosophy put to the extreme
As we have a very good way to objectively test performance as a teacher, which is the relative performance of the students during the year - we should incentivise heavily. Given the huge importance of a highly educated workforce Sweden should throw some serious money into this. The money should go to:
- Rewarding the high performing teachers, the best ones should get rich
- Severance packages for low performing teachers
- Scholarships for top talents to study for teacher at the University
Calculation - High level input numbers:
- Total number of teachers estimated to 110 000
- According to skolverket (Swedish School age 7-15 we have 90 000 teachers and 1 million pupils (I´ll come back to the efficiency potential). In high school we have some 300 000 pupils, cannot find teacher number but would guesstimate it to be 20 000.
- Some teachers excluded from whole scheme
- I would exclude subjects teachers such as "household knowledge (cooking)", sports, crafts (e.g. sewing), music and some special needs teachers (e.g. "home language teachers")
- For ease of calculation say we are left with 100 000 teachers
Incentive scheme:
| Relative teacher performance | Performance pay | Comment | Total Cost |
| 0-10% | 1 MSEK | Severance package based on time in service | 9 Billion SEK |
| 10-20% | 0 | 0 | |
| 20-30% | 0 | 0 | |
| 30-40% | 0 | 0 | |
| 50-60% | 0 | 0 | |
| 60-70% | 100 KSEK | 1 BSEK | |
| 70-80% | 200 KSEK | 2 BSEK | |
| 80-90% | 400 KSEK | 4 BSEK | |
| 90-99% | 1 MSEK | 9 BSEK | |
| 99%+ | 5 MSEK | 5 BSEK | |
| Total | 30 BSEK |
Some thoughts on how to form the system:
- Need to be really careful of course in forming the system, looking into second and third order effects
- Of course it is much harder to move a class from the 90 to the 99,9 percentile than moving a class from 40 to 50, so you should build in that into the system.
- Also we should take absolute performance into account as well. If a class that last year scored in the 99 percentile and end up with the 99% the year after it is a good teaching performance. However, you would want to skew the system somewhat to ensure that "problem classes/schools" far away from potential end up with good teachers.
- We should bias the system giving a teacher with large classes higher score than ones with small classes for efficiency. Also, we should have a ranking system in each geographical area, sending students to the best teachers until they do not accept more. This would take out the teachers of poor quality from the system.
- If class size really does not matter this will make the class size drift towards larger classes for great teachers. There is quite a runway here with Swedens average class size is 18 compared with an average of 32 world leading school systems in Japan and South Korea
Think about what behavior this would drive. I would probably apply to become a teacher and take the test myself, in competition with many other talented people in our society. Hopefully we would also get some high skilled Finish teachers moving to Sweden (many speak Swedish). How would I and other teachers behave differently?
- I would think every day on how I can improve the teaching and pupil learning. I would seek out and implement best practices like a religion
- I would rally the parents to support their children in all way possible
- I would put in long hours to improve performance, for instance offering voluntary study sessions during the year and especially in the period before the test
- I would spend much less time on the teaching approach that got us here (freedom, creativity and group work) and spend time on what works. You would get back the discipline and teacher authority that the system lacks today.
4. Scholarships for top students admitted to become a teacher
The incentives to become a teacher is not there today. Average pay is perhaps 25-30K SEK per month with no upside, slightly higher than an unskilled industrial worker. For this the teacher needs 5 years of University education (should be 3-4 years) and take on student debt of perhaps 250k SEK.
Pay in Finland is not substantially higher than in Sweden, but they still manage to attract many high talent students to the teacher profession. This really shows what a competitive advantage it can be to have high status workplace, attracting much higher quality individuals at similar cost than a comparable workplaces. The status of the teaching profession in Sweden is gone and the only way to higher status is to up the incentives.
Although the (political impossible) incentive scheme above would attract increase attractiveness the ultimate no-brainier is teaching scholarships for top talent.
| Top high school students in Science and Civics | Scholarship | Comment |
| National top 2% of students | 450 K | Student loan fully paid for and a 100k grant each at admission (repayable if not compleated) and graduation |
| National top 10% of students | 250 K | Student loan fully paid for |
3 billion SEK to this purpose and you would fill the Universities talented teachers. There are 10k students admitted per year, which comes down quite a lot as you only include subjects of high value in the knowledge economy to this scholarship (e.g. Science, Civics, Language)
Business case for this is really easy. Think about how many students a teacher have during the career, calculate the cost the of scholarship per student. Then go back and look at the graph on high quality teacher versus a low quality teacher and estimate what the value difference is for society (in future taxes and so on). This should not be done in Finland, but is the ultimate no brainier in Sweden today
Funding all this
It would require some 30 billion SEK to just get the incentive systems in place. Then we have the independent testing system, It should not be that expensive to get in place as it on balance should remove a lot of work from the teachers and should be more efficient to do centrally. Even if no savings were found I would still do it, I cannot think of a better way for a country to invest money in today's global knowledge economy than to fix a broken school system and secure that the children gets world class education.
But I think there is a lot we can do on the cost side here as well, that the proposed system should drive out naturally. Sweden's average class size is 18 compared with an average of 32 world leading school systems in Japan and South Korea. If we would go to a class size of 25 we would be able to reduce number of teachers with 28% (poor quality teachers) which alone almost could fund the initiative.
The biggest lever though is that I am convinced that there are a lot of non-student facing staff working in Swedish school system, where we can find a lot of money, The Swedish School Administration (Skolverket) have on there website a figure of 90000 teachers for 1 million pupils in the "age 7-15" school system today, an average "class size" of 11,1. This does not add up and all non student facing staff should be cut to the bone, which of course does not happen in government institutions. Although you could just remove or automate a lot of activities there are no incentives for it and people just sit there and invent stuff to do.
The system we must have is to get the right teachers in the classrooms, measure and incentivise effectively and leave them be to run the show.
Final Word
I leave the final words to Charlie Munger on what is wrong with the Swedish School System:
"I think I've been in the top 5% of my age cohort all my life in understanding the power of incentives, and all my life I've underestimated it. Never a year passes that I don't get some surprise that pushes my limit a little farther."
"The most important rule in management “Get the incentives right”‘
“Never think about something else when you should be thinking about the power of incentives.”
Inga kommentarer:
Skicka en kommentar