That is the question asked by Scott Alexander and John Cochrane in discussing high school education, college and infrastructure spending. Despite rising funding, it is not clear outcomes are improving.

Scott highlights the example of K-through-12 public education where spending has increased substantially since 1970 but test scores have remained stagnant. He asks:

Which would you prefer? Sending your child to a 2016 school? Or sending your child to a 1975 school, and getting a check for $5,000 every year?

On college he presents a similar counterfactual:

Would you rather graduate from a modern college, or graduate from a college more like the one your parents went to, plus get a check for $72,000?  (or, more realistically, have $72,000 less in student loans to pay off)

He also highlights the rising cost of infrastructure spending through the example of a New York City subway:

1900…it’s about the inflation-adjusted equivalent of $100 million/kilometer today… In contrast…a new New York subway line being opened this year costs about $2.2 billion per kilometer

As Scott outlines, the underlying crisis here is made all the worse by the fact that new technologies and globalization should have put downward pressure on the costs of provision.

Two questions arise: why is this happening and what can be done about it?

This requires a huge amount of research. Certainly it cannot be answered in a blog post. But I want to suggest an analytical framework for thinking about these examples that can be applied in each case to work out what is going wrong. This is all the more necessary because the absence of meaningful prices in the public sector makes measuring productivity much more difficult than in the full market sector of the economy.

Rather than merely comparing money spent to outcomes, we can break things down as follows:

Taxpayer dollars -> Inputs -> Production process -> Outputs -> Outcomes (quality-adjusted outputs)

Take schooling. We pay money in through taxes.  These are used to fund the labor (teachers, administrators etc), to build schools, and to pay for the goods and services used within schools. The schools then operate. And those inputs work to produce measurable outputs in terms of number of children being taught, hours of teaching, exams prepared for etc. But what we really care about is outcomes, which are linked to but not quite the same thing (think test scores). This is best thought of as a measure of quality-adjusted output. Productivity (to the extent we can measure it) can be thought of as the ratio of outputs to inputs, whereas what we ultimately care about here is improving the effectiveness of money spent (outcomes over taxpayer dollars).

This framework allows us to posit different theses (which are not mutually exclusive) for why taxpayer dollars have gone up but outcomes stagnated, which we can test empirically:

1. The cost of inputs, such as teachers etc, may have risen substantially
2. The number of inputs necessary may have risen (more administrators and other inputs not previously needed)
3. Productivity (the ability to turn inputs into outputs) may have fallen
4. Quality of outputs may have declined (e.g. children’s lessons are now worse than they were before)

My hunch is that there are probably a lot of people doing things in education and infrastructure preparation that have added to the number of inputs necessary but which do little to affect the quality-adjusted outputs we care about directly (think a lot of environmental audits and reports, compliance with regulation etc.)

But before jumping to conclusions, we should really try to measure outputs and inputs directly. In the UK, it was historically assumed that public service outputs were the same as public service inputs (implying stagnant productivity). But in recent years the Office for National Statistics there has put in a lot of effort to try to measure the quality and quantity of public service outputs, albeit imperfectly. It has actually proven very useful. They have produced interesting work which found public service productivity improved in each of the first four years of so-called “austerity,” for example.

Unless I have missed it entirely, similar indices are not currently constructed here. But if we really want to get to the bottom of why taxpayer funding is not producing better outcomes, we need to shine a light on the public sector production process to see where things are breaking down.