Monday 12 June 2017

Public Service Productivity

The Productivity Commission is to be looking at state-sector productivity in a coming inquiry. 

Here's Patrick Nolan on it:
Measuring the productivity of public services is complex. The question is whether evidence of a lagging productivity performance tells us something about public services or more about the measures being used. In particular, the Statistics New Zealand data for the education sector do not account for changes in quality, yet the quality of products and services varies over time.

Adjusting estimates of public sector productivity for quality changes is a challenging task. To illustrate, the Productivity Commission has just published work on quality adjusted productivity measures for New Zealand schools using sector-level data. This work shows the benefits and risks of different approaches and highlights differing results emerging from different quality measures. The differences point to the need to better understand what measures reflect the performance of New Zealand schools.

There is still much more to do to better understand the productivity of New Zealand’s public services. And it is important not to simply point to poor performance, but to develop practical insights into how measures can help improve services. No single approach will work in all cases and no single organisation will have all the answers.

While understanding public sector productivity may be challenging this does not mean giving up. Nations cannot ignore the need to measure and improve the productivity of their public services. As David Cameron once noted, improving public services are “not about theory or ideology – they are about people’s lives.”
Public sector productivity is a big problem for a few reasons.

One of those reasons is that a reasonable chunk of what the public sector does destroys value, and making the government more effective and efficient towards those outcomes is itself destructive. Imagine a world in which the government were tremendously efficient at identifying and prosecuting people who:
  • Smoked a joint;
  • Provided assistance to a terminally ill person in constant pain to help end their suffering;
  • Gave somebody a ride in their car, not for free but for cash, and without a P-endorsement on their license;
  • Blasphemed;
  • Watched an R-13 TV show with their 12 year old on DVD (illegal) rather than on broadcast (legal);
  • Rode a bicycle without a helmet;
  • Was homosexual before the government finally got around to deciding that it shouldn't be prosecuting people for being homosexual.
Efficiency in the pursuit of public bads is bad. Government produces many public bads. The same government that funded Alan Turing, killed him. The first part of that was a public good; the evil of the second part outweighs the first. Views on the desirability of efficient government depend on views on that government's benevolence. And there's much to be said for government stopping doing evil things before doing good things better. 

But leaving all that aside, you need to put some value on the outputs. There's plenty of stuff you can do to get closer towards one sense of cost-efficiency. Outsourcing, competitive tendering, and a lot of the outcome-based contracting under the Social Investment Approach will all give us a much clearer picture of the cost per unit of outcome. And the Social Investment Approach also gives a sense of the value to the government of those outcomes in terms of reduced long-term fiscal liability - which is a darned sight better than what we have had before. But it still doesn't really say how much people really value the outcomes.

Patrick's link does go to some exceptionally interesting work on productivity in education; I'll follow up on that in a later post.

It'll be interesting to see where Prod Comm takes the inquiry. I'm an optimist about getting better cost-per-outcome efficiency. I'm a pessimist that we'll ever get the right mix of outcomes. But I'm optimistic that we will eventually at least be able to stop the purchasing of bad outcomes. 

No comments:

Post a Comment