Monday, 25 October 2010

Can 180,000 public sector workers really be doing nothing?

Many macroeconomists are calling for more UK public sector spending,. with Paul Krugman joining the fray (though note that he calls for the postponement of cuts to a later period, rather than denying their necessity altogether) .  At the same time, many commentators complain about public sector waste. The excellent Steve Richards is, I think opposed to cuts in general, but wants sackings of middle managers in the BBC whom he says are useless. One hears similar sentiments about managers in the NHS and, dare I say it, universities.

I think this is an unexplored area: people don't like the cuts, but they can identify inefficiencies, often in the very organisations where they work.  Can economists bring any facts to the table here?

I think they can, but I think they are a puzzle.

The UK stats authorities are leading on the effort to measure public sector productivity. In most countries the output of doctors is the number of doctors.  In the UK, the output of the health service is the waiting time  adjusted output of a vast collection of medical procedures, all weighted by their cost shares.  And so on for other hard-to-measure sectors like education output (exam-result adjusted kids in school).

What's interesting about this is that these measures expand back to 1997, just before the start, in 2000, of the huge expansion of the public sector (from around 5m workers then to 6million now). What do they find?  The latest Economic and Labour Market review makes depressing reading.  Here are the facts on the table:

  1. For 1997-2008,  using the quality adjusted figures (that raise output, since e.g. waiting times have fallen), output has risen, 2.9% per annum. Inputs (that is, labour, capital and intermediates) have risen by 3.2%pa. So "multifactor" productivity (that is output per input of labour, capital and intermediates) has fallen by 0.3%.  Note that quality adjustment raises output; if we did not quality adjust, output would be even lower and productivity would have fallen even faster.
  2. Another paper here looks instead at non-quality adjusted value added per input, 1995-2008.  The market sector multifactor productivity grows at 1%pa but the industries dominated by the public sector (but not wholly public) falls by 1% pa, even more.
What do we draw from this?
  1. One possibility is that all the data are wrong. Maybe. 
  2. So lets be the most conservative. Let's suppose that public sector productivity is falling by 0.3% pa.  That means this. In 2008, to get the same output in the public sector as 10 years before when you employed, say, 100 labour and capital inputs you now have to employ 103.  If this was the private sector you would only need 90 units. 
  3. So if we ignore capital and materials, if these data are right, then in the public sector 3 people in 100 are doing nothing, whereas the private sector is getting the same output with 10 less people.  
  4. Again, if these data are right, what does that imply?  I see two depressing implications.  First, if we just asked the public sector to deliver zero percent productivity increase since 1997, we would have been able to produce the 2008 output with 3% fewer workers.  3% of 6million is 180,000.  Second, if we asked the public sector to deliver the same productivity increase as the private sector then we would have been able to produce the 2008 output with 10% fewer workers, that is 600,000 workers.
  5. (Finally, to give a yardstick, I note that the newly-formed Office for Budget Responsibility forecast earlier this summer that 490,000 government jobs would be cut by 2014-15).
Can this really be right?  To take the most optimistic number, can it really be that if we went back to 1997 productivity levels, we can get rid of 180,000 workers and have exactly the same output?  If the productivity numbers are correct, what are these 180,000 workers doing?  Can they really be doing nothing?  But if they are doing something, why doesn't it show up in their output?  I am puzzled.

No comments:

Post a Comment