It also provides an overview of methods used to create the statistics and guidance on how they should be used.
- What is productivity?
- Does productivity measure value for money?
- Does productivity measure efficiency?
- How do we measure output of individual public services?
- How do we add together different outputs to derive an index of overall output?
- Do the output measures account for quality change?
- Why adjust output measures for quality change?
- Do these public service output statistics measure simply what is produced by the public sector?
- What about services like defence which are not directly provided to households?
- What inputs do ONS measure?
- How do we measure inputs into public services?
- How do we add the different inputs up to get an overall inputs index?
- Can we use these figures to compare public with private productivity?
- Can we use these figures to compare the efficiency of public sector production of public services with the efficiency of using the private sector to acquire those services instead?
- Can we use these figures to compare UK output with that of other countries?
- Can we use these figures to compare UK productivity growth with that of other countries?
In more realistic cases there are more than one, possibly many, outputs and more than one input that have to be added together (see below).
Productivity is then measured as an index number, set to equal 100 in the base year.
Movements in the index show how total output is moving relative to inputs.
What end-users have to pay for an output depends broadly on two considerations.
The first is how much input is needed to produce a unit of output. The higher productivity is the less input is needed to produce a given amount of output, so rising productivity will tend to reduce what end-users have to pay.
But what end-users actually have to pay depends also on a second consideration: how much the producers of the output have to pay for the inputs they use.
Returning to the simple steel case above, the cost of a tonne of steel depends both on how much steel is produced in an hour (productivity) and how much an hour of labour costs.
What a single measure cannot show is whether this is the best that could be done with current technology, that is, whether existing practice is best practice.
Answering that question would require finding out what best practice is, which would involve detailed service-based analysis. But if existing practice changes over time to get closer to best practice, then this would show up as a rise in productivity.
Also, see above, productivity does not show whether the public sector is buying the inputs it uses as cheaply as possible.
ONS uses data from administrative sources which records the amount of certain activities.
For example, in healthcare there are detailed records of health procedures carried out, broken down into well-defined Health Resource Groups (HRGs); in education there are pupil attendance records separately available for primary and secondary education; for adult social care there are records of numbers of people being looked after in care homes and of attendances at various day-care settings.
For education and healthcare these quantity measures are supplemented by quality measures (see below).
To get the overall output of goods produced in the market sector national income accountants weight each output by its price in a given year to get an overall value in that year. So cars count for more than apples or pears in the final total because cars have higher prices than apples or pears.
To get the increase in real output between two (consecutive) years national income accountants value each output in the second year at the prices ruling in the first year, add them up and subtract the value of the previous year.
Most public services do not have a market price, as they are not sold on a market. So what ONS do is find the cost of a unit of activity, as defined above and use these costs instead of prices.
So the amount of each activity is multiplied by its cost weight and the total can be added to the value of other public service activities, and also the value of pear, apple and car output to produce a figure for total output of the economy.
In education this means ONS adjusts the quantity/activity measure in the light of information on changes in academic attainment.
For healthcare the activity measures are adjusted to take account of health gain, post-operative mortality, the age-mix of the treated population, patient experience and a measure of the proportion of people on GPs’ lists who are receiving appropriate treatment for certain named conditions.
The high quality outputs sell for more than the lower quality because of the way the buyers value the higher quality. So quality is accounted for by the price differential.
This is much harder to do for public services because service users do not pay directly for the services and there is no user-driven differential to use.
ONS have therefore used quality adjustments based on factors such as academic attainment (for education) and post-operative survival and health gain (for healthcare).
For most areas of government spending most of the output is produced by the public sector, for example, healthcare delivered in NHS hospitals or education delivered by state schools.
However, sometimes the public sector provides services to households by buying in the service, for example if the NHS uses independent treatment centres to deliver knee operations or cataract treatments. These outputs are also included in the ONS figures.
For these services ONS follows the international convention that the amount of output produced is equal to the amount of input that is used up in this activity.
This means, of course, that, by definition, productivity cannot change for these services.
So, for example, ONS distinguishes between doctors, nurses and other staff in healthcare, and between teachers and teaching assistants in education.
But the overall input measure also includes inputs of goods and services used up in production, such as heating and lighting costs, textbooks, bandages and dressings. And it includes the inputs of private sector providers of healthcare, such as independent treatment centres or hospitals made available under the Private Finance Initiative.
Inputs also include an estimate of the annual use of fixed capital assets such as the school and hospital buildings and the IT equipment.
But for some areas of labour and all areas of goods and service inputs ONS estimates the quantity of input by taking figures for total spending on an input area and dividing by an appropriate price index.
If the price index perfectly reflected the mix of the relevant goods and services this division would give the exact actual quantity. In practice the price indices used (the price deflators) do not exactly match.
For those measured by deflating expenditure the component parts have already been accounted for by the construction of the price index used to deflate. So, in effect, the inputs are added together using cost weights.
The ONS public sector services statistics refer to services provided to households through the agency of government.
Some of these services may be produced by the private sector, for example, knee operations performed for the NHS by independent treatment centres.
Moreover these ONS statistics refer to total output (value-added plus goods and service inputs) whereas most other published productivity estimates only look at value-added.
The closest to a direct comparison is to compare productivity growth in those industries, such as healthcare and education, where production is dominated by public sector producers, with growth in industries in the market sector, such as manufacturing, where production is almost entirely by private sector producers.
Such comparisons however are not straightforward public/private comparisons.
Can we use these figures to compare the efficiency of public sector production of public services with the efficiency of using the private sector to acquire those services instead?To do this would require separate productivity figures for publicly-produced and provided public services on the one hand with privately- produced but publicly-provided services on the other.
These statistics are not currently available on this breakdown.
Simply comparing total spending by converting into say dollars at current exchange rates does not give a sensible answer, because the differences in prices between currencies need not be the same as the difference in exchange rates.
Alcohol in France, for example, is relatively cheap, whereas in Sweden it is relatively expensive. So alcohol in any common currency costs different amounts in the two countries. The solution to the problem is to calculate special 'purchasing power parities' (PPPs) which take account of the actual price differences.
The Economist magazine has a light-hearted version where the PPPs are the ratio of the prices of a big Mac in the relevant countries.
In public services the problems are compounded by the fact that the quantity of activity in areas like healthcare is measured differently in different countries.
So far OECD, which leads on the calculation of PPPs, has effectively valued the inputs into public services (for example, using the earnings of health professional etc for healthcare) rather than the outputs.
OECD is now undertaking a pilot project to try and use the new direct output measures to calculate PPS for healthcare.
The problems are much less for comparing growth rates, where much of the level effects drop out.