Stats NZ has a new website.

For new releases go to

As we transition to our new site, you'll still find some Stats NZ information here on this archive site.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
6 – Health care

This section discusses the availability of data for measuring change in health care output and inputs, along with the issues associated with these data sources for those interested in measuring change over time in health care productivity. Separate sub-sections are devoted to output quantity, output quality, output weights, labour inputs, capital inputs and intermediate consumption. However, before launching into these descriptions, here follows a brief description of the health care system in New Zealand, followed by a summary of concepts and a description of the existing analyses in New Zealand of health care output (along with critiques of their sources and methods).

6.1 The health care system in New Zealand

Possibly the single most important question to be decided before work on government productivity could move forward is that of scope. Economic statistics published by Statistics NZ such as Gross Domestic Product and productivity are produced on an industry basis, without regard to the finance of those industries. Health care output estimates consistent with the System of National Accounts are required to expand Statistics NZ’s existing suite of industry-based official productivity estimates. Users may also wish to answer questions about productivity with regard to government expenditure on and/or delivery of health care services. The selected scope has implications for data requirements, compilation methods, and coverage. Refer to section 5.2 for discussion of scope as it relates more broadly to government sector productivity measurement.

This section describes the health system in New Zealand to help the reader understand, for example, the scoping issues as well as the general context for any ensuing productivity measure.

Figure 2, from the Ministry of Health’s Health and independence report 2008 (MoH 2008), presents the structure of the health system in New Zealand.


Figure 2 Structure of the health system in New Zealand

Source: Health and Independence Report 2008


The New Zealand Public Health and Disability Act 2000 established 21 District Health Boards (DHBs), governed by boards of directors that include locally elected members and ministerial appointees. The 21 DHBs are responsible for planning, funding and delivering most publicly funded health services to New Zealanders. DHB providers offer most secondary and tertiary hospital services, including for all acute and most elective cases. Private hospitals offer elective services on contract to DHBs and on a private basis, generally for those cases that do not meet the need thresholds established by DHBs.

The first primary health organisations (PHOs) were introduced in 2002 as the cornerstone implementation of the Primary Health Care Strategy. There are now 82 PHOs with more than 4.04 million enrolees (more than 95 per cent of the New Zealand public), involving the vast majority of general practitioners and practice nurses. Governed by non-profit boards of directors, PHOs contract with DHBs to offer a range of preventive and curative services, as well as an increasing array of population health services. From 1 July 2007 all New Zealanders enrolled with PHOs could benefit from low or reduced-cost primary care services.

Much of the health care in New Zealand is delivered by non-government organisations (NGOs). These include providers with national contracts, such as the Royal New Zealand Plunket Society, and providers who contract with their regional DHBs, such as community-based NGOs providing services to people with experience of mental illness.

Overall, 77 percent of health expenditure is funded by the public through taxes, 5 percent by private health insurance premiums, and less than 1 percent by non-profit organisations; the remainder (17 percent) is paid directly by those receiving the service. In addition to programmes funded by the Ministry of Health and District Health Boards, the other main public funder is the Accident Compensation Corporation (ACC), which pays providers to treat New Zealanders who suffer from accidents or injuries.

The Ministry of Health is a policy advisor to the Minister of Health, an agent of the Minister for monitoring and overseeing DHBs, a funder of DHBs and national services such as national screening services, and a provider of regulatory and other functions (eg public health).

6.2 Health care output top

6.2.1 Summary of concepts

The basic framework for measuring health care (and other non-market) output is set out in the SNA 1993. It recommends that current price consumption of those goods and services for which prices are not economically significant, or indeed zero, should be measured as the sum of input costs (compensation of employees, intermediate consumption, consumption of fixed capital, and taxes less subsidies), taking into account all sources of finance. Measuring change in output according to change in inputs in this way is referred to in this feasibility study as the ‘output=inputs’ method.

The SNA 1993 goes on to encourage measurement of the actual volume amounts of goods and services produced and consumed (referred to in this paper as the direct volume method), but acknowledges that the use of deflated expenditure on inputs is an option; that is, the ‘output=inputs’ method.

The other main publications available, listed in section 5.1, provide further detail on implementation. As each publication has appeared, from the Eurostat Handbook to the Atkinson Review and most recently to the OECD’s forthcoming manual, the guidance has been incrementally improved and refined, each publication building on what has gone before.

The common thrust in these publications suggests a growing international consensus on what would constitute a conceptually pure method of measuring change in the volume of health care output over time. However, the literature acknowledges that measuring health care output according to this ideal is a tall order. Table 3 sets out this possible conceptually pure method, as well as the acknowledged limitations.


Table 3. Aspects of a conceptually pure method of measuring change in the volume of health care output over time, and corresponding limitations

Aspect of conceptually pure method


A. Health care output should be defined from the perspective of the consumer: a unit of output covers a whole course of treatment for a particular condition or disease, rather than the individual activities that make up a treatment.

A.1 Typically, countries' health information systems do not readily provide information on the whole course of treatment, in particular joining up activities in primary care (including appointments with a general practitioner) with those in secondary care (including day care appointments and inpatient operations and stays).

A.2 Even where information systems do provide such joined-up information, there are significant conceptual issues that need to be addressed for defining the unit of output for large parts of the health system. While it is relatively simple to define the whole course of treatment for curative care (typically 'one-off' type treatments, including treatment for broken leg and heart attack), this is not the case for other types of treatment, for example, what is the whole course of treatment for those with mental health problems, with chronic health conditions, with multiple health problems (co-morbidity); what is the output of preventative care?

A.3 This concept cuts across the usual National Accounts methods, which distinguishes between the value added provided by different parts of the economy. In the case of health, primary care providers in the private sector are working alongside secondary care providers in the public sector – unpicking the relative contributions of each of these providers (their individual value added) is a non-trivial task.

B. Any measure should be as comprehensive as possible, covering all of the different types of treatment provided by the different parts of the health system.

B.1 Many countries have good information on inpatient hospital stays, but relatively less information on day care, and may have almost no information on primary care treatment that is easily accessible.

C. The relative importance of different health treatments should be given by the marginal valuation

C.1 There is little conceptual basis for judging relative importance in the absence of a competitive market that allows for clearing or marginal prices:(i) average costs provide information from the producer perspective and are generally easily available (hence this is what is recommended in the literature); use of costs can lead to what some commentators see as perverse results. For example, in the case of cheap wonder drugs ‘undervaluing’ significant health benefits (while this is true, it is also the case for any good or service; eg the micro-chip);(ii) revealed preference studies could provide information from the consumer perspective;(iii) measures of health status eg QALYs do not combine information on all aspects of health treatment that might be pertinent.

D. Change in health care output can occur because of either a change in the number of treatments and/or a change in the quality of treatment. Some of the quality change can be picked up by differentiating between different types of treatment. Other types of quality can be picked up by examining the contribution of treatment to outcome.

D.1 Differentiation will not pick up quality change beyond what can be picked up in casemix.

D.2 Quality is multi-dimensional and the choice of which dimension or dimensions are relevant depends on the type of treatment and diagnosis. For example, for invasive surgery the main dimension of quality may be survival, whereas for non-life threatening events the comfort of the patient, and the extent to which the patient is informed etc may be more relevant.

D.3 There is no agreed method for combining two or more quality dimensions for treatments where this is apropriate: how relatively important is health improvement against, say, waiting time?

D.4 Distinguishing the role played by a health system from other factors affecting health status (healthier lifestyles, smoking cessation, etc) is difficult and needs further work.


This summary of concepts merely presents a broad picture of how health care output should be measured. Many aspects of health care output need to be dealt with in a consistent and orderly manner. The remainder of this sub-section covers the most important of these aspects, and relates mainly to a measurement approach involving individual activities as the units of measurement, rather than the health care pathway (the units of measurement for the health care pathway need further consideration at a global level).

How to deal with multiple diagnoses

Many patient records contain a number of diagnosis codes representing possible co-morbidities (multiple, concurrent illnesses) or simply providing further information on the characteristics of the single health problem that the patient has. The first in the list of diagnoses is, in principal, the ‘main’ diagnosis and this is the one that should be used for determining the type of health care problem.

How to deal with misdiagnoses

Occasionally, the diagnosis offered by the first physician, typically a General Practitioner, will be over-written by a specialist on referral. Ideally, the diagnosis used for the spell of illness should be the correct code as determined by the specialist, with all connected health care activities being given this correct code.

How to deal with death during treatment

Some commentators have suggested that death during treatment should be treated the same as a ‘broken brick’ is treated in the market sector: consumers would not pay for a broken brick; it has no value. This would be unfair for the health care sector, which is doing what it can to improve the health of patients who may have a high risk of death. A better analogy would be with the services of a lawyer: what consumers are paying for are the services of the lawyer to maximise their chances of avoiding a guilty verdict. Similarly, patients (and other funders on their behalf) are paying for the hospital and other actors to maximise the patient’s chances of recovery.

How to deal with missed and cancelled appointments

These should be dealt with according to the cause. For example, if the cause of the cancellation is that the hospital had not organised the necessary resources, then it would be inappropriate to record this as an activity (but to record any use of resources, when part of the resources needed have been organised). If the cause of the cancellation is to do with the patient not turning up, for example, then it could be appropriate to record this as an associated activity (with associated inputs).

How to deal with first and follow-up appointments top

Many health care problems, especially those provided in outpatient and day patient settings, require ongoing, repeated treatment. One way to deal with these is to record each activity separately. An issue arises with such an approach if the medical best practice guidance changes. For example, if the recommended periodicity of repeat appointments changes from once per month to once per fortnight; other things being equal, this would make for an apparent doubling of activity (although updating the cost weights would counter some of the effect). Another way to deal with these repeat appointments would be to aggregate them, and alter the unit of measurement from ‘an individual appointment’ to ‘all of the appointments in a year for a patient with a particular diagnosis’. Identifiers on patient records that distinguish between first and follow-up appointments, as well as the patient identifier on the record, can be used to group appointments in this way.

Such an approach might also be taken with repeat prescriptions, for example, as well as repeat appointments with General Practitioners, with mental health services, and so on.

How to deal with complications

In some cases, complications arise from mistakes made by the health institution. A well-publicised example of this is hospital-acquired infection. It might be appropriate to record only activity associated with the illness as presented on first contact, and record activity following on from errors only as extra resources required to treat the patient. From a productivity measurement perspective, this would make sense: reductions in, for example, hospital-acquired infection (and other things being equal) would mean the same quantity of activity but reduced inputs and therefore improved productivity.

In other cases, the complications arise from the nature of the patient and their illness. In such cases, it would be appropriate to record any extra activities as output. Indeed, hospital inpatient recording systems do distinguish between some types of activity that have complications and those that do not.

How to deal with different types of patients

There are greater risks associated with elderly patients admitted to hospital for invasive procedures compared with younger people. Where this results in more activities per person, then the activity-based measure of output will pick this up, as the greater number of activities for the elderly will be recorded. Where this results in more expensive versions of the same activities, this will not necessarily be picked up by an activity-based measure, and it would make sense to deal with this in a different way. For some inpatient treatments, the classification system does indeed distinguish between the elderly and others. Such an approach could be extended to other parts of the health system where that makes sense.

6.2.2 Existing analyses and data collections in New Zealand top

This section provides a summary of the two existing, and available, methods for measuring health care output:

  • the Ministry of Health compiles health care output estimates as part of its hospital productivity work, which is published annually in the Health and independence report, the last one being published in 2008 (MoH 2008).
  • Statistics NZ also compiles information on health care output so that the National Accounts can comprehensively cover the total economy. However, Statistics NZ does not publish a separate health care output index.

Other studies have been published, most notably a study by the New Zealand Business Roundtable, Productivity performance of New Zealand hospitals 1998/99 to 2005/06, (NZBR 2008) but information on sources and methods have not been made available for this feasibility study.

6.2.3 Ministry of Health work to develop and improve measures of health care productivity: an update provided by the Ministry of Health

The Ministry of Health produces and publishes a measure of hospital productivity. This section, which describes their approach as well as potential future areas of work to improve the approach, has been written by colleagues at the Ministry of Health.

An approach to measuring national hospital productivity, based on centrally collected data to assess sector wide trends, has been developed by the Ministry of Health. The measures produced using this approach have been adopted as one of a set of headline indicators of systems performance, and have been published annually for the last three years in the Health and independence reports for 2007 and 2008 under ‘Progress on Headline Indicators’, and in the 2009 Ministry of Health Annual Report under ‘Efficiency and Value for Money’.

The national hospital productivity measure includes two views of productivity based on medical and surgical output, labour personnel cost input growth, and labour force FTEs input growth, for medical and nursing personnel working in medical and surgical services in DHB provider-arms. A technical paper describing this productivity measurement work is available from the Ministry on request.1

The Ministry is currently reviewing how productivity measurement could be incorporated into future work programmes. Potential areas of work to generate aggregate health system and hospital productivity measures, so that they are consistent with international best practice include:

  • a first-stage expansion of the scope of the existing DHB provider-arm hospital-based measure to include maternity and neonatal, mental health, health of older people, and disability support services. This includes capturing other staff groups providing these services, including allied health personnel (for example, occupational therapists and physiotherapists), and inclusion of non-labour inputs for capital and consumables.
  • a second-stage expansion of the scope to capture services provided outside of hospitals, including primary and community health care.
  • review of and improving on the methodology, assumptions, and data sources used in the productivity measure, including the means of weighting output for the cost and quality or relative value of care provided.

This is a challenging piece of analytical work, reflecting the complexity and resource intensive nature of productivity measurement work, and the challenges posed by the data sources that the Ministry has to work with. The same key issues that Statistics NZ has identified in its feasibility study are also very relevant to Ministry productivity measurement work: namely those of scope, definitions of health sector output and quality of services provided, and the absence of prices for many services provided within the public health system.

Other related productivity measurement work in the Ministry of Health and work to improve hospital productivity

There are other, related, areas of work that are being conducted within the Ministry. These are:

  • development of DHB-level productivity measures for accountability purposes used to assess DHBs planned delivery on productivity
  • work on productivity benchmarking in DHBs
  • a focus on gains in hospital productivity, including more efficient and productive wards, improved day surgery and theatre utilisation, improved workforce productivity, better use of joint procurement, and reduced cost of back office functions
  • other initiatives to improve hospital performance, as set out in the Statement of intent 2009-12, include the Ministry working with DHBs to develop strategic partnerships with private sector providers in order to make more effective use of resources and capacity that span both public and private sectors.

6.2.4 Statistics NZ’s health care output method top

An implicit part of Statistics NZ’s National Accounts economy-wide measure of change in output is the health care industry’s contribution to output. However, this is not separately published and is used to ensure comprehensiveness of the economy-wide measure.

It is a direct volume measure, which was introduced in the late 1990s as a replacement for the previous (output=inputs) measure. It is calculated using information from three data series, which are combined using fixed weights to form an index that is used as a proxy for the whole of public sector health care production. The three data series, along with their weights, are:

  • inpatient discharges, casemix adjusted (IP), 85.5%
  • day patient discharges (DP), 7%
  • mean length of inpatient stay (ALOS), 7.5%

An inpatient stay is defined in New Zealand as one where the patient is admitted into hospital and stays overnight. A day care patient is one where the patient is admitted into hospital, usually for more than three hours, but does not stay overnight. An outpatient appointment is one where the patient is not formally admitted and is usually in hospital for less than three hours. Note that these definitions are not necessarily shared by other countries or indeed by the OECD's classification of 'mode of production', as set out in its publication A system of health accounts (OECD 2001b).

(i) inpatient dischargestop

Up to 1993/94, casemix-adjusted data on inpatient discharges were available from the Ministry of Health. Since then, only data combining day patients and inpatients have been acquired on a casemix-adjusted basis, from the publication Ministry of Health’s Hospital throughput report, the latest report including data for 2003/04 (MoH 2006). From 1993/94, growth in the quantity of casemix-adjusted inpatient discharges is calculated by applying the annual growth rate in the combined number of inpatient and day patients to the old series ending in 1993/94.


where j denotes the DHB and the asterisk (*) indicates that IP and DP are casemix-adjusted. Casemix adjustment of the kind carried out on the data used by Statistics NZ means distinguishing between different types of treatment, thus allowing growth in treatments with different costs to be properly accounted for:


where i denotes the type of hospital activity and c are the mean costs of each of the different kinds of activity in each DHB.

Calculation of equation (1) is carried out by Statistics NZ.

Calculation of equation (2) is carried out by the Ministry of Health.

(ii) day patient discharges top

The total number of day patient discharges, without any casemix adjustment or other breakdowns, are taken from the Ministry of Health's Selected morbidity data for publicly funded hospitals publication, the latest report providing data for the year 1 July 2002 to 30 June 2003 (MoH 2006b). This is the volume measure.

(iii) mean length of inpatient stay

The mean length of inpatient stay, without any casemix adjustment or other breakdowns, are also taken from the Ministry of Health's Selected morbidity data for publicly funded hospitals publication (MoH 2006b). This is multiplied with the number of inpatient discharges to arrive at a total number of bed-nights, which is the volume measure used.

(iv) aggregation

In order to combine the three indicators into a single health care output growth series, each indicator series (only inpatients is shown below) is turned into an index with 1996 set at 1000:

(3) Value of index for IP = 1000 for t = 1996;

The aggregation method involves combining the value of each of the three indices for each year using fixed weights:

(4) Value of health care output index = 1000 for t=1996;


Issues with current sources and methods

(i) inpatient discharges

The coverage of the ‘total throughput’ figures used changes over time, with the early part of the series including both inpatient and day patient activity, and the later part including only inpatient activity. The calculation takes this into account, with the changes in the later inpatient-only part of the series being used as a proxy for change in the combined inpatient plus day patients series. Information on the number of day patients is available, and could be incorporated into the calculation to improve the quality of the estimates.

Recommendation H1

The available information on the number of day patients should be incorporated into the existing method of calculation of Statistics NZ’s health care output.

The year-on-year percentage changes in casemix-adjusted throughput do not match the corresponding levels of activity reported in successive annual reports. For example, the 2003/04 report puts casemix-adjusted total throughput for private providers at 25,681.2, which it says is 10.72 per cent higher than the previous year, suggesting that the total throughput in 2002/03 was 23,194.7. The 2002/03 report puts the figure at 20,251. This suggests that there may be revisions to the figures following publication. Where appropriate, these revisions should be taken into account when compiling health care output estimates.

Recommendation H2

Revisions to estimates of casemix-adjusted throughput should be incorporated into the existing method of calculation of Statistics NZ’s health care output.

(ii) day patient discharges

The day patient discharges figures are not casemix adjusted, and ideally would be in order to take into account aspects of quality that disaggregation can pick up (higher cost treatment is generally higher value treatment). Day patient discharges are available with a breakdown by, for example, type of service. This should be taken into account.

Recommendation H3

Changes in the number of day patient discharges should be broken down by type of service. Along with information on average costs of these different types of service, this information will help to introduce an element of quality change into Statistics NZ’s measure of day patient output.

Throughout OECD countries and beyond, there has been, and continues to be, a drive to treat patients in the most efficient setting. This means that, when patients can be treated just as effectively in a day care setting as in an inpatient setting, it makes financial sense to do so. The patient is receiving the same output irrespective of whether there is an overnight stay or not (and it could even be argued that patient experience is better). This suggests that day patients and inpatients should not be distinguished. An argument for making the distinction is that the inpatient treatment is delivered to, for example, patients who have more severe symptoms or patients who require a greater degree of care (the elderly or those with co-morbidities), in which case the output could be argued to differ. If the breakdown of day patient discharges is consistent with the breakdown of inpatient discharges AND the breakdown distinguishes between severity of cases, then inpatient and day patient discharges should not be distinguished.

Recommendation H4

Consideration should be given to combining the number of inpatient and day patient activities, where these are substitutes, in order to improve the price / volume breakdown.

Day patient figures may also be affected by revisions, which are not being taken onboard in the National Accounts.

Recommendation H5

Revisions to estimates of the number of day patients treated should be incorporated into the existing method of calculation of Statistics NZ’s health care output.

Large quantities of activities are excluded from analysis, many of which are health interventions and which are mainly of a diagnostics nature (colonoscopies, gastroscopies). The number of excluded cases in 2000/01 was 158,825. Ideally, there should be consideration of whether these excluded cases should be included in a measure of health service output.

Recommendation H6

Consistent with recommendation 5.3.5 on comprehensiveness and representativeness, consideration ought to be given to incorporating all of the available information on activities in hospitals and other settings in order to maximise the comprehensiveness of Statistics NZ’s measure of health care output.

(iii) mean length of inpatient stay

Number of bed days (calculated as the product of mean length of stay and number of inpatient discharges) is a poor indicator of output. An efficiency drive in health care across the globe is leading to health care production shifting from the relatively expensive inpatient hospital setting, to outpatient and primary care settings. This is also leading to reducing lengths of stay for those people who require an inpatient stay. Where this improvement in efficiency has not lead to poorer health outcomes, this should mean that hospital output has not fallen. However, the inclusion of bed-nights in the National Accounts measure of non-market health care output means that these improvements in efficiency are leading to measured output falling.

There are two reasons for including bed-nights as an indicator of output:

  • The first is that, in a static system with no changes in the number of bed-nights due to changes in efficiency etc, an increase in bed-nights would be an indicator of hospital activity. However, this effect will already be picked up in the other two measures: the number of casemix-adjusted inpatient discharges and day patient discharges.
  • The other reason is that number of bed-nights might be an appropriate indicator for ‘boarders’ (a person who is receiving food and/or accommodation, but for whom the hospital does not accept the responsibility for treatment and/or care. However, a hospital may register a boarder. This excludes all babies born in hospital.). The extent of inclusion of boarders is not clear from the available definitions and data descriptions. It seems likely that boarders will have long ALOS.

Recommendation H7

The number of bed-nights should not be used as part of a measure of health care output for all types of hospital patient. It might be appropriate to consider using number of bed-nights as an appropriate indicator of the volume of health care output associated with ‘boarders’.

(iv) aggregation

The ratio of inpatient to day patient care expenditure from the Ministry of Health's Health expenditure trends in New Zealand 1994-2004 (MoH 2005) is 1:18, while the ratio of the (fixed base) weights used in the National Accounts is 1:122. For aggregating growth in non-market output, international guidance states that the use of cost weights is acceptable (see section 6.2.1). As such, the weighting scheme ought to be revised to take into account the actual costs of providing these services – or the costs of the services for which these series are considered to be suitable proxies. Given the ongoing shift of health treatments from inpatient to day and outpatient care, it seems likely that the unit cost of day care and outpatient treatment will have increased (it is also likely that the average cost of inpatient care will also have increased, but more slowly). This would mean that the weighting structure has been changing as the proportion of patients treated in day care and outpatient care has increased.

Recommendation H8

The weighting scheme should be updated, possibly as frequently as annually, to reflect the changing relative costs of providing the different services.

The weights provide information on the relative importance of growth in the different quantities that comprise health care output, and are usually applied to the growth rates in those quantities. In practice, the weights have been applied to the values of the indices in individual years. To make the index number methodology more consistent with standard practice, the calculation should weight together growth in the volume indicators, rather than indices of the volume indicators.

Recommendation H9

The method for aggregating the different sub-components of the health care output index should conform to the standard method involving weighting together changes in the volume of different activities using relative weights (rather than weighting together different index series). 

(v) other issues

The coverage of the health care indicators used in the measure is very partial. Many types of health care are not covered, including outpatients, mental health, preventative care, primary care and long term care. According to Health expenditure trends in New Zealand 1994–2004, of the Ministry's $8.5bn expenditure on health care, only $2.7bn (31 per cent) was spent on inpatient and day patient care. Outpatient care cost $2.9bn (34 per cent) and long-term care cost $1.7bn (20 per cent). As mentioned above, the global drive to deliver health care to patients in the most efficient setting may mean that these are relatively high growth areas. Extending the coverage of the health care output measure is desirable

The timeliness of data used is less than optimal. The underlying health series is based on health indicators up the 12 months ending June 2004 (after when the inputs-based figures are used to provide recent history).

6.3 Health care inputs top

This section discusses the availability of data for measuring change in health care inputs, along with the issues that these data sources present for those interested in measuring change over time in health care productivity. This section focuses on the availability and quality of appropriate data, as the concepts and methods are rather less contentious than is the case for output. That said, this section does include a summary of the concepts and methods, as well as descriptions of existing analyses of health care inputs (along with critiques of their sources and methods). Measuring productivity (OECD 2001) covers concepts and methods more thoroughly.

This section is organised differently from section 6 on health care output, which had separate sections on quantity and quality. Instead, this section distinguishes between labour, capital, and intermediate consumption as the three types of input into production, each of which has its own sources and methods.

The concepts and methods for calculating inputs are the same, irrespective of the exact specification of the productivity equation – single, multi, or total factor productivity.

6.3.1 Labour inputs: summary of concepts

Ideally, the best measure of labour inputs to production is the number of hours (actually) worked, differentiating between different types of labour. The weights to be used to aggregate changes in the number of hours worked by the different types of labour, should be the total employment costs of the different types of labour.

Hours (actually) worked is better than simple numbers of people working because the latter ignores changes in what labour is contributing to production (for example a shift to or from part-time working, greater or lesser overtime working, and so on). It is also better than counts of full-time equivalents (FTE), as the information used to convert to FTE is not usually based on actual hours worked but on contracted hours, paid hours, or even simpler assumptions; such as, that part-time working is half full time working.

6.3.2 Labour inputs: existing analyses: Statistics NZ’s measured sector labour volume data

Statistics NZ’s measured sector productivity estimates draw on a wide range of labour market data. As the measure only includes the measured sector, it excludes the health care and education industries entirely (along with public administration and defence). Nevertheless, the sources used also provide information that could be used to compile measures of labour inputs for the health care and education industries.

The labour volume series is an estimate of paid hours for all employed persons engaged in the production of goods and services in the measured sector in New Zealand. Paid hours, rather than actual hours, is used because only the former is available from the Quarterly Employment Survey. The series is compiled using a number of data sources, from which the best characteristics of each are used for productivity measurement.

Throughout the series, there are three components that are summed to an industry level:

  • employees in industries covered by employment surveys
  • employees in industries out of scope of employment surveys
  • working proprietors.

For each of these components, the labour volume series is constructed by multiplying together the following two variables to give total weekly paid hours for the measured sector:

  • job/worker counts
  • weekly paid hours per job/worker.

For the first of the three components, data from the Department of Labour (DoL) Employment Information Survey is used up to 1980, when it became the DoL Quarterly Employment Survey (QES). The DoL data was the sole source for employee counts and hours paid for this component until 1989, from which point annual Business Demography counts are rated forward by quarterly movements in employee counts from the QES. The resulting quarterly series of employee numbers is then multiplied by average weekly paid hours from the QES to achieve a quarterly series for paid hours. In 1989, Statistics NZ assumed responsibility for administering the QES. From 2000 onwards, monthly Linked Employer-Employee Dataset (LEED) has replaced Business Demography as the sole data source for employee counts, and is combined with QES data on average weekly paid hours.

The second component includes employees in the following ANZSIC industries that are omitted from the coverage of the surveys above:

  • A01 – Agriculture
  • A02 – Services to agriculture
  • A04 – Commercial fishing
  • I6301 – International sea transport
  • L7711 – Residential property operators
  • M813 – Foreign government representation
  • Q97 – Private households employing staff.

Prior to 2000, Census of Population and Dwellings data provides benchmarks for employee counts and average weekly hours for this component. Prior to 1986, counts are interpolated using data from the Agriculture Census where appropriate. From 1986 to 2000, quarterly estimates of change from the Household Labour Force Survey (HLFS) are used to interpolate weekly hours between census benchmarks. From 2000 onwards, LEED provides monthly data on employee counts, while the average hours methodology remains unchanged.

For working proprietors, the third component, prior to 1986, census benchmarks are used to calculate both counts and average hours for almost all industries, supplemented by data from the DoL employment surveys and the Agriculture Census where appropriate. From 1986 to 2000, both hours and count data are benchmarked using totals from the census and interpolated using data from the HLFS, as in the previous component. From 2000 onwards, LEED provides annual benchmarks for working proprietor counts, supplemented by data from the HLFS and QES. Census data continues to provide average hours benchmarks during this period.

LEED employee count data are unavailable for the last quarter of the series and LEED working proprietor count data are unavailable for the last year of the series, so the latest movement in the employee count data is estimated as the latest movement from the QES, and the latest HLFS movement is used for industries outside the QES scope. Working proprietor counts are rated forward using HLFS movements. Adjustments are made to the QES and HLFS data where necessary. Average hours worked per industry is calculated as in previous years, however the data are adjusted to account for the proportion of secondary jobs for employees in industries out of scope of the QES and working proprietors.

The labour input index

The industry volume series are aggregated to the measured sector level by means of a chained Törnqvist index. The quantity relatives in the index are two-period ratios of industry labour volumes. Industry two-period mean shares of measured sector nominal labour income form the exponential weights.

Use of LEED

LEED is the main data source of counts of employees and working proprietors from 2000 onwards. The LEED dataset is created by linking a longitudinal dataset from the Statistics NZ Business Frame with longitudinal data from administrative taxation sources. Statistics NZ sees LEED as the best available data source for measuring labour counts for the reasons outlined below.

For measurement of employees, LEED data differs to the previous Business Demography Database (BDD) in the following ways:

  • LEED employee count data are monthly, whereas under the previous approach, quarterly data was used. Therefore LEED captures the seasonality of labour volume better.
  • Unlike the previous approach, LEED counts are not interpolated using survey information, reducing the effect of sample error on the series.
  • LEED data includes information about secondary jobs for industries outside of the scope of the Quarterly Employment Survey (QES). These jobs were previously excluded from the series.

For measurement of working proprietors, LEED data differs to the previous census or HLFS measurement in the following ways:

  • The majority of the working proprietor data are based on LEED annual benchmarks, based on a working proprietor's main income source over the year, ie it is not a point-in-time estimate. It is modified to incorporate seasonality using the HLFS and QES; however, the annual average counts remain the same.
  • LEED data includes information about people with secondary jobs (based on income) as a working proprietor. These jobs were previously excluded from the series.
  • Under the previous methodology, census benchmarks could be extrapolated forward for up to five years before being finalised. However, LEED provides annual benchmarks and, at most, it is only the latest year which will be extrapolated forward.
  • Working proprietors who pay themselves a salary can now be identified more accurately using LEED.

Composition-adjusted labour input

Composition-adjusted productivity measures account for the impact of changes in the skill composition of workers. As multifactor productivity (MFP) is measured residually, when change in skill is incorporated as part of labour input, it provides a theoretically better productivity measure, as it would otherwise be allocated to MFP.

Composition-adjusted labour is calculated by adjusting the Labour Volume Series using movements in a labour composition index, which estimates changes in skill composition using proxies for skill, namely education attainment and work experience. The labour composition index is calculated using the HLFS to estimate the proportions of each skill category of worker, while the New Zealand Income Survey (NZIS), an annual supplement to the HLFS, is also used to compile income shares for each of these groups.

Due to the availability of NZIS data, the series runs from 1998. For further background on composition-adjustment, and details on the methodology, see the Accounting for changes in labour composition in the measurement of labour productivity (Statistics NZ 2008).

6.3.3 Existing analyses: Ministry of Health’s labour inputs top

As a gross output-based labour productivity measure, the Ministry of Health’s methodology does not include either capital or intermediate consumption as part of inputs. The labour inputs are tightly defined around doctors and nurses working in medical and surgical units. The information is sourced from returns completed by DHBs.

This measure of labour inputs is not separately published: only the calculated productivity measure is published, in the annual Health and independence report (see, for example, MoH 2008).

The labour measure used is based on numbers of full-time equivalents. The definition of a full-time equivalent is not standardised and would typically be based on, for example, contracted hours or possibly hours usually worked. Ideally, a measure of labour inputs should be based hours actually worked. Any changes over time in overtime worked, leave entitlements, the balance between full-time and part-time working, and so on, may not be properly incorporated in the full time equivalent measure.

The way in which contracted (as opposed to permanent) staff are accounted for in the measure may be a source of bias: whilst there is good information available on the expenditure by DHBs on contracted (temporary) staff, there is little information available to decompose this into price and volume components. In order to do this, the price and volume breakdown from employed staff is used.

The measure takes some account of the different inputs from different types of staff by distinguishing between, say, doctors and nurses and senior versus junior staff. It may be that further differentiation between types of staff, along the lines of the analysis carried out by Statistics NZ (Statistics NZ 2008) could improve the measure further.

6.3.4 Existing analyses: Ministry of Health Financial Templates(used up to 2005/06)

The Ministry’s source of information on the number of staff working in New Zealand health care, along with associated costs, for use in the published productivity measure up to 2005/06 is the Financial Templates, which are used as the mechanism for DHBs to report their ongoing expenditure to the Ministry. The Financial Templates report numbers of FTEs and costs for all salaried staff along with costs (not FTEs) for outsourced staff, although the Ministry’s currently published productivity measure only focuses on doctors and nurses in surgical and medical areas.

The level of disaggregation available from the Financial Templates is relatively low (at least compared to the level of disaggregation of information from other sources, for example HWIP (see below). For example, the breakdown by type of staff is limited to 5 categories: medical personnel; nursing personnel; allied health personnel; support personnel; and management/administration personnel.

The definition of costs does not extend to what is usually required for weighting together different types of labour: what is available is salary, rather than total compensation, although salaries may be a reasonable proxy.

The lack of information in the Financial Templates on the number of outsourced FTEs is resolved by assuming that the average cost of and average hours worked by an outsourced FTE is the same as a salaried FTE (total expenditure on outsourced staff is divided by the average cost of an FTE to yield the number of outsourced FTEs).

6.3.5 Existing analyses: Health Workforce Information Programme (HWIP) (used from 2006/07)

The Ministry’s main source of information on the number of staff working in New Zealand health care is the Health Workforce Information Programme (HWIP), which is run by DHBNZ. Example reports, along with more detailed information about HWIP, are available in the quarterly Health Workforce Information reports (DHBNZ 2008).

The HWIP includes a regular quarterly survey of DHBs, which asks about the staff employed. There are a number of mandatory fields which are listed in table 4. Table 4 also presents those percentages of the mandatory fields which are valid, pre- and post- a data correction phase.

For the productivity measure, the same data elements on staff numbers from HWIP as for the Financial Templates are used. HWIP provides an opportunity to expand coverage, as well as to increase the level of disaggregation (see section 5.3.3 for a discussion of the level of disaggregation in relation to output: this is also apt for inputs).

The collection of two further pieces of information has been under discussion, and data are beginning to be collected. These are the Australian New Zealand Standard Classification of Occupation (ANZSCO) and Workforce Strategy Groups mapping.


Table 4. Mandatory fields in DHBNZ and HWIP staff survey, along with indicators of data validity

Mandatory Fields

Valid Raw Data

Valid Data after correction

Birth Date



Common Chart of Accounts



Employment Start Date



Employment Status









Health Service



Iwi Affiliation



Scope Of Practice



Sex Total



Total Contracted Hours



Union Status



Average (over all fields)



Source: Reproduced from Table 1 in Health Workforce Information (DHBNZ 2008)


Over recent years, much effort has been put into standardising definitions, for example of the mapping between ANZSCO and the Workforce Strategy Groups, and the definition of what is a full-time equivalent (FTE), the latter being defined as a 40 hour working week, taking into account annual leave, statutory holidays, and time off in lieu.

Together, the mandatory fields along with the occupational information provide a great deal of information that is of use in measuring the quantity of labour in the New Zealand health care industry. In particular, the numbers of staff and hours provide a quantity measure of labour, and ANZSCO as well possibly, as length of service and other variables provide an opportunity to incorporate differences in the quality of labour using a disaggregation approach.

The detailed data dictionary is published on the web, entitled the Health workforce information programme (HWIP) code set (DHBNZ 2006).

This information on staff quantity and quality would need to be used alongside appropriate weights, or staff costs, when used as a measure of labour input. It would be useful to compare and reconcile the information from the HWIP source with the Financial Templates that are the source for the staff numbers (see section 6.3.4) used in compiling the Ministry’s current productivity measure.

6.3.6 Capital inputs: summary of concepts top

Ideally, the best measure of the capital input to production is the flow of capital services that are generated from the capital held. Capital services are not directly observed, but instead are estimated from a measure of the capital stock – assuming that the flow of capital services is directly proportional to the underlying stock of the capital being considered. This approach typically assumes constant capacity utilisation rates, and in New Zealand the capital stock data are predominantly sourced from the productive capital stock series from the National Accounts, which are derived using a perpetual inventory model (PIM).

The relative weight of capital services is given by the user cost of capital. The user cost of capital can be seen as an imputed rent: it is the rent that the owner of the capital might notionally charge themself for use of the capital. In some cases, there may be fully functioning capital rental markets, and the imputed rents may be inferred from equivalent actual rents. For many assets, however, there are no fully functioning capital rental markets, and the imputed rent has to be calculated indirectly. The user cost of capital can be seen as being made up of two basic terms: the cost of financing and the change in the value of the capital. The cost of financing is made up of two further parts: an estimate of the interest payment on a loan to purchase the capital and the cost of depreciation. See section 5.4.6 for discussion of the appropriate rate of return to use when undertaking this calculation for the public sector.

The OECD’s Measuring productivity (OECD 2001) sets out the methodology and concepts. Implementation for market sector industries in New Zealand is set out in Productivity statistics: sources and methods (Stats 2009).

The scoping issues discussed in section 5.2 are particularly difficult in respect of measuring capital. This is due to the need in a productivity measure to match the inputs to the output within scope: were the scope of the health care productivity measure to be delineated by whether the government or private individuals are the source of funding, then difficult questions arise about which capital assets should be included as part of government production. While (publicly-owned) DHB provider hospitals would clearly be in scope, what about (private sector) general practices? The solution to this has already been proposed, and is covered by recommendation G13 in section 5.4.3.

6.3.7 Statistics NZ’s measured sector capital services

The capital services input index measures the flow of capital services generated by the use of the stock of capital assets for a given March year. No allowance is made for differences (across industry and time) in asset capacity utilisation rates.

As capital service flows cannot be directly measured, industry level flows are modelled, based on the productive capacity of industry capital stock. The industry level flows are aggregated to the measured sector level using industry shares of the measured sector current-price capital income as weights. More specifically, the following steps occur:

  • The starting point is the annual constant-price productive capital stock series. An asset's productive capital stock is its gross capital stock adjusted for the decline in its efficiency.
  • Measured in constant prices, the productive stock represents standardised efficiency units and can be interpreted as a measure of the potential capital services that the asset can contribute to the production process. The productive capital stock series are built up using a perpetual inventory model (PIM) that generates productive capital stock estimates for 26 asset types by industry, of which only 24 are used in the capital services index.
  • The model specifies for each asset type a mean expected useful life, a retirement function based on a distribution about this life and its pattern of (hyperbolic) efficiency decline. These parameters, and gross fixed capital formation in constant prices, are used to estimate an asset type's productive capital stock in volume terms.
  • In addition to the PIM-derived fixed asset stocks, the range of capital included in the productivity measures is supplemented by estimates for seven other assets, namely livestock, exotic timber grown for felling, and five different types of land: agricultural and forestry; commercial; industrial; mining; and other non-agricultural land.
  • Capital service flows are assumed to be proportional to these productive stock estimates, and are aggregated to the industry level using a Tornqvist index, with weights based on implicit rental prices (or user costs) which are a function of an exogenous real rate of return, depreciation, net taxes on production and asset price changes.

The PIM produces estimates of capital services for the health care industry. The underlying data on capital assets that are fed into the model are collected by Statistics NZ from the Ministry of Health and through the Crown Financial Information System (CFIS). A fine level of detail is collected and used.

6.3.8 Intermediate consumption: summary of concepts top

Intermediate consumption, sometimes referred to as consumables, consists of all other items incurring expenditure other than labour and capital. The ideal way to measure the change in intermediate consumption over time is by deflating current price expenditure with suitable price indexes. Generally speaking, the quality of measures of volume change in intermediate consumption improves with the level of disaggregation of the expenditure and price deflator data. See section 5.3.3 for a discussion of the level of disaggregation of output, which is just as relevant for measuring intermediate consumption.

A measure of intermediate consumption is only needed as part of an input measure if a gross output total factor productivity measure is being constructed: it is inappropriate to include intermediate consumption if the output measure is a value added one. However, it is important to note that in calculating value added, intermediate consumption is subtracted from gross output.

6.3.9 Statistics NZ’s health care industry intermediate consumption

Statistics NZ compiles information on health care industry intermediate consumption in current prices as part of the New Zealand National Accounts.

The source of information on intermediate consumption in current prices is the expenditure reports that the Ministry sends to Statistics NZ’s Government and International Accounts unit, and information available in the Crown Financial Information System (CFIS). These contain a great deal of information on the type of goods or services incurring expenditure.

6.3.10 Ministry of Health information on expenditure

The Ministry collects information on all the (public) expenditure incurred by DHBs on consumables, as well as labour and capital. Part of the rationale for collecting these data are to provide Statistics NZ with the information that is needed in constructing estimates of health care industry intermediate consumption and the health care industry Producer Price Indices (PPIs). The level of the disaggregation of this information is as provided to Statistics NZ (see section 6.3.9).

6.3.11 DHB accounting systems

DHBs are actively seeking ways to improve their accounting systems. Some have procured off-the-shelf systems which allow the tracking of expenditure and to examine the reasons for changes over time, including the ability to distinguish between price and quantity change. Currently, there is no single nationwide system, with each DHB and the Ministry collects information on all the (public) expenditure incurred by DHBs on consumables, as well as labour and capital. Part of the rationale for collecting these data are to provide Statistics NZ with the information that is needed in constructing estimates of health care industry intermediate consumption and the health care industry Producer Price Indices (PPIs). The level of the disaggregation of this information is as provided to Statistics NZ (see section 6.3.9).

1 Technical documentation on the 2009 update of the Ministry of Health Performance Assessment and Management Steering Group (PAM) productivity metric, Ministry of Health, (unpublished).

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+