Enhancing measures of non-market output in economic statistics: A roadmap

Jason Annabel - Economic Research Hub

The ABS defines the industries of Health care and social assistance, Education and training, and Public administration and safety, as being predominantly non-market in nature. Non-market economic activity occurs when goods and services are provided free of charge, or sold at highly subsidised prices. 

These three industries make up around 17%¹ of the Australian economy and support 3,782,100 jobs² . These industries are of great importance to the Australian community – the services they provide directly affect the lives of many. 

Given the increasing importance of government spending in health and education, and against the backdrop of an aging population, the share of the economy represented by these industries and the employment opportunities they will provide in the future is expected to grow. 

With Government funding as part of the Enhancing Economic Statistics New Policy Proposal, the ABS Economic Research Hub is undertaking work to enhance ABS measures of non-market activity. The work aims to build robust and sustainable indicators for measuring growth in non-market output volumes for the health and education industries. These indicators will be: 

• not directly based on input costs;
• appropriately weighted; and
• reflective of quality change over time.

These indicators will implicitly reflect the impact of technological progress and innovation underpinning the delivery of non-market services.

The work will leverage off work done by other National Statistical Offices (NSOs) and in response to changing policy dimensions. 

The 2008 System of National Accounts discusses three main methodsfor compiling output volume estimates of non-market goods and services³. For ease of referencing and implementation we label them Bronze, Silver and Gold as follows: 

1. Bronze: This approach assumes that the volume of output is largely dependent on the volume of inputs used in production. Since outputs and inputs are strongly correlated, productivity growth over time is concealed.
2. Silver: This approach involves deriving output volume indicators that are independent of input volumes. Such indicators would normally be constructed by directly observing outputs, using the costs of providing specific services as the aggregation weights.
3. Gold: Gold is similar to silver, except the cost shares are replaced by value shares reflecting consumer utility. This involves estimating a ‘pseudo’ price index for non-market output. An ideal ‘pseudo’ price index should explicitly account for quality changes.

From a national accounts perspective, the gold standard is conceptually ideal, but impractical to implement for measuring non-market outputs, as the data required to construct a ‘pseudo’ price index does not exist. While techniques such as hedonic regression could be used to estimate ‘pseudo’ prices, the ABS is not aware of any country that has implemented the gold standard in its national accounts. In line with other leading NSOs, the ABS is ultimately aiming to extend the current silver approach for all aspects of health and education.

Following a review of the Australian data landscape, two papers were published in 2018 which articulated some of the measurement challenges from the perspective of hospitals. 

This paper framed the measurement problem for non-market health output – constant innovation and technological progress, combined with a lack of market transaction prices. It outlined work done by other NSOs in developing adjustments to reflect quality change. 

The paper outlined Australia’s current data sources and methods. Where possible, quantity indexes are used to measure non-market output. These are weighted by input costs. The 'mix' of health products changes over time as a result of consumer demand, new treatments coming online, technological progress and other factors, but it doesn't explicitly take patient outcomes into account. This is known in the literature as ‘implicit’ quality adjustment. It is a way of building a proxy for utility into the measurement, but the drawback is that it is observed from the producer’s perspective and not the consumer’s. The paper concluded by framing ABS’ preferred short term approach, which led directly into the paper discussed below. 

This paper constructed disease based output measures for hospital services in Australia over the period 2004-05 to 2014-15. The primary source data was Australian Institute of Health and Welfare (AIHW) summary statistics of hospital admitted patient care expenditure. 

Hospitals provide three broad types of services – admitted patient care, emergency department care, and non-admitted patient care. The paper focused on admitted patient care, which makes up 74% of total hospital spending. 

The paper stratified by diagnosis, because it allows for ‘substitutions’ across different types of treatments to be accounted for over time. It also allowed more focus on treatment outcomes, as opposed to the specific types of goods and services provided. The paper further stratified by patient age. 

The paper outlined the data used and the methods employed in constructing a volume index, as well as some of the complications encountered. It framed an alternative method which used the source data at a more dis-aggregated level. The paper found that growth in output volumes is marginally slower when more detailed data is used. This is because ‘substitutions’ across treatments through time is more accurately accounted for, though the source data is probably less reliable at more granular levels. 

The paper found that while the Australian population increased by 1.7% per year on average between 2004-05 and 2014-15, aggregate hospital output increased by 3.4% per year on average. This represents an increase in real expenditure on hospital admissions per person during this period. This could be driven by demographic change, though this question was not considered in the paper. 

It also found a disparity of growth across different types of treatments provided by hospitals. Treatments of diseases of the eyes, blood and nervous system increased much more quickly than aggregate hospital output, whereas treatment of circulatory system diseases and the provision of pregnancy and childbirth services grew more slowly than population. It further found that private hospital output increased at a faster average annual rate than public hospitals, mainly driven by stronger growth in the number of private hospital admissions. 

By June 2020, the ABS will review the work done to date on hospitals, with a view to expanding the measure beyond admitted care, and subsequently exploring the feasibility of implementing the methodology in the national accounts. The ABS will also design an independent output volume measure for ambulance services and residential aged care, and commence designing a measure for disability care. Work on disability care and residential aged care is not expected to be completed by June 2020. 

By June 2020, the ABS will design independent output volume measures for pre-school, primary, secondary, tertiary and special education. These measures will in part reflect the impacts of technological advancements that have allowed for education to be delivered in a more customised manner, such as distance learning, and on-line learning platforms. 

Owing to measurement complexities, the work on producing independent volume measures for special education and tertiary education is not expected to be completed by June 2020. 

The ABS is not proposing to investigate output volume measures for non-market services that are consumed collectively. Very little progress has been made internationally in terms of measuring volumes of these services. However there is potential to investigate experimental ‘silver’ volume measures for selected non-market activities where services are consumed individually, such as fire and rescue services and justice services, in a future research program. 

The ABS is consulting with peer NSOs and other subject matter experts, including academics, as part of this research program. Output measures for government education and health services produced by the UK’s Office for National Statistics (ONS) based on recommendations from the Atkinson Review (2005)⁶ and a recent feasibility study published by Statistics New Zealand⁷ are of particular relevance to this research project. 

The ABS will also work with the Productivity Commission, the Australian Institute of Health and Welfare, and the Commonwealth Departments of Health and Education. 

The ABS will publish a progress paper in December 2019 outlining work done to that point, followed by a results paper in June 2020. 

For further queries please email economic.research@abs.gov.au