The ABS faced considerable unwelcome attention this week after it announced substantial revisions to the seasonally adjusted labour force figures for July and August. Reacting to the announcement, Treasurer Joe Hockey expressed dissatisfaction with the volatility in the labour force figures, suggesting funding constraints on the ABS were to blame, and indicating that a shift to a “user-pays” model may be a way of restoring the organisation’s financial position.
There are certainly indications the ABS is facing acute resource constraints. It is reportedly cutting 115 staff and former Australian statistician, Brian Pink, had blamed cuts to its survey program on funding cuts.
Ironically, the volatility in labour statistics that appears to have precipitated ministerial action to address the funding inadequacies does not seem to have had anything to do with those inadequacies.
There was no cut to the sample size, no change in fieldwork or data processing procedures to reduce costs and, indeed, no failure of systems or people that led to the wild fluctuations in employment and unemployment numbers in July and August.
In fact, the ABS has not revised the unadjusted (original) numbers. Rather, it is the series that is adjusted for the monthly seasonal variation in employment and unemployment that has been revised—because, for reasons unknown, the typical seasonal patterns of July and August (and September) appear to have changed this year.
That aside, it remains the case that all is not well at the ABS. But just how bad is the funding situation?
Keeping the lights on
The budget papers show the federal government allocating A$341 million to the ABS for fiscal year 2014-15. The ABS is expected to raise a further A$30 million from other sources, most of which presumably derives from data collections funded by other government departments and agencies.
To be sure, this a sizeable budget. But, as our national statistical agency, the ABS plays a critical role in the functioning of our society, supplying the economic and social statistics that are vital to policymakers at all levels of government and across a broad spectrum of portfolios. The data and statistical outputs are, moreover, extremely valuable to academic researchers, to various non-government organisations, and — ultimately — to the general public in understanding how Australia is faring and what needs our attention. Without the ABS, informed public debate and policy formulation are not possible, and forward planning by private businesses is severely hampered.
Of course, as is true of any publicly funded body, funding constraints are an ever-present issue, and it is very difficult to determine what is the “right” level of funding.
A basic economic principle is that funding be increased to the point at which the marginal benefit equals marginal cost (on the assumption of diminishing marginal benefits). However, while the marginal costs of the ABS’s activity should be reasonably easy to measure (albeit ignoring unexploited opportunities for improved efficiencies), the marginal benefits are extraordinarily difficult to gauge. In no small part, this reflects the public good properties of much of the statistical output – once the unemployment rate is measured by the ABS, we all benefit from the improved public policy that results.
Nonetheless, as the chart below shows, it is clear funding of the ABS has in recent years been relatively low and declining.
The figure presents a five-year moving average of direct federal government funding since 1997-98, deflated using the public sector Wage Price Index (a measure of labour costs, produced of course by the ABS) and indexed to 100 in the base period. A five-year average is presented because of the year-to-year variation in funding created by provisions for the census (conducted every five years).
Excluding the two funding years most affected by census funding (Figure 2) presents a much starker picture of funding decline: in real terms, funding received over the 2012-13 to 2014-15 period was only 85% of the funding received over the 1997-98-1999-2000 period. Significantly, all governments from the Howard Government on have played a role in the cuts to funding.
While improved efficiencies, such as gained from use of new technologies, could justify some reduction in funding, it seems highly likely that, to accommodate the budget cuts, the ABS would have to be producing fewer outputs now than it was in the late 1990s. It is hard to view this as a good development: Australia is a larger and richer country now than in the late 1990s, and the needs for good economic and social data are more likely to have grown than decreased.
The possible solution to this funding predicament aired by the Treasurer is to move towards a user-pays system for ABS statistical outputs. While it is true that there are benefits of ABS statistics that accrue to private businesses, the public good properties of these outputs (and their subsequent uses by others) suggest this is not a wise strategy. It will reduce research activity using ABS data, it will reduce the degree to which public debate and indeed policy is informed, and ultimately it will reduce the value of the ABS to the Australian community.
Moreover, the available evidence is that the market for these products is relatively thin and therefore pricing these outputs is unlikely to raise much revenue. A user-pays regime was in place until 2005, when Treasurer Costello abolished it. Budget papers show ABS revenue (excluding direct allocations from the federal government), expressed in 2014 dollars, of $25 million in 2003-04 and $21 million in 2006-07, implying it could have had no more than a negligible impact in a budget of over $300 million.
The government is right to place pressure on the ABS to find ways to lower costs by doing things more efficiently. But if we want the ABS do keep doing what it is doing (and we probably should), let alone more, increased direct funding from the federal government will be required.