Anthony O’Donnell reviews Daniel Mulino’s Safety Net: The Future of Welfare in Australia, La Trobe University Press, 2022 i + xii; 372pp


What do we talk about when we talk about ‘the Australian welfare state’? For some, it’s almost a contradiction in terms. We have a minimalist, means-tested, flat-rate system of income supports that in many cases fail the measure of adequacy for some of our most vulnerable citizens, and nothing to approach, say, the national health system of Britain. For others, we can be seen as pioneers, instigating aged and disability pensions in 1908 and offsetting our meagre welfare spending with a system of arbitration and conciliation that purported to deliver workers a ‘living wage’.

Daniel Mulino’s Safety Net: The Future of Welfare in Australia is welcome because it embraces the plurality of welfare state initiatives, their different motivations and logics, suggesting there is no single paradigm when it comes to constructing a welfare state. But for some readers — including policy historians — this might be as much a weakness as a strength. Mulino’s concept of ‘social insurance’ floats a little too freely from any historical grounding. Which is fair enough: Mulino’s book is, after all, subtitled ‘The Future of Welfare in Australia’.

Mulino uses ‘social insurance’ to refer to just about any welfare state spending. And it’s 70 pages in before he offers a ‘precise definition of social insurance’: all government payments or in-kind benefits contingent on loss or harm, and not limited to those based on contributory taxes or premia.

Well, maybe. For policy historians in Australia, ‘social insurance’ has a particular meaning: a series of failed attempts to institute some sort of tripartite contributory scheme of earnings-related benefits in the interwar period. And we have some good histories of this. Thinking of only book-length treatments, there’s Tom Kewley’s meticulous but studiedly unopinionated account, Social Security in Australia 1900–72; Rob Watts’ revisionist The Foundations of the National Welfare State; and John Murphy’s more recent A Decent Provision: Australian Welfare Policy, 1870–1949. On top of that, there’s a host of good articles by Sheila Shaver, William de Maria, Carmel Black, Murphy, and others.

Mulino does provide the reader with a romp through the history of risk management more broadly, and the Australian, US and British welfare states more particularly, but his account tends toward the technocratic and economistic. It lacks the nuance and engagement of Dan Bouk’s excellent and eminently readable How Our Days Became Numbered (2015). Although focussed on the United States, which came late to the social insurance scene compared with Germany or Britain, Bouk’s book explains how actuarial and statistical principles around risk expanded from a pre-occupation with life insurance in the 19th century to a concern with public health management in the early 20th century, the promise of extending life overtaking the mathematics of predicting death.

But as life became longer and safer, it also became less economically secure. For Bouk, the significance of Roosevelt’s social insurance was that it applied the statistical imagination to daily economic existence — work and wages — and more people fell under the actuarial gaze. Similarly, writers in the Foucauldian ‘governmentality’ tradition — I’m thinking of people like Pat O’Malley or William Walters — have interrogated these technologies of social insurance.

This all helps us tell a story of ‘social insurance’ in Australia that supplements Mulino’s fairly anodyne account. Social insurance was a favoured policy of conservative governments precisely because contributory schemes weren’t simply a means to address a technical problem concerning financing; they also embodied a particular societal model based around prudence, the alliance ‘of labour with frugality and thrift’, sobriety and an early night, with benefits appearing as a legal right linked to contributions, ‘an exchange between equals that did not vitiate independence’.[1]

Former Australian Prime Minister Robert Menzies (Universal Images Group via AAP Photos).

Australian conservatives remained committed to the ideology of insurance rather than social assistance, with the Liberals going into the 1949 election insistent on such a policy even after the framework of our social security system was well and truly put in place by the Labor governments of the mid-1940s. Liberal opposition to Medibank — later Medicare — suggests how much Liberals adhered to a model of private, contributory insurance, subsidised by government in a way that largely benefited high income earners. There was a social media meme going around last year quoting Menzies in parliament in March 1944, pronouncing that people should be able to obtain social security benefits ‘as a matter of right’, with no more loss of self-respect than ‘collecting from an insurance company the proceeds of an endowment policy on which they have been paying for years’. The idea was to show up the Morrison government’s punitive approach to welfare, in particular JobSeeker payments, as betraying some sort of Menzian tradition. But it missed the point. Menzies’ comments entirely captured the Liberal Party’s distaste for non-contributory, redistributive, tax-funded social security payments.

In all this, the nature of the ‘unemployment’ risk — that extension of actuarial logic from health and mortality into the quotidian world of lost work and lost wages — proved particularly intractable across the interwar period. The frequency of joblessness inevitably varied between industry sectors, and depended on management practices, unionisation, product markets and so on. If ‘unemployment’ meant more or less absolute inactivity in response to a cyclical downturn, then an insurance scheme would only gain purchase in those trades where lay-off and redundancy were the normal response to slack demand, rather than reduced hours or variations in the use of casual and intermittent workers. Indeed, the British social insurance legislation did initially limit itself to workers in construction, shipbuilding, engineering and the metal trades: that is, those trades where ‘the line between being in or out of work was fairly clear cut’.[2]

The Bruce–Page government appointed a Royal Commission to examine the possibility of an unemployment insurance scheme in the mid-1920s. The Commission ran up precisely against that problem as to which groups of workers to include in the scheme. It suggested a distinction be made between seasonal unemployment which was the normal characteristic of many trades, and which was partially compensated for with casual loadings in award wages, and unemployment ‘over and above this’. The latter was a matter for insurance, but not the former. The alternative would be to revise awards to remove loadings for intermittent work and to include the entire workforce in such a scheme, with the exception of those intermittent and casual workers who were habitually excluded from such schemes operating overseas.[3] While it was recognised that ‘employers who give, and employees who receive, constant employment could not equitably be included in the scheme’, the issue of short-time or work rationing, common in such trades as clothing and textiles as a response to trade fluctuations, was left unaddressed: would these workers be forced to pay contributions while rarely being in a position to claim benefit because management practices meant they were rarely ‘unemployed’ according to most insurance definitions? Nevertheless, the Commission perfunctorily recommended, and roughly costed, a scheme to insure against those risks of unemployment ‘found to be unavoidable’.

But when the government introduced a National Insurance Bill in 1928 following on from the Commission’s reports, it contained no provision for unemployment insurance, concentrating instead on sickness, disability, old age and widowhood. José Harris observes in the British context that while the problems surrounding unemployment insurance were largely technical, those around sickness and disability insurance were overtly political.[4]  The latter tended to draw the ire of vested interest groups, in particular doctors and friendly societies. Within 12 months, in the face of opposition primarily from employer groups, the bill was shelved.[5]  John Murphy links employer opposition to the bill with the existence of the arbitration system: employers were worried that workers’ contributions to the scheme would end up being taken into account in the cost of living adjustments to wages, so in effect the scheme would simply end up being financed largely by industry.[6]  Ten years later, the Commonwealth government again introduced similar legislation for a social insurance scheme, and again it excluded any form of unemployment insurance. This time resistance to the bill came from the Labor opposition as well as the medical profession, women’s groups and the United Australia Party’s coalition partner, the Country Party. The legislation was again abandoned.

Labor had signalled its opposition to contributory schemes as far back as its 1912 national conference—which makes it odd that the only actually implemented social insurance scheme came from a Queensland Labor government, enacting an unemployment insurance scheme in 1922. The 1922 legislation is best seen as a compromise after a more radical 1919 bill failed to get through the Queensland upper house. That earlier proposal was to be funded solely by a levy on employers, which would have resulted in a genuinely radical and redistributive outcome, consistent with Labor policy, but it also spelled the legislation’s political death-knell. The 1922 legislation excluded striking workers along with professional and high-income workers, but was designed to specifically include seasonal workers in the sugar and pastoral industries as well as dock workers. Receipt of benefit could be suspended where a jobseeker refused to accept an offer of work, refused to join a union as a condition of accepting work, or became unemployed through their own fault. Benefits were flat-rate rather than earnings related, varied according to region, made allowance for dependants, and were capped at half the minimum award wage or the prevailing rate in the trade. The legislation also included provisions relating to the funding of public works and labour farms. In short, it more or less replicated the 1919 bill except in two significant ways that defused employer opposition: benefits were now to be funded by way of equal tripartite contributions from employers, employees and the government; and receipt of benefits was capped at 15 weeks in any one year.[7]

Mulino rightly recognises that at one level it is easy to overstate the differences between ‘social insurance’ models of welfare and Australia’s ‘social assistance’ model. Without delving very deeply into the history, he does note that the parliamentary Joint Committee on Social Security, convened in the first years of the Second World War, managed to sidestep the impasse by arguing in effect that there was little material distinction between contributions on the one hand and progressive income taxes on the other, with claimants in the latter case having ‘earned’ their benefits through the taxes they paid. This meant extending income taxes down the income scale so that those likely to receive the new social security benefits were being asked, for the first time, to pay for them, and to pay for them via a tax based largely on participation in the employment relationship. So in the space of a single paragraph, the Committee could refer to ‘the obligation of all the potential beneficiaries to contribute to the scheme’ and then go on to propose that ‘the simplest and most equitable plan in the present circumstances is to impose a [graduated] general tax on every income-earner in the community’.[8]

The Committee’s Second and Third reports reiterated the shortcomings of established social insurance schemes. Those reports were mostly written by the Committee’s research officer, Ronald Mendelsohn. An economist who had worked for a major bank and an appointee of UAP Minister Stewart, Mendelsohn was initially viewed with suspicion by Labor members of the Committee. But Mendelsohn had definite socialist sympathies and was happy to make the case for a non-contributory scheme of social security.[9]

The reports relied on only a handful of submissions from within the public service, most notably from Jim Nimmo, a junior Treasury official.[10] Nimmo proposed the idea of a ‘national minimum of real income’: that is, a minimum limit below which no citizen need fall in the advent of adverse circumstances. It was to be made up of cash transfers and financed from a progressive income tax. The resulting cash payment would be means-tested and set considerably below prevailing wage rates. In this last aspect, the proposal mirrored the old British poor laws, whereby the recipient of relief was not to be ‘made really or apparently so eligible as the situation of the independent labourer of the lowest class’.[11] The Committee ended up recommending a weekly payment of 80 shillings for an unemployed adult male who was supporting a wife and one child. The basic wage at that time, notionally meant to cover the same type of household, was 86 shillings a week, so Nimmo’s suggestion regarding a parsimonious payment seems to have gone unheeded by the Committee. In the end, though, the government would legislate for a far lower rate of 50 shillings a week for such a household, which amounted to less than 60 per cent of the basic wage. It was ‘more than New Zealand provided, a little less than that Beveridge had proposed recently in the United Kingdom’.[12]

The Committee’s Third Report, published on 25 March 1942, affirmed the merits of a non-contributory scheme, financed from a progressive income tax.[13]

The ALP government went further, doing an accounting fiddle by setting up the National Welfare Fund during World War II and then, in 1945, legislating for an earmarked social services income tax contribution and a pay-roll tax on employers to pay for the Fund, with all pensions and benefits becoming a charge on the Fund. It gave the mirage of a tripartite contributory scheme, but it remained a mirage: eligibility and entitlement were not linked to contributions, nor were pensions and benefits free of a relatively strict means test, and they continued to be financed from consolidated revenue — as the aged and invalid pension had been since 1908.

Many studies of the Australian tax-transfer system show that as a redistributive tool, it is capable of alleviating poverty, through what is, in essence, a ‘Robin Hood’ strategy of taking money from well-off households and giving it to poor households, taking money from households with no children and giving it to households with children, and taking money from working age households and giving it to aged households.

Put in this way, as a simple point-in-time analysis, it’s easy to support an ‘us-and-them’ outlook, feeding further resentment toward welfare recipients and undermining broad-based political support for the system. If you think about it, it in fact suggests that many people will move from being net contributors to the tax-transfer system to net beneficiaries depending on life circumstance. A good proportion of the redistribution effected by the Australian system can’t help but be redistribution of income across individuals’ own lifetimes as many of the risks that the system compensates for are fairly strongly associated with particular life course events — ageing, in particular, and also the costs associated with having children.

Similarly, many other countries besides Australia place importance on the alleviation of poverty regardless of income smoothing over the life course. This means that although ‘classic’ insurance models favour lifetime redistribution over interpersonal redistribution, they by no means eliminate it, using things like caps on earnings-related payouts and contribution credits for low income earners.

And in any case, a stylised view of Australian arrangements as conforming to a social assistance or ‘Robin Hood’ model can seem too restrictive. Frank Castles famously used the notion of ‘path dependency’ to suggest Australia for decades was stuck in the pattern established in the first decade of the twentieth century: flat-rate payments, means-tested, and paid out of consolidated revenue. But it is worth remembering that the full panoply of income maintenance and welfare schemes in Australia exhibits a greater diversity than this model suggests. State-based systems of workers compensation, which remain Australia’s earliest income maintenance schemes, follow a social insurance model based on contributary financing and offering earnings replacement rather than flat rate benefits. Medicare follows a quasi-insurance model based on a nominal ‘levy’. Both superannuation and the higher education contribution scheme follow a compulsory savings model—the first prospective, the second retrospective.

Again, this hybridity is unsurprising. As Julia Perry has noted, the reasons for people’s less-than-full participation in the labour market can fall into four broad categories. There are those normal, predictable and universal barriers (such as childhood and old-age). There are those sanctioned and valued non-market activities that can keep us away from paid work, such as child rearing and other forms of caring. Then there are those times we dip out of the labour market to improve our employability and productivity (having lost one job, we spend time looking for another that can make use of our skills — what economists call frictional unemployment — or we return to education and training in order to get a better job). And there are those undesired misfortunes such as sickness, disability and long-term unemployment.[14]

Mulino’s book offers a comprehensive, albeit sometimes wonkish, survey that suggests direct transfers through a generous social assistance scheme or social insurance or individual savings accounts or contingent loans all represent different strategies that can be more or less appropriate for any given set of risks. Any social welfare system is necessarily a complex amalgam of decisions as to what risks are insurable, who to compensate, and how to pay for it all. And if close attention to our welfare history reveals Australia as more a mongrel welfare state than a pedigree, then the future of welfare in Australia seems a lot more interesting.



[1].       Pat O’Malley, Risk, Uncertainty and Government (Glasshouse Press, 2004) pp 30, 32.

[2].       Noel Whiteside and James Gillespie: ‘Deconstructing Unemployment: Developments in Britain in the Interwar Years’ Economic History Review, Vol. 44, No. 4, 1991, pp 665, 675.

[3].       Commonwealth of Australia, Royal Commission on National Insurance, Second Progress Report: Unemployment, Parliamentary Paper No. 79 (1926–28), pp 22–23.

[4].       Jose Harris, William Beveridge: A Biography (Clarendon Press, 1977) p 169.

[5].       For a detailed account, see John Murphy, A Decent Provision: Australian Welfare Policy, 1870 to 1949 (Routledge, 2016); Rob Watts, The Foundations of the National Welfare State (Allen & Unwin, 1987); TH Kewley, Social Security in Australia 1900–72 (Sydney University Press, 1973) pp 143–49.

[6].       John Murphy, ‘Path Dependence and the Stagnation of Australian Social Policy Between the Wars’ Journal of Policy History, Vol. 22, No. 4, 2010, pp 450, 458.

[7].       Passage of the 1922 legislation was also made easier by the abolition of the Queensland Upper House (the Legislative Council) in that year. Voted out of existence by a largely Labor-appointed majority of Legislative Councilors, the Council met for the last time in 1921.

[8].       Commonwealth of Australia, Joint Committee on Social Security, Second Interim Report, Parliamentary Paper No. 71 (1940–43) para 12, emphasis added.

[9].       Stuart Macintyre, Australia’s Boldest Experiment: War and Reconstruction in the 1940s (NewSouth, 2015) pp 179, 205.

[10] .    Watts, p 76.

[11].     Report of the Poor Law Commission (1834), cited in Simon Deakin and Frank Wilkinson, The Law of the Labour Market: Industrialization, Employment, and Legal Evolution (Oxford University Press, 2005) p 134.

[12]      Macintyre, p 207.

[13]      Commonwealth of Australia, Joint Committee on Social Security, Third Interim Report, Parliamentary Paper No. 72 (1940–43) para 25.

[14].     Julia Perry, ‘Alternative Sources of Welfare’ Australian Bulletin of Welfare, Vol. 32, No. 3, 2006, p  280.

Anthony O’Donnell
Anthony O’Donnell

Anthony O’Donnell is an adjunct senior lecturer in the La Trobe University Law School. He researches and publishes in the areas of Australian labour law, labour history and social policy. His most recent book is Inventing Unemployment: Regulating Joblessness in Twentieth-Century Australia, Hart-Bloomsbury, 2019.