Removing the “Long-Service” from Long Service Leave

I recently left a job I’d held for the better part of a decade. Looking back, I can see that as I approached the point where my long service leave would become available, my plans to move into a new career lost their urgency. It seemed foolish to walk away from a large benefit I had spent years accumulating. If I held on for another year, I would be entitled to a significant chunk of paid time off.

That sounded like a pretty sweet deal, but the rational part of my brain argued that it wasn’t worth delaying a career transition that would be far more satisfying, would be a better professional fit, and would more than outweigh this loss of leave. It was this moment of hesitation that made me wonder how many people subtly arrange key moments in their working lives around their long service leave. Many readers will have had a conversation like this with a friend, parent, or colleague—or have weighed this tradeoff themselves.

Long service leave entitlements are substantial. A worker who reaches the ten year vesting threshold receives nearly nine weeks of paid leave at that point, worth $18,100 at Australia’s average full-time earnings of $108,400. A worker who exits before vesting generally forfeits it entirely. The annual accrual amounts to an additional 22 per cent on top of the standard four weeks of annual leave. Over a 40-year working life, an average-earning employee who happens to spend their career in several long-tenured positions can accumulate up to $72,000 more in paid leave than an employee who moves more often.¹ But much of the long service leave Australians accrue is never received, because it isn’t structured for how we work today.

The case for reforming long service leave rests on two arguments.

The first argument is about fairness. Standard economics suggests that mandated benefits are funded by workers themselves through lower wages, rather than by employers through reduced profits.² Only those who reach the threshold receive the leave that those reduced wages paid for. Those who exit before reaching it are not simply missing out on a benefit; they are cross-subsidising the benefit for those who stay, and the data suggests women are more often on the losing side.

The second argument is about productivity. Long service leave functions as a shadow tax on job switching: workers approaching the threshold stand to forfeit a substantial entitlement if they leave, and incur an opportunity cost if they stay only to collect. I hypothesise that this discourages job switching at the margin, which matters because the evidence indicates that labour mobility raises productivity in aggregate.

There are several ways we could address these issues, such as replacing the leave with extra annual leave, or abolishing it entirely. I make the case for removing the “long-service” from long service leave, and simply enabling workers to access the leave over the course of their employment. This option would preserve the ability to save up for a well earned sabbatical, provide fairer access to time off for all workers, and create a more dynamic labour market in which a portion of workers no longer stay in jobs to hit an arbitrary milestone that delays the next chapter of their careers.

Long service leave is a colonial artefact

Long service leave in Australia exists thanks to a quirk of our colonial past. It is, as the Productivity Commission has called it, an “Antipodean idiosyncrasy”, though one that “now has strong community support”.

In the nineteenth century, when Australia was still a collection of British colonies, long-serving public servants were granted extended paid leave so they could visit their homes in Britain by sailing ship. It evolved from there, and this colonial furlough was extended around the time of Federation to all public servants after 20 years’ work as a “reward for their diligence”. In the 1950s, all states extended long service leave to the private sector, often with justifications we would find strange today: the NSW Minister for Labour and Industry extended long service leave with the explicit goal of reducing labour turnover, “which today is costing employers a considerable waste of effort and money”. One Western Australian MP railed against a proposed shorter version of a long service leave scheme, arguing that a “more static market” would be a desirable state of affairs.

Although the thresholds have fallen over time, the concept remains: work for the same employer for an extended period—between seven and ten years depending on the state—and you receive a substantial block of paid leave with ongoing accrual thereafter. But if you switch employers before the threshold, you typically forfeit it.³ Australia appears to be the only country in the world with a universal statutory entitlement of this kind, and, given these origins, we shouldn’t be surprised that long service leave is a poor fit for the modern economy.

Most Australians are unlikely to receive their long service leave

The ABS tenure distribution shows that the typical Australian worker has been with their employer for well short of the current vesting thresholds, with only around one in four having reached the ten-year mark.⁴ These thresholds bear little relationship to how the average Australian works.

Graph 1: Australian workforce tenure at current employer

Share of employed people, by time with their current employer

• Less than 1 year16.4%
• 1–4 years39.3%
• 5–9 years18.7%
• 10–19 years15.5%
• 20 years +10.0%

Long service leave is not a single national institution. It is eight overlapping, inconsistent regimes. Vesting thresholds range from seven to ten years.⁵ Most jurisdictions also pay a pro-rata amount earlier, typically from five to seven years, but the rules vary: Western Australia and South Australia pay it on almost any exit, while the Australian Capital Territory, New South Wales, Queensland, the Northern Territory and Tasmania pay it only under specific conditions, such as illness or employer termination, the latter of which is intended to deter bosses from strategically dismissing workers just before they vest. What matters here is the “forfeit cliff”: the point after which a worker can leave voluntarily without losing their entitlement.

Long service leave allocates leave unfairly

Long service leave distributes a communally funded quantity of paid leave to a lucky subset of workers, on a basis that is at best arbitrary, and at worst actively perverse. Those who change employers or have a career break earlier than their jurisdiction’s forfeit cliff bear the cost of an entitlement that other workers receive. If this state-mandated loyalty scheme were in service of some genuinely equitable or productive outcome, this could be considered a justifiable means to an end. But the system as it works today is more likely to work against these goals than in service of them.

Across 12 years of data in the four jurisdictions that have a 10 year forfeit cliff, women consistently reach the 10-year tenure milestone at a lower rate than men: a persistent, statistically significant gap averaging roughly 2.5 percentage points. As of February 2026, the rates of currently employed workers stood at 24.9 per cent for women compared with 26.9 per cent for men.⁶

Graph 2: Men and women who have reached ten or more years with their employer

Share of employed people, pooled for the jurisdictions with a ten-year forfeit cliff (NSW, Qld, Tas, NT)

MenWomen

This gap is likely downstream of the uneven range of factors that impact men and women’s tenure differently. For example, leaving a job for family reasons accounted for 14.6 per cent of women’s job exits but only 3.9 per cent of men’s.⁷ While a 2.5 percentage point gap may seem small, had women’s tenure matched men’s, around 75,000 more women in these four jurisdictions would have access to this entitlement as of February 2026, with the shortfall averaging around 80,000 over the past 12 years. Across all jurisdictions, we should expect this number to be even higher.

Long service leave puts a price on switching jobs

The current design of long service leave presents workers with a trade-off between two distinct costs that could influence their decisions about job switching: forfeit cost, incurred by workers who exit before the forfeit cliff; and opportunity cost, incurred by workers who choose to stay.

The forfeit cost for workers who leave

Workers who choose to exit before their long service leave forfeit cliff lose the entitlement they have spent years accruing. As Table 1 above shows, these are significant figures, and behavioural research suggests the distortion is even larger than the raw financial figures would predict. Losses are typically felt more acutely than equivalent gains.⁸ A worker who has mentally accounted for their accrued leave would not experience forfeiting it as a neutral trade-off, but as losing something that’s already theirs. Even a worker who leaves a job for a new role with a pay rise that more than compensates for the forfeited value has still lost something that they implicitly paid for, cross-subsidising the value for their colleagues who stay. And the effect does not end once a worker has vested: moving resets the clock, so the leave they accrue in a new job is at risk until they reach the threshold again.

The opportunity cost for workers who stay

The closer a worker is to vesting, the larger the financial and psychological loss they would incur by leaving, and the more powerful the incentive to stay even when a better opportunity is available. The opportunity cost is the difference between the worker’s actual trajectory and the trajectory they would have followed had the entitlement not exerted an influence: a higher-paying role declined, a more interesting job not pursued, a career change deferred. Strictly, this is a loss to the worker only when the pull of the entitlement outweighs its true value, as loss aversion suggests it can. But there is still a cost when the entitlement’s value makes staying the better private deal for the worker; it just falls instead on the wider economy, which forgoes a potentially more productive match.⁹

This is a straightforward case of people responding to incentives, and it is well documented empirically in a closely analogous setting: employee stock options, which share long service leave’s threshold vesting structure. Value accrues over time and an early departure forfeits what has not yet vested. If a forfeit cliff distorts the timing of departures, stock options are where that distortion should be visible, and it is.¹⁰

The economic case for mobility

The economic case for mobility rests predominantly on efficient labour market matching. When workers can move freely towards their highest-value job, skills flow through the economy as workers move between organisations and employers compete for talent through pay and conditions. The evidence that job switching delivers these gains in Australia is now reasonably well established.¹¹

Two caveats are necessary. First, not all switching is productivity-enhancing. Research using administrative microdata finds that almost half of job changes involve workers moving to less productive employers, often for legitimate personal reasons such as shorter commutes or better work-life balance.¹² The productivity premium is an average, not a guarantee. Second, mobility has costs as well as benefits, including disrupted teams and onboarding friction. The argument is not that individual mobility always increases productivity, but that Australia’s current level is below the economically optimal level. The aim is to remove the bias against moving, not to encourage movement for its own sake. For most workers the role they already hold is the most productive one available to them, given the specific knowledge and networks built up over time, and the proposed reform would do nothing to disturb that.

We can remove the “long-service” from long service leave

Reforming long service leave is relatively simple: remove the vesting threshold, such that pro-rata accrual of long service leave can be taken from the first day of employment, rather than after seven to ten years. The accrual rate, administration, and the way the leave is taken would all remain as they are. The one thing that stops fitting is the name: an entitlement that is available from the first day of employment is no longer leave for “long” service, though that is a labelling question rather than a substantive one. What changes is that a worker of any tenure can access the time off they have accrued, and will no longer lose it if they switch employers. Instead, it would be paid out pro-rata to all workers upon exit, exactly as it is for workers who reach the threshold under the current system. A gradual phase-in approach can resolve the concern that balances accrued but not yet vested would become payable at once, causing a one-off step up in liabilities for all employers.

Averaged across the whole workforce, workers would get just over three extra days of leave a year, without having to stick around with a single employer for a decade, and without losing out if they change jobs.

There is, of course, a productivity concern here: it would mean that Australians will on average work less. If everyone is given access to their long service leave from day one of employment, Australians will work roughly 1.4 per cent less: 226 days per year instead of 229¹³. However, a 1.4 per cent reduction in working days does not translate to a 1.4 per cent reduction in output. The relationship between time worked and output is non-linear; the hours forgone are likely to be among the less productive ones at the margin.

Graph 3: Total working days per year

Net of statutory annual leave, public holidays and, for Australia, workforce-weighted long service leave

Australia nowAustralia under the proposalOther countries

Median 228

• Austria
• Malta
• Cyprus
• Luxembourg
• Denmark
• Finland
• France
• Iceland
• Spain
• Sweden
• United Kingdom
• Slovak Republic
• Australia (Proposed)
• Portugal
• Germany
• Croatia
• Czech Republic
• Romania
• Bulgaria
• Italy
• Latvia
• Lithuania
• Poland
• Slovenia
• Australia (Current)
• Estonia
• Greece
• Hungary
• New Zealand
• Norway
• Belgium
• Chile
• Korea
• Ireland
• Japan
• Netherlands
• Switzerland
• Turkey
• Costa Rica
• Israel
• Canada
• Mexico
• United States

220225230235240245250

This change would not make us an outlier among peer nations. Under the proposed reform, Australia would move from slightly above the OECD median of 228 days to slightly below it. This is hardly a seismic shift in days worked, and one worth making for the fairness and mobility gains.

Why not replace long service leave with extra annual leave?¹⁴

Australians have had this entitlement for their entire working lives, and are accustomed to a kind of leave that annual leave does not provide in all cases. The distinction is control. Employers can direct that annual leave be taken to run down the liability on their balance sheet or to align with forced shutdown periods, and it is usually taken in fragments across the year.

Long service leave, on the other hand, typically accumulates, with employees having more say over when to take it, allowing for a sustained uninterrupted break. A reform that folded it into additional annual leave would potentially strip this character away, not as a neutral simplification but as the loss of something workers currently have and value. There is a second problem with a straight swap. For workers whose time off is already structured through a larger scheme, additional annual leave can simply be absorbed into the existing arrangement and deliver nothing. Teachers are the clearest case: a teacher has the same four weeks of annual leave as anyone else, but it is pinned to the school vacation periods, and their longer holidays are paid non-attendance scheduled around the school calendar rather than leave the teacher can deploy at will. A teacher who surrendered long service leave for extra annual leave folded into that calendar would give up a real entitlement for a notional one, with nothing to show for the exchange. Retaining long service leave as a distinct, standalone entitlement avoids both problems. It keeps the leave under the worker’s control, and it ensures the entitlement cannot be subsumed by whatever system already governs a worker’s time.

Why not abolish long service leave altogether?

The purest economic response would be to abolish the policy and return its value to workers as higher wages, since a benefit that must be mandated is, absent some market failure, one that workers value at less than its cost.¹⁵ But that value would return slowly and invisibly—through faster wage growth rather than an explicit rise—and guaranteeing it would mean mandating a pay increase, a difficult and clumsy intervention. Abolition would also diminish the potential for workers to take the kinds of long, self-directed breaks discussed above.

The impetus for reform

As for me, I ended up staying in my job for another few years. I took my long service leave, had a wonderful sabbatical, and have since moved on to pursue an exciting career change. I wouldn’t say that long service leave was the single deciding factor for why I stayed, but I certainly felt its pull.

The concept of long service leave is a humane idea: a long working life deserves genuine breaks. That idea is worth defending. But the current design is unfair to the millions of workers who implicitly pay for long service leave but don’t get to enjoy it, and the scheme as it exists may be quietly holding a portion of workers back from moving to better jobs. The simple fix is to remove the one feature responsible for both faults—the vesting threshold—and keep everything else. The states and territories have the power to do this: it simply requires an amendment to legislation they already administer. In the 1950s, legislators wanted a more static labour market. Seventy years on, we are wise enough to know that’s the last thing our nation needs.

References

Footnotes

1. These figures assume the most common accrual of 0.867 weeks per year of service (8.67 weeks at the ten-year threshold) and average full-time earnings of $108,400. The entitlement at vesting is worth about $18,100 (8.67 ÷ 52 × $108,400); the annual accrual is roughly 22 per cent of the four weeks of standard annual leave; and over a 40-year career uninterrupted by forfeiture it totals about 35 weeks, or $72,000. ↩

2. Summers (1989) sets out the theoretical framework. Bozio, Breda, Grenet & Guillouzouic (2026), in a meta-analysis of international evidence on social security contributions, find full pass-through to wages where there is a strong and salient link between contributions and the worker’s expected benefit, and substantially lower pass-through where no such link exists. LSL fits the high-linkage category: the entitlement accrues to the workers, and is received as paid leave or a cash payout rather than flowing into general revenue. The theoretical prediction is therefore full pass-through, meaning the vesting weighted cost of LSL is borne by workers through correspondingly lower wages rather than by employers through reduced profits. ↩

3. Some industries with high levels of subcontracting, such as construction and cleaning, have portable long service leave schemes, which pay out based on time worked in an industry, rather than with a single employer. These complex schemes are typically managed by specific government authorities. ↩

4. Australian Bureau of Statistics (2026b). The ABS Job Mobility data only records the tenure of people who are still employed at the moment of the survey. It captures a snapshot rather than the completed tenure of all workers, so the published distribution, which displays stable rates over the medium term, only implies what full-length tenure looks like, rather than stating it directly. The heading refers to individual jobs, it does not mean most Australians never receive LSL at some point in their careers. ↩

5. Fair Work Ombudsman (2025), summarising the state and territory schemes. ↩

6. This was calculated by pooling the tenure data for the jurisdictions that have a 10 year forfeit cliff (NSW, QLD, TAS, NT). SA and WA also have a 10 year threshold but they pay out LSL for most reasons after 7 years of service upon exit. Due to the way the data is aggregated by the ABS, only showing the totalled 5 to 9 year bracket, it was not possible to measure the 7 year forfeit cliff jurisdictions (ACT, SA, VIC, WA) against 7+ years with their employer. ↩

7. Australian Bureau of Statistics (2026b), Chart 16, “Main reason left or lost a job, by sex.” ↩

8. Tversky & Kahneman (1992). Their formulation of cumulative prospect theory estimates a loss-aversion coefficient of roughly 2.25, so a loss is weighed about twice as heavily as an equivalent gain. ↩

9. The claim that the forfeit cliff distorts job switching has not been tested directly using Australian data. Measuring it requires longitudinal data that follows the same workers over time. If it deters departures in the approach to vesting, that should appear in the data as a discontinuity: a measurable dip in employer separations among workers in the lead up to their vesting threshold. The states also offer a natural experiment due to the difference in forfeit cliffs. This is research worth commissioning to determine if the hypothesised distortion appears in the data, and if so, how large it is. ↩

10. Chou & Shiu (2024), studying a firm whose options vested in equal annual batches, found turnover bunched sharply on the vesting dates, as workers who would otherwise have left earlier waited until the forfeit cost dropped. Aldatmaz, Ouimet & Van Wesep (2018), examining broad-based option grants across a large panel of firms, found the same pattern over a longer horizon: turnover fell while options vested, then rose once they had, and the eventual increase matched the earlier decline almost exactly. ↩

11. Buckley (2023); Wong (2024). Research from the e61 Institute found that workers who switch jobs move to employers that are on average 13.1 per cent more productive than those they leave. Job switchers also secure larger pay rises, with e61 estimating the gain at around 9 percentage points more than staying put. These findings cannot quantify the effect of LSL reform, because the switches it would unlock are driven by different underlying forces, but they do show that moving between firms is a real source of productivity gains. ↩

12. Albagli et al. (2025), using Chilean administrative microdata. ↩

13. OECD (2020) and the author’s calculations incorporating LSL. Net of 20 days of annual leave, roughly 10 public holidays depending on the jurisdiction, and roughly one day a year of LSL when weighted across the workforce, using the rough estimate that only about a quarter of accrued time vests, based on current tenure data. ↩

14. Readers may be interested in Jeff Borland’s article discussing the surprisingly low cost of Australia adding a fifth week of annual leave. ↩

15. Summers (1989): where pay is freely negotiated, benefits are provided up to the point where each dollar spent is worth a dollar to workers, so a mandate binds only where they value the benefit below its cost. Summers uses a leave policy as his example, and identifies paternalism, that people undervalue some benefits, as the leading case for mandating them regardless. ↩

Credits

Thank you to Jonathan O’Brien, Editor-in-Chief of Inflection Points, for his careful editing of this essay.

Thank you to Professor Jeff Borland AO, Redmond Barry Distinguished Professor & Truby Williams Professor of Economics at the University of Melbourne, for reviewing an earlier draft and for his thoughtful comments.