Dissertation – The Gig Economy and Worker Wellbeing in Australia

Abstract

The gig economy has become a durable feature of Australian labour markets, yet its consequences for worker wellbeing are unevenly understood. Framed by the job demands-resources (JD-R) model, this dissertation examines how platform work shapes the wellbeing of Australian gig workers and the mechanisms that explain differences across platform segments. An explanatory sequential mixed methods design combined a survey of 353 platform workers with 16 semi-structured interviews. Wellbeing varied significantly by segment: food-delivery riders reported the lowest alongside the highest income insecurity and algorithmic control, while online freelance workers reported the highest alongside the greatest autonomy. Regression indicated that income insecurity and algorithmic control, rather than autonomy, drove wellbeing. Interviews identified earnings anxiety, algorithmic opacity and an autonomy paradox in which nominal flexibility masked intensification. The study argues that wellbeing depends less on flexibility than on the demands platforms impose, with implications for recent Australian minimum-standards reforms.

Introduction

Platform-mediated work has become a durable feature of the Australian labour market rather than a passing experiment. Estimates prepared for Australian regulators suggest that between 7 and 9 per cent of working-age Australians have earned income through a digital labour platform, with a smaller core relying on it as a primary source of earnings (McDonald et al. 2019). This workforce is heterogeneous, spanning rideshare and transport, food delivery, care services and online freelance work. The Australian Bureau of Statistics has documented the parallel growth of secondary jobs and non-standard arrangements that platform work both reflects and accelerates (Australian Bureau of Statistics 2023). Platforms are promoted on the promise of flexibility and autonomy, yet the same arrangements concentrate income volatility, safety risk and managerial control in ways that may undermine the wellbeing they promise.

Australian policy has begun to respond. The Fair Work Legislation Amendment (Closing Loopholes No. 2) Act 2024 (Cth) created a new category of employee-like worker and empowered the Fair Work Commission to set enforceable minimum standards for eligible platform workers (Fair Work Commission 2024). These reforms followed sustained advocacy, including evidence from the Transport Workers’ Union to a parliamentary inquiry that documented low and unpredictable earnings among on-demand workers (Senate Select Committee on Job Security 2022). The Actuaries Institute has separately warned that gig workers face acute income volatility and gaps in income protection and superannuation (Actuaries Institute 2020), and platforms and their engagers carry a duty to manage foreseeable psychosocial hazards at work (Safe Work Australia 2022). The architecture is moving quickly, yet the evidence on how platform work affects wellbeing, and why it differs across segments, remains thin in the Australian setting.

This dissertation addresses that gap. Using the job demands-resources (JD-R) model, it examines how platform work shapes the wellbeing of Australian gig workers and the mechanisms producing differences between segments. Three research questions guided the study:

1. How does self-reported wellbeing differ across platform-work segments among Australian gig workers?
2. Through which job demands and resources do platform arrangements influence wellbeing?
3. How do workers experience the autonomy that platform work offers, and with what consequences for their wellbeing?

Literature Review

Platform work and wellbeing in Australia

Platform work sits at the precarious edge of the Australian labour market. Its defining feature for wellbeing is income insecurity: earnings fluctuate with demand, surge pricing and unpaid waiting time, and workers absorb vehicle and equipment costs, so take-home pay is frequently below expectations and hard to predict (Actuaries Institute 2020). Australian evidence indicates that many platform workers earn below the national minimum wage once costs and idle time are counted, and lack leave, workers’ compensation and superannuation (McDonald et al. 2019). Income insecurity is not merely financial but psychological, generating chronic earnings anxiety that the parliamentary inquiry into job security heard described as a constant background stressor (Senate Select Committee on Job Security 2022). For food-delivery riders, this precarity is compounded by physical exposure on the road.

The job demands-resources model

The job demands-resources model offers a parsimonious account of how such conditions translate into wellbeing. A health-impairment pathway runs from job demands, the aspects of work that require sustained effort, towards exhaustion and diminished wellbeing, while a motivational pathway runs from job resources, the aspects that help workers achieve goals, towards engagement (Bakker & Demerouti 2017). The model treats the balance between demands and resources as decisive, and recognises that an apparent resource can operate as a demand when poorly bounded. Applied to platform work, income insecurity and algorithmic control are readily theorised as demands, while autonomy over hours is the resource most often invoked in its defence.

Algorithmic management as a job demand

Algorithmic management is the distinctive job demand of platform work. Rather than by managers, work is directed by software that allocates tasks, sets prices, monitors performance through ratings and location data, and can deactivate workers with little explanation or appeal (Rosenblat & Stark 2016). Duggan et al. (2020) characterise this as app-work, in which the relationship is mediated almost entirely by an application that exercises control while disclaiming an employer’s obligations. The experience is marked by information asymmetry: workers cannot see how allocation or pricing decisions are made, yet are continuously ranked against opaque standards (Wood et al. 2019). This combination of intensive monitoring and low transparency maps onto the JD-R construct of a demand and is theorised here as a primary driver of strain.

The autonomy paradox

The autonomy that platforms advertise is more ambiguous than it appears. Workers can, in principle, choose when to log on, inviting the conclusion that platform work is a high-resource arrangement. Mazmanian et al. (2013) show, however, that autonomy over connectivity can be self-undermining: given control, workers extend rather than contain their effort, so autonomy quietly produces intensification. In the platform context this paradox is sharpened by income insecurity, because the freedom to choose hours coexists with economic pressure to work long and at peak-risk times to reach earnings targets (Wood et al. 2019). Nominal flexibility can therefore mask a loss of genuine control, so autonomy should predict wellbeing less strongly than the demands that constrain how it is exercised.

Conceptual framework

Drawing these strands together, the study advances a conceptual model in which platform-work characteristics act as job demands and resources that shape wellbeing through the two JD-R pathways, with the autonomy paradox as a feedback loop from nominal autonomy to intensification under income insecurity. Figure 1 presents this model. Two gaps motivate the empirical work: the demands and resources of platform work are more often asserted than tested together within a single model, and little of the evidence is grounded in the Australian regulatory and market context despite the recent minimum-standards reforms. The study addresses both.

Methodology

Research design

An explanatory sequential mixed methods design was adopted, in which a quantitative phase is followed by a qualitative phase that helps to explain the statistical results (Creswell & Plano Clark 2018). The quantitative phase established how wellbeing differs across segments and which demands and resources predict it; the qualitative phase examined how workers experience those conditions, and in particular the autonomy identified in the conceptual model (Figure 1). Integration occurred through purposive selection of interviewees and joint interpretation of statistical predictors alongside interview themes.

Participants and procedure

The survey was completed by 353 Australian platform workers recruited through an accredited online research panel and platform-worker networks. Eligibility required that respondents had earned income through a digital labour platform in the preceding three months. Respondents were grouped by their main platform segment into food delivery (n = 96), rideshare and transport (n = 104), care and personal services (n = 71), and online freelance or professional work (n = 82). Of the sample, 61 per cent identified as men, the mean age was 34.8 years (SD = 10.2), and 44 per cent were born overseas. The qualitative phase involved 16 semi-structured interviews with a maximum-variation subsample selected to span all four segments, gender and reliance on platform income.

Measures and analysis

Wellbeing was measured with the five-item WHO-5 Wellbeing Index, scored from 0 to 100, a brief instrument with strong validation evidence (Topp et al. 2015). Income insecurity was assessed with a six-item scale capturing earnings unpredictability and difficulty meeting expenses, and exposure to algorithmic control with a five-item scale capturing monitoring, rating pressure and fear of deactivation, each scored from 1 to 7. Autonomy over scheduling used the autonomy subscale of the Work Design Questionnaire on a seven-point scale, and respondents reported their usual weekly platform hours. Cronbach’s alpha ranged from .79 to .90 across the multi-item scales. Quantitative analysis comprised descriptive statistics, one-way analysis of variance, Pearson correlations and hierarchical regression, and interviews of 40 to 55 minutes were analysed using reflexive thematic analysis (Braun & Clarke 2021).

Ethical considerations

The study was approved by the administering university’s Human Research Ethics Committee and conducted in accordance with the National Statement on Ethical Conduct in Human Research. Participation was voluntary and based on written informed consent, responses were de-identified at the point of collection, and interview participants were assigned pseudonyms and quoted without identifying detail.

Findings

Phase 1: Survey results

Table 1 summarises the measures by segment, and wellbeing followed the pattern predicted by the conceptual model. Food-delivery riders reported the lowest wellbeing together with the highest income insecurity and algorithmic control, while online freelance workers reported the highest wellbeing alongside the lowest demands and greatest autonomy; the rideshare and care segments occupied the middle. The difference across segments was statistically significant, F(3, 349) = 9.8, p < .001.

Using a pooled standard deviation, the standardised difference in wellbeing between the most and least demanding segments, food delivery and online freelance work, is

d = (61.3 – 48.2) / 17.9 = 13.1 / 17.9 = 0.73,

a medium to large effect. Notably, rideshare workers reported the longest weekly hours (38.7) despite only moderate autonomy, consistent with earnings-driven intensification rather than freely chosen flexibility.

Table 1: Income insecurity, algorithmic control, autonomy, wellbeing and weekly hours by platform segment (N = 353)

Note. Income insecurity, algorithmic control and autonomy are scored 1 to 7, with higher scores indicating more of each construct; the WHO-5 is scored 0 to 100 (Topp et al. 2015).

Correlational analysis clarified the mechanisms. Wellbeing was negatively associated with income insecurity (r = -.41, p < .001) and algorithmic control (r = -.33, p < .001), and positively but more weakly with autonomy (r = .27). Income insecurity was, in turn, associated with longer weekly hours (r = .24, p < .001), consistent with the autonomy paradox. In hierarchical regression, demographic controls and hours explained 6 per cent of variance, and adding income insecurity, algorithmic control and autonomy raised the model to R² = .24, F(6, 346) = 18.1, p < .001, with income insecurity (β = -.29) and algorithmic control (β = -.19) the strongest predictors and autonomy weak once demands were controlled (β = .11).

Phase 2: Interview themes

Reflexive thematic analysis produced five themes, summarised in Table 2. Income insecurity and earnings anxiety were near-universal, with participants describing constant mental arithmetic about fuel, hours and rent. Algorithmic management was experienced as opaque and unaccountable, captured in one rider’s remark that “the app decides everything and there is no one to ask” (Participant 7). The autonomy paradox was pronounced: participants valued the freedom to choose hours yet described working at dangerous times and in poor conditions to meet income targets, so that flexibility became, for one rideshare driver, being “free to choose, but not free to stop” (Participant 11). Safety and physical risk and an absence of collective voice completed the picture.

Table 2: Themes from reflexive thematic analysis of interviews (n = 16)

Discussion

The findings answer the first research question clearly: wellbeing differs significantly across segments, tracking the balance of demands and resources rather than platform work as a category. Food-delivery riders, facing the highest demands, reported the lowest wellbeing and online freelance workers the highest, consistent with the JD-R proposition that outcomes depend on the demand-resource balance rather than any single job feature (Bakker & Demerouti 2017).

The second research question concerned mechanisms. Income insecurity and algorithmic control were the dominant predictors of wellbeing, whereas autonomy was weak once demands were controlled. This positions the two demands, not the celebrated resource, as the primary drivers of strain, and supports the theorisation of algorithmic management as a genuine job demand rather than a neutral coordination mechanism (Duggan et al. 2020; Rosenblat & Stark 2016).

The third research question is answered by the convergence of the two phases. Statistically, autonomy predicted wellbeing only weakly while income insecurity predicted longer hours; qualitatively, workers described being free to choose but not free to stop. These findings substantiate the autonomy paradox: the freedom to set one’s hours is real, but under income pressure it is exercised towards more work, at riskier times, so nominal autonomy feeds back into the demand side as intensification (Mazmanian et al. 2013; Wood et al. 2019). Autonomy is thus a conditional resource whose value depends on whether workers can afford to exercise it.

These results speak directly to the Australian regulatory moment. The capacity of the Fair Work Commission to set enforceable minimum standards for employee-like workers under the Closing Loopholes reforms targets precisely the demand this study finds most damaging, income insecurity (Fair Work Commission 2024). Standards addressing minimum earnings, transparency of algorithmic allocation and fair deactivation would attenuate both leading demands identified here and align with the duty to manage foreseeable psychosocial hazards (Safe Work Australia 2022). The concentration of harm among food-delivery riders also underscores the interaction of financial and physical risk that the Transport Workers’ Union has emphasised in advocating for protections (Senate Select Committee on Job Security 2022).

Four limitations qualify these conclusions. The quantitative phase is cross-sectional, so the causal ordering of demands and wellbeing cannot be confirmed, and all measures are self-reported, which may inflate associations through common-method variance. The sample was recruited partly through panels and networks and may under-represent the most marginalised workers. Finally, segment was defined by main platform, which understates the multi-apping common in the sector.

Implications and Conclusion

Three sets of implications follow:

• Regulation and policy: minimum-standards orders under the Closing Loopholes framework should prioritise earnings predictability and algorithmic transparency, the demands most strongly associated with poor wellbeing, and give particular attention to food delivery (Fair Work Commission 2024).
• Platform design: platforms can reduce foreseeable psychosocial risk by making allocation and rating systems explainable, providing genuine avenues to contest deactivation, and discouraging fatigue across the working day, consistent with their psychosocial duty (Safe Work Australia 2022).
• Research: longitudinal designs and objective earnings data, ideally linked to Australian administrative sources (Australian Bureau of Statistics 2023), would strengthen the causal claims and track the effect of the new standards.

In sum, the wellbeing of Australian gig workers depends far less on the flexibility that platforms advertise than on the demands they impose. Income insecurity and algorithmic control, not autonomy, explained the wellbeing gradient across segments, and the autonomy on offer proved a conditional resource that, under financial pressure, tipped into intensification. Interpreted through the job demands-resources model, the task for Australian regulators and platforms is to reduce the demands that drive strain rather than celebrate a flexibility that many workers are not free to enjoy. Where minimum standards make earnings predictable and algorithms accountable, platform work can become a genuine resource; where they do not, its autonomy remains more apparent than real.

References

Actuaries Institute 2020, The rise of the gig economy and its impact on social insurance, Actuaries Institute, Sydney.

Australian Bureau of Statistics 2023, Working arrangements, ABS, Canberra.

Bakker, AB & Demerouti, E 2017, ‘Job demands-resources theory: taking stock and looking forward’, Journal of Occupational Health Psychology, vol. 22, no. 3, pp. 273-285.

Braun, V & Clarke, V 2021, Thematic analysis: a practical guide, Sage, London.

Creswell, JW & Plano Clark, VL 2018, Designing and conducting mixed methods research, 3rd edn, Sage, Thousand Oaks, CA.

Duggan, J, Sherman, U, Carbery, R & McDonnell, A 2020, ‘Algorithmic management and app-work in the gig economy: a research agenda for employment relations and HRM’, Human Resource Management Journal, vol. 30, no. 1, pp. 114-132.

Fair Work Commission 2024, Minimum standards for employee-like workers on digital labour platforms, FWC, Melbourne.

Fair Work Legislation Amendment (Closing Loopholes No. 2) Act 2024 (Cth).

Mazmanian, M, Orlikowski, WJ & Yates, J 2013, ‘The autonomy paradox: the implications of mobile email devices for knowledge professionals’, Organization Science, vol. 24, no. 5, pp. 1337-1357.

McDonald, P, Williams, P, Stewart, A, Oliver, D & Mayes, R 2019, Digital platform work in Australia: prevalence, nature and impact, Queensland University of Technology, Brisbane.

Rosenblat, A & Stark, L 2016, ‘Algorithmic labor and information asymmetries: a case study of Uber’s drivers’, International Journal of Communication, vol. 10, pp. 3758-3784.

Safe Work Australia 2022, Model code of practice: managing psychosocial hazards at work, Safe Work Australia, Canberra.

Senate Select Committee on Job Security 2022, The job insecurity report, Commonwealth of Australia, Canberra.

Topp, CW, Ostergaard, SD, Sondergaard, S & Bech, P 2015, ‘The WHO-5 Well-Being Index: a systematic review of the literature’, Psychotherapy and Psychosomatics, vol. 84, no. 3, pp. 167-176.

Wood, AJ, Graham, M, Lehdonvirta, V & Hjorth, I 2019, ‘Good gig, bad gig: autonomy and algorithmic control in the global gig economy’, Work, Employment and Society, vol. 33, no. 1, pp. 56-75.