Mapping wage theft with data science

What started as a few underpayment cases revealed by Australian media in 2015 has turned into an epidemic of ‘wage theft’.

Wage theft is the popular term that has come to describe “under-or non-payment of minimum wages and entitlements that are rightfully owed to a worker.”

Wage theft broke into the public consciousness at large in 2015 with a joint investigation by Fairfax Media and ABC’s Four Corners into underpayment of 7-Eleven workers.

Many cases of underpayment by major businesses have come to light since the initial investigation, and the limited evidence available points to an increasing prevalence of wage theft over the last several years.

A 2019 report by the McKell Institute, which collated evidence from several surveys, found that on average 60 per cent of survey respondents had experienced wage theft. However, in surveys specific to young people, this number could be as high as 76 per cent.

This is more than just a few rogue employers not complying with the correct payment awards. Researchers and campaigners have described wage theft as a norm across several industries, while the Australian Council of Trade Unions has described it as a ‘business model’ for many employers.

Since 2015, Victoria has criminalised wage theft – with some other states following - and the federal government has recently proposed jail time and large fines for deliberate underpayment of workers.

But the problem is that detecting wage theft is difficult, relying largely on employees reporting a case to their union or the regulator, the Fair Work Ombudsman. And neither the regulator or unions have sufficient resources to combat it.

One answer is to give workers and regulators better tools to uncover wage theft. Our Fair Day’s Work project, backed by the Paul Ramsay Foundation, is developing software that will use predictive algorithms to help regulators identify the high-risk areas and businesses they need to focus resources on.

At the same time we are developing an online portal to support working people to keep track of their workplace rights. It will be particularly aimed at young workers who are among the most vulnerable to wage theft.

Employees in industries at high-risk of wage theft – such as the hospitality and retail industries – are predominantly young people aged 15 to 24 years old. Australian Bureau of Statistics data shows that 66 per cent of retail workers and 70 per cent of accommodation and food service workers are aged between 15 and 24.

These are also industries that employ large numbers of young migrant workers – who are at an even greater risk of wage theft. A 2019 survey by Unions ACT found experiences of wage theft have increased every year since 2017, but only 25 per cent of young workers reported trying to address wage theft.

It suggests that many young workers chose not to act due to fears over losing their job, especially as these jobs are usually casual roles.

While wage theft obviously harms employees, the Australia Chamber of Commerce and Industry notes that business also faces negative social and economic consequences of wage theft.

For example, wage theft creates an anti-competitive effect, where businesses exploiting workers get a competitive edge over employers that do the right thing, which further normalises wage theft in certain industries.

As a result, good employers doing the right thing are being undercut by employers doing the wrong thing.

The economic downturn that has been caused by the restrictions put in place to counter the COVID-19 pandemic only increases the risks that employers may deny young workers their entitlements as they try to address shortfalls in revenue by intentionally or accidentally cutting wages and conditions.

This risk is further exacerbated by casualisation and other forms of precarious work which mean unscrupulous employs can easily fire workers who complain.

Similarly, in an economy where jobs are scarce, people may be less likely to report wage theft for fear of losing work.

Currently, strategies to address wage theft differ across jurisdictions. Many approaches are purely deterrent based, such as the criminal liability legislation we discussed above – they focus on penalties for those who commit wage theft.

However, the available evidence shows that deterrent approaches involving penalties alone aren’t necessarily the most effective way to address wage theft.

This is in part due to the difficulty of detecting underpayment and other breaches, but there is also evidence to suggest that higher penalties are less of a deterrent compared to increasing the likelihood of detection.

One approach to improving the detection of wage theft is using data and predictive algorithms to support regulators and unions’ compliance checking and enforcement activities. Recent research work in the United States has seen predictive algorithms used to support the prioritisation of wage theft investigation.

Collecting and sharing data about wage theft is also important to raise public awareness of wage theft’s systemic nature. These uses of data aim to increase the likelihood that wage theft will be detected either by the regulator, workers or unions.

The Fair Day’s Work project, recently launched by the Melbourne School of Government, will bring together business, government, academia, unions and NGOs to develop a set of data-driven tools to increase the likelihood of wage theft detection.

The project will involve developing an online portal, public dashboard, wage theft database and wage theft prediction tool.

The online portal will support young people to access tailored information about their employment rights, upload data about their employment and receive a tailored wage theft risk assessment and preventive measures.

The public dashboard will provide information for Australians regarding wage theft prevalence in their community. A core part of this project is assessing the availability of wage theft data in Australia and developing a wage theft database that can enable research, policymaking, and improved detection.

We already have a good understanding of the major risk factors for wage theft in terms of industry and employee characterises – however there is other data, including macroeconomic and microeconomic data that could factor into the risk of wage theft.

The project was selected by the Paul Ramsay Foundation for support as part of the Inclusive Growth and Recovery Challenge led by data.org.

Launched in partnership with the​ ​Mastercard Center for Inclusive Growth​ and ​The Rockefeller Foundation​, the data.org Challenge identified breakthrough projects that harness the power of data science to help people and communities thrive.

Unless we enhance the abilities of regulators to detect wage theft and empower workers with better knowledge of their rights, the risk is that wage theft will simply remain “normal” for many industries and businesses.

Banner: Getty Images