The Productivity Commission report – Immigration Fit For The Future looks at what working-age immigration policy settings would best facilitate New Zealand’s long-term economic growth and promote the wellbeing of New Zealanders.

The Commission carries out an interesting analysis of the role that migration plays in the New Zealand labour market which underscores the complex nuances of how skills are accessed, developed and utilised. This nuanced understanding of how the labour market operates is crucial for evidence based immigration and tertiary education policy. Unfortunately, New Zealand’s current suite of labour market information is not up to the task.

The Commission examines the extent to which migrants displace domestic workers. Different labour market conditions can lead to different outcomes. However, using data to identify migrant displacement of domestic workers is not easy. You can look at how many migrants are entering an industry compared with domestic workers. But, does a rising share of migrants indicate displacement or difficulty recruiting domestic workers? And, is this difficulty the result of a genuine shortage or working conditions that are not attractive to domestic workers?

The dairy industry is an example. The Commission reports that, over the past decade, the share of entry-level roles occupied by migrant workers has increased significantly. This may indicate displacement or reflect challenges attracting New Zealanders to work in remote regional areas.

There may be cases where the New Zealand labour market and education system are too small to develop and utilise certain highly-specialised skills. In which case, employing migrants can be the best option. Education providers need to meet minimum student numbers to run a course at a profit. And there needs to be enough work to sustain the employment of specialised domestic workers over several years. In cases where this isn’t possible, skilled migrants can be brought in for a finite period to provide specialist skills for specific projects. The tunnelling engineers that were brought in for public transport infrastructure projects in Auckland is one example of this.

There may be contexts in which employing migrants in certain kinds of roles, such as managerial, can enable the employment of domestic workers in other roles. There could also be instances where migrant workers complement the use of technology and higher-skilled domestic workers. The Commission uses fruit pickers as an example where migrant workers enable apple orchardists to invest in automated packing technology and employ higher-paid technicians.

These different context all require a different response from the immigration and tertiary education system. Immigration New Zealand regularly consults with industry bodies to determine the extent and nature of labour and skill shortages. But we lack a more objective and comprehensive information collection programme.

The role of migrants in the skills development of domestic workers is also complex. The Commission explains that there are several ways that migrants can influence the skills development of domestic workers. Migrants can be brought in to train or pass on knowledge to domestic workers, or they can be brought in to provide cover in the workplace for domestic workers to go away and train. The Commission uses the meat processing industry as an example where the provision of lower-skilled migrant workers allows domestic workers to be taken off the chain and receive training.

Unfortunately, access to a migrant workforce can also reduce the incentive for employers to recruit and develop domestic workers. The Commission points to the technology industry as an example of this. In recent (pre-COVID) years, easy access to migrant IT workers has made it difficult for domestic graduates to gain a foothold in the industry.

In terms of data collection, you can look at training delivery in industries that are taking on increasing numbers of migrants. But does falling training delivery indicate that employers are disincentivised to train their domestic workers because they have easy access to migrants with the skills they need, or is the training being offered not meeting employer skill needs?

Skill Shortage Lists (SSLs) are a good example of how labour market information is translated into immigration policy. SSLs cover occupations where skills are not readily available in the local labour market. SSLs offer easier access to migrant workers by enabling employers to forego Labour Market Tests. Labour Market Tests require employers to check that a domestic worker is not available before they recruit a migrant.

The Commission concludes that SSLs can reduce the need for firms to raise wages to attract local workers or to innovate and invest in capital equipment to raise labour productivity. SSLs can also disincentivise locals to train for these positions and discourage employers to invest in training locals. The Commission also argues that SSLs can be useful for meeting highly specialised skills needed for a fixed period of time such as the tunnelling engineers example mentioned earlier. But the SSLs need to be continuously reviewed to ensure that occupations are removed when these skills are no longer needed or in shortage.

But there is an argument for including occupations on the SSLs for which there are longer-term, systemic shortages either because the work is not attractive to domestic workers, or because the economy is at full employment and there simply aren’t enough domestic workers to meet employers’ needs.

The Commission argues that over the longer-term, a shortage of workers will lead to employers offering higher wages or improving working conditions which will attract people to develop the skills necessary for these jobs. But this only works if workers are sufficiently well-informed as to where these shortages are. The Regional Skills Leadership Groups (RSLGs) and Workforce Development Councils (WDCs) have been set up to carry out this role. And the Tertiary Education Commission is busy giving effect to its Careers System Strategy. Only time will tell if they are effective. Supports also need to be in place to enable workers to re-skill and, if necessary, relocate to where the jobs are. Labour markets have become much more ‘flexible’ over recent decades. But inflexibility persists. For example, the Commission points to regulatory settings that limit eligibility for student allowances for students aged 40 and over, which creates training barriers for mid-career workers.

Finally, if the economy is at full-employment as it probably is at the moment, raising wages in response to skill shortages will only lead to poaching of workers from other firms or industries. If this goes on for any length of time, wages (which ought to be driven by productivity increases) become decoupled from productivity leading to a reduction in labour productivity.

The Commission explains that occupations on SSLs are decided according to information provided by affected industries, employers, unions, and other interested parties. Consulting with various parties is a process intended to identify unmet labour demand and test whether employers are doing enough to attract workers from the local labour market. However, the Commission points out that this process can become cumbersome and open to lobbying. Also, SSLs are by their nature backward looking, in that they reflect yesterday’s labour shortages.

The Commission believes that SSLs can be improved by relying less on information provided by the affected parties and more on empirical data and transparent indicators such as vacancy data, wage movements, and patterns of demand for migrant labour. To inform this view, the Commission asked Taylor Fry to explore the feasibility of developing a quantitative occupational skills shortage model in New Zealand. Taylor Fry recommend that skill shortages should primarily be measured by determining whether employers are able to fill specific vacancies.

Taylor Fry recommend that vacancy information be collected as part of the Business Operations Survey (BOS). The BOS already asks businesses about the difficulty they are having recruiting across three broad occupation groups: managers and professionals, technicians and associate professionals, and tradespersons and related workers. The BOS surveys about 7,000 businesses in New Zealand that have 6 or more employees and have been active for one year or more out of a population of 40,000-50,000. It being a sample survey limits the extent to which results can be disaggregated to a fine level of occupational detail. The BOS coverage also means that data for small businesses, or new businesses is not captured. On the other hand, having recruitment data collected as part of a larger survey means it can be analysed against measures such as firm size, industry, and degree of change in the businesses’ use of technology.

The UK’s Employer Skills Survey (ESS) offers an example of how the BOS could be expanded. The ESS asks respondents if employers have vacancies, whether any are hard to fill, which occupations these hard to fill vacancies are in, and the reasons why the vacancies are proving hard to fill.

Taylor Fry also list a number of secondary indicators that together could be useful for predicting skill shortages. The list includes measures of employment, wages, hours worked, unemployment, and migrant employment. I disagree with Taylor Fry that the Household Labour Force Survey (HLFS) gives reasonably good resolution of occupations nationally. Stats NZ publishes HLFS employment for the eight ‘1-digit’ Australia and New Zealand Standard Classification of Occupation (ANZSCO) occupation groups, which is the highest level of aggregation. Ideally, data is required for the 1,000 or so ANZSCO ‘6-digit’ occupations. With the HLFS surveying 15,000 households every three months, it’s hard to see how it could offer data on ‘6-digit’ occupations that had small enough sampling errors. In fact, I consulted Stats NZ about this. They said that 6-digit ANZSCO was too granular to derive any meaningful interpretation of HLFS data.

I do agree with Taylor Fry that the Integrated Data Infrastructure (IDI) offers possibilities because data is available at unit record level. The Migrant Employment Dataset published by MBIE is one example of the IDI being put to good use. However, the IDI is only as good as the data it contains. And, broadly speaking, in New Zealand only employment data is collected at the level of detail and regularity required to indicate skill shortages. Data on hours worked and wages, for example, are not detailed enough.

The IDI also suffers from a lack of occupation data. There is occupation data from the 2018 Census in the IDI, but that is obviously well out of date now. Occupation data needs to be collected in detail on a regular basis and incorporated into the IDI. Inland Revenue Department employment data would be one way to do this. However there are significant practical challenges in having employers submit to IRD data about the occupations their staff work in, and having Stats NZ classify this data to the ANZSCO framework.

New Zealand needs to build on the ANZSCO classification framework. For too long we have relied on measures of employment across ANZSCO occupations as a proxy for skills demand. But there is very little information in the ANZSCO framework about the skills required for each occupation. This is a serious drawback for Workforce Development Councils (WDCs) and Regional Skill Leadership Groups (RSLGs) whose responsibilities include determining skills demand across industries and regions and translating this into demand for qualifications.

The Australian Skills Classification offers information about core competencies, specialised skills, and digital skills required across ANZSCO occupations. We could start using it right now. It’s not perfect. The ASC operates mostly across ‘6-digit’ ANZSCO occupations. But there are a number of areas where the information is aggregated to ‘4-digit’ occupation groups. A small number of occupations are not covered by the ASC, mostly in agriculture. There is also the question of whether skill needs in the Australian Job Market are the same as in New Zealand. And even if they were, it’s likely that skill needs vary across individual jobs even within a single occupation.

But the ASC covers 95% of employment in New Zealand. It is reviewed regularly to ensure it remains up to date. And it could be tweaked by WDCs and RSLGs to ensure that skill needs within individual occupations broadly reflect what employers are saying they need. Work is going on in New Zealand to create our own version of a skills classifications. But, it is not close to being completed and the WDCs and the RSLGs need something now.

Like an employer advertising a hard-to-fill vacancy, I feel like I have been banging on about better labour market data for some time now. The Productivity Commission report highlights the inadequacies in the suite of data available to us (perhaps inadvertently) by explaining the nuances of labour market dynamics that need to be understood for well-designed immigration and tertiary education policy. I hope somebody in Government is paying attention.