Some visa policy questions on Cooks, students and working holiday makers

Earlier this week, I took issue with the methodology in some recent research titled Immigration and Unemployment 2014 by Dr. Birrell and Dr. Healy of Monash University.

While the numbers produced from their report drew most of the headlines, I’d like to discuss some policy concerns they raise in their paper which I largely agree.

I want to preface this discussion. I think more migration, in the abstract, is a very positive thing and working towards this goal is important. However visa programs should be used the way they are intended to ensure public and political support. This is sometimes overlooked but something I consider critical for the long-term success of immigration policies. It’s up to immigration advocates to call for visa systems that can be abided by and build immigration. Allowing visas to be abused is pernicious in a democratic society.


As the authors – Bob Birrell and Ernest Healy – highlight, the single largest occupation in 2012-13 for migrants seeking skilled visas was cooks. This includes 8,449 permanent residency visas and 3,040 temporary 457 visas. However there is no apparent overwhelming employment demand for cooks in other indicators, such as ABS labour market surveys or Department of Employment forecasts.

I’m less concerned about the permanent visas. This is a hangover from past policy mistakes around student visas. The migration system is slowly but surely working through this. As a result, the period of 2008-2014 may not produce the best labour market results from highly-skilled migrants. Despite this, I believe the issues have been largely addressed and future migration outcomes will be vastly improved (see below).

However there are issues with 457 visas and cooks. The central reason is pay. Under the 457 visa program, the minimum salary threshold is $53,900. As Birrell and Healy point out, the market salary rate for a cook is less than this, typically between $40,000 and $50,000 when you account for penalties and hours worked.

An important principle of the 457 visa program is equal pay for equal work. You can argue about whether this principle should exist (I believe it should) but one should recognise it is embedded in legislation and has bipartisan support. Either employers are paying migrant cooks above the market wage – something I find difficult to believe for such a number of people – or underpayment is occurring once visas have been approved to firms. Unfortunately I believe this is likely the case, at least for a substantial number of the cooks being sponsored.

Under program rules, technically employers should not be able to hire migrants on 457 visas if the market salary is below the threshold salary. This means an occupation like Cooks with a market salary of below $50,000 should have few 457 visa holders and should definitely not be the number one occupation under the program. Cooks is the most obvious outlier to this issue and demands a closer response from government. It’s not just cooks either. I have written previously on a range of occupations in the hospitality industry where there appear to be some serious disjuncture of visas and labour markets.

All of this raises important questions. This is not to say remove or cap the 457 visa program immediately. The program is vital to our immigration framework and fills an important niche for the labour market. But a more thorough approach to granting visas is apparent, particularly in relation to wages, conditions and enforcement. Coordination enforcement by using visa data should be (relatively) simple. Given all of this is well known in the bureaucracy, hopefully there is some attempt to address this in the Abbott government’s response to the recently concluded 457 visa integrity review.


The interaction between student visas and residency visas may not seem apparent on face value. Student visas are to obtain an Australian education, not to live here. However immigration isn’t that simple. Many international students seek permanent residency, some even value this above obtaining an education qualification.

For sometime Dr Birrell has successfully prosecuted the case about how student visa settings in the 2000s enabled permanent residency in a manner which had massive unintended consequences. Poor quality education providers thrived, gorging students for money as they attempted to become permanent residents. This is now accepted as a failure of immigration policy making, perhaps the most prominent in the past generation (outside of asylum policy).

As a result, there are former international students living legally in Australia who have applied for permanent visas and meet all the criteria yet remain without a visa after years of waiting. The number of these former students was once much higher (numbering above 100,000) but has since dwindled to less than 15,000. Birrell calls this process ‘warehousing’. In essence, by accepting these former students with low quality educational qualifications, the skill level of these new permanent residency is questionable.

However, it is important to note. These former international students were simply playing by rules governments established. Refusing their visas would have been a substantial policy shift and, from my point of view, highly unfair. It is also an issue which has nearly worked its way through the system. The method of obtaining residency via a student visa pathway has changed substantially in the past four years and is now well structured.

Working Holiday Makers

The evidence obtained in Immigration and Unemployment in 2014 is perhaps strongest on Working Holiday Makers.

In Table 6 (p.21) it is shown nearly 600 Working Holiday Makers living in Australia were sponsored by employers as ‘Program or Project Administrators’ in the 12 months to August 2012. This is over 35 per cent of program or project administrators sponsored onshore in Australia. This is emblematic of the complexities of visas.

It’s unlikely these young migrants are working in the occupation of program or project administrator as intended. More likely, they were working for an employer (allowed for up to six months on a WHM) and both the employer and migrant sought a method to work full-time, transferring to a 457 visa. If these migrants have the skills and qualifications required to perform the job, this is a positive use of the program. However the scope, size and trend increase of backpackers moving into the skilled workforce calls into question what is happening at the macro-level as this has not occurred before.

The number of WHM visas has been increasing over the past three years from 185,480 in 2010-11 to 249,231 in 2012-13. Unlike other immigration programs, the WHM is not driven by the demand of the labour market but by the push factor from overseas. How many young people want to visit Australia? Work in Australia? In the case of some European and Asian countries – many do.

Unlike Dr. Birrell, I believe capping the program would be a mistake. This program is built on reciprocity and a more restrictive program may impact on the ability of Australians to undertake similar opportunities overseas. Too often this benefit is completely overlooked in migration analysis.

However there are methods to improve the integrity of the program. The transition from a WHM visa to a sponsored visa should be a point of interest and subject to higher verification. DIPB should work with labour market enforcement officials elsewhere to better analyse and monitor where working holiday makers do work and what their salaries and conditions are, particularly if they transition to a more substantial visa. If we don’t even know the basics about the program (and we don’t), undertaking analysis to better tailor outcomes to government policy intentions is very difficult.

There appears to be cause to consider removing the second year of eligibility. This additional time period was only introduced as a favour to the National’s by the Howard government as a source of labour supply for regional agriculture. Little evidence has been provided (or research undertaken) to understand this policy issue. However since that decision was made, the Seasonal Worker Program has been introduced. In my opinion, this is a more appropriate method to address labour concerns given the development benefits for Pacific countries and the additional integrity embedded in the program.

Addressing an ageing population

Immigration is important for countries with fertility rates below replacement levels to maintain and grow population. This is undisputed. However the degree of importance certainly is.

I believe a positive net flow of migrants will help – at the margins – reduce the dependency ratio of taxpayers to younger (<15) and older people (>65). However migrants will not “solve” an ageing population. This was highlighted by the ABS yesterday, with their “Does Size Matter” article detailing population projections. Even spectacular inflows of migrants unlike Australia has ever seen before will only have a degree of impact on an ageing population. This is something Dr. Birrell notes but conflates with his argument about employment.

There are other reasons apart from an ageing population why I think a larger population is beneficial but we’ll leave it at that for now.

Birrell and Healy’s research included some dodgy numbers. But like all matters migration, this is a complex area where they note some important concerns we should better pay attention to.

Immigration and Unemployment in 2014: Did they take our jobs?

There are major issues with Bob Birrell and Ernest Healy’s latest publication, Immigration and Unemployment in 2014.

The central claim (p.3):

“Net Overseas Migration (NOM) is running at some 240,000 a year. The result is that, as of May 2014, the number of overseas born persons aged 15 plus in Australia, who arrived since the beginning of 2011, was around 709,000. Most of these people are job hungry. According to the Australian Bureau of Statistics (ABS) Labour Force Survey, 380,000 of these recent arrivals were employed as of May 2014. Over the same three years, the net growth in jobs in Australia is estimated by the ABS to have been only 400,000. This means that these recent overseas born arrivals have taken almost all of the net growth in jobs over this period.” (My emphasis)

This research was picked up extensively. Birrell’s own op-ed ran in Fairfax papers. Fairfax picked it up as a story, including it in a series on Thursday called “Broken Borders”. Business Spectator ran Callum Pickering. MacroBusiness ran Leith Van Onselen. All of these articles and op-eds focused exclusively on the theme of Australian’s missing out on jobs being created in the last three years. MacroBusiness even ran the title “Migrants take 95% of jobs created since 2011”.

But is it true? Have recent overseas born arrivals taken almost all of the net growth in jobs between 2011 and 2014? Have only 20,000 new jobs out of 400,000 since 2011 have been filled by Australians?

This is highly unlikely.

By conflating the number of recent overseas arrivals with the net growth in jobs, the authors are comparing apples with oranges. By using the gross number of migrant arrivals, they are excluding a key variable: overseas departures. Using the net employment figure while dismissing parts of the labour market which make it a “net” figure leads to some confusion.

To explain. Migrants come, many stay and work. However some leave the workforce also. This has an impact on employment but is hidden from the net employment growth figures Birrell and Healy source their research on, if the migrant who departs is replaced by another worker.

A basic example:

There are three jobs at a firm. Each person is a migrant. They quit their jobs and leave the country. Another three migrants are hired who move to Australia.

We have three new arrivals working in the labour market but as the firm has the same number of jobs available, no change appears in the net employment figure.

This has been understood for sometime. Some of Australia’s best migration researchers – such as Graeme Hugo – have worked extensively since the 1970s on how circular migration has evolved over time. Circular migration means people come and go as opposed to settle in one place. Labour market churn coupled with circular migration mean figures based only on arrival numbers are misleading.

Using this Birrell and Healy methodology, we do not know one way or the other how many recently arrived migrants ended up in newly created jobs for the period 2011-14. We don’t have enough information about the labour market or migration trends. It is possible newly arrived migrants accounted for 380,000 of the 400,000 created jobs from 2011-14 but it could also be any number from 0 to 399,999. The point is: we don’t know from this analysis. The authors certainty is misplaced and damaging.

So what is a more appropriate figure for these types of questions?

DIBP provide good forecasts of overseas departures to key into assumptions and estimates. Looking at a detailed breakdown of the Net Overseas Migration figures (table 3) we can conservatively estimate about 50 per cent of new arrivals are offset by new departures.

Using this NOM data and using the same figures as Birrell and Healy, 380,000 migrants are employed who arrived between 2011 and 2014. However we also assume 190,000 people departed Australia in the same time who were in employment. By using the net migration figure (arrivals minus departures) – 190,000 – for the 2011 and 2014 period, we have a better understanding of potentially how many migrants filled newly created employment positions.

This leaves us with approximately 190,000 net migrants filling about half of the 400,000 net employment growth for 2011-2014. Is this high? For comparison, this migration contribution to the net jobs figure is actually less than the net migration contribution to our rising population, which runs at about 60 per cent migration to 40 per cent natural increase (births minus deaths).

Perhaps we don’t have enough migrants in the labour market?

However I’d like to issue a note of caution regarding this entire methodology. Birrell and Healy themselves are aware of the key flaw in this methodology when they state (p.8):

“DIBP cannot provide data on the labour market participation of the people added to Australia’s population each year. Thus, it is not possible to identify precisely the contribution of NOM to the growth in the labour force shown in Table 1.”

“It is not possible to identify precisely the contribution of NOM to the growth in the labour force shown in Table 1”. By itself, this is a pretty damning statement against their own research findings. Unfortunately this was not noted in any of the publicity the research received.

As I said, I find this research misleading. From experience, the authors know a report like this will reach a broad audience, many of whom are more worried about daily life than they were previously (despite rising prosperity). Their message provides a simple answer to the wrong question: The government is at fault because there are too many migrants in the labour market.

In fact, this is likely not the case at all. Bob Birrell knows about how net migration operates and should have done better to justify these claims. Mistakes have been made in immigration policy – particularly in student policy which Birrell helped document – but debates over the labour market where passions run high should be argued on evidence grounded in rigorous analysis, not numbers thrown together in haste.

A deep and growing global research agenda shows the impact of immigration on the domestic population is either neutral or benignly positive. This is especially the case when discussing skilled migration for a developed country. Simultaneously, the benefits for the migrants themselves and often the countries they come from is large and significant.

Work such as Birrell and Healy’s research crowds out space for the questions we should be asking on immigration policy. How does this global agenda relate to Australia? Should immigration be used to reduce inequality? What role can immigration play in global development?

The authors use this research to push a singular, particular message: to limit further immigration to Australia. Sadly, it appears to have been taken on face value and pushed through the media cycle without question lending credence to this message. This feeds into the public discourse about immigration as something to fear. There are reasons to have a debate about immigration and Australia’s population but researchers have an obligation to present their best attempt to explain what is going on and then move onto policy preferences.

A more thorough analysis and the headline could have read: “We don’t know. But highly unlikely migrants are filling 95 per cent of new jobs.”

(This is my 200th post. Thanks to those who have provided feedback, commented and/or shared content. I appreciate it and I hope you enjoy reading.)

What is the labour force participation rate for secondary 457 visa holders?

Senate Estimates question BE14/118:

Senator Carr (L&CA 85) asked:

Senator KIM CARR: Are you able to provide me with a breakdown of the number of secondary 457 holders actively participating in the workforce?

Dr Southern: We would have to take that one on notice. I am not sure.


The department does not record the number of secondary subclass 457 visa holders who are actively participating in the workforce.

The Department of Immigration should not be tasked with collecting the labour market status of partners of visa holders. Fair enough.

However the Department does have some information about this specific question which could (should?) have been provided to Senator Carr.

In 2012, a survey was commissioned on the 457 visa program. In this survey are a handful of questions about the partners of 457 visa holders. A quick summary:

Do you currently have a partner living with you in Australia? 67 per cent responded yes.

If yes, did this partner migrate with you? 68 per cent responded yes.

(Asked if yes to previous two questions) Does your partner work in a job, business or firm? 61 per cent responded yes.

If no, has your partner looked for work in the past four weeks? 32 per cent responded yes.

Using the first two questions, we can assume 45 per cent of 457 visa holders migrate with a partner. In the parlance of government, these people are called secondary migrants, a group which also includes children.

We know from this question that at the end of April 2014, there were about 110,000 primary visa holders and 89,000 secondary visa holders. 45 per cent of 110,000 is 49,500 partners, leaving about 39,500 children.

Assuming 49,500 partners migrated with the primary visa holder and were living in Australia in April 2014, we can use the third and fourth survey questions to estimate a very rough labour force participation rate.

Partners of 457 visa holders: 49,500

In work: 30,195 (61 per cent of 49,500)

Looking for work: 6,180 (32 per cent of 39 per cent of 49,500)

Not looking for work: 13,125 (68 per cent of 39 per cent of 49,500)

Labour force participation rate: 73 per cent (In Work plus Looking for work divided by total Partners).

Unemployment rate: (approximately) 17 per cent (Looking for work divided by Looking for work plus In work).


First of all, people can reply to surveys in strange ways. This survey is asking one person (the primary visa holder) about a different person (their partner). While we can assume most people know their partners very well, the question about looking for work should be taken with a grain of salt. Maybe these migrants thought it would be good to say they were looking for work. Maybe the primary visa holder has no idea about their spouse.

Second, we are using percentages from one survey to infer about a different population of people. While the survey group from 2012 and migrants of 2014 are alike, they are not a perfect match. Thankfully the survey sample was very large meaning we don’t need to worry too much about serious bias but should keep this in mind particularly when we notice big shifts in trends.

Third, there is some uncertainty about the percentages relating to partners. The questions do not allow for the nuance of real life. Partners can migrate at the same time as the primary visa holder or at anytime afterwards. I don’t know what is the most common. I’d guess most people migrate together however I can think of a variety of reasons why this might not be so. The purpose of asking these questions – to sort out migrants from Australian born partners – is important however leaves us with a degree of uncertainty.

Fourth, we don’t know if the partner/children split of 49,500 / 39,500 is completely accurate as we’re using an assumption from the survey. From my experience, this sounds approximately correct it’s hard to nail down. The Department does not release the age of secondary visa holders so this is an assumption, not a fact.

Finally, this still does not answer Senator Carr’s question. He asked about secondary visa holders and this data only relates to partners. Of the 39,500 children – also secondary visa holders – we should assume at least some of them are working or are looking for work. We simply don’t know how many meaning we can’t fully answer Senator Carr’s question.

Despite these possible reasons for doubt, I think this information is useful. It provides a rough estimate for Senator Carr for his question. The labour force participation rate appears in the ballpark of what I consider reasonable. It is higher than the average for the Australian labour market because temporary migrants have no access to welfare support and probably need more income to live than an average household.

We can also assume about two thirds of partners are women as primary visa holders are two thirds male (we have no way to account for same-sex couples from the survey or other data released by the department). We know about half are from non-English speaking backgrounds.

Deducing from these two demographic details, a 17 per cent unemployment rate stands out. Non-english speaking female workers are more likely to be exploited in the labour market and have difficulty finding work. This certainly wasn’t the focus of Senator Carr’s question but I feel this is important to note.

What can Indian IT workers on H1B visas teach Australian immigration policy makers?

IT specialists, very often from India, make up one of the larger groups of temporary migrants, both in Australia under the 457 visa program and in the U.S. under the H1B program.

Unfortunately in Australia we have a lack of empirical insight into labour mobility for 457 visas. However in the U.S. new research suggests temporary migration programs are complex, with a degree of labour mobility.

This paper on the H1B visa program and Indian IT workers – Inter-Firm Mobility and Return Migration Patterns for Skilled Guest Workers by Peter  Norlander, Briggs Depew and Todd A. Sørensen – presents some evidence to suggest low wages for temporary migrants act as the same incentive for domestic workers, namely they seek other, higher paying jobs or return home:

Our results cast doubt on the claim that these workers face severe mobility restrictions outside of the Great Recession. They reveal that during periods of full employment, inter-firm mobility of these workers is comparable to other estimates in the literature obtained from presumably more mobile workers in other labour markets, suggesting that competitive market forces provide some check against firms dramatically underpaying these workers.

There is a fair amount to unpack here from a policy perspective.

First, is this research applicable in Australia? The study analysed six large India IT firms many of which operate in Australia. The type of employees are also similar. The visa costs are similar also. One major difference may be the visa programs in Australia have a more strict method to match migrant wages to domestic wages than the U.S. As highlighted in the research, one study found as soon as H1B visa holders received permanent residency, they experienced a 20-25% pay increase. I don’t believe this occurs here however there is a lack of research to suggest one way or another. Obviously there are inherent differences in the broader labour market between Australia and the U.S. however the particulars discussed – temporary migration and labour mobility – seem a decent, but imperfect, match.

Two main policy considerations jump out. What is the difference of labour mobility compared to other workers? What happens to labour mobility in a weaker labour market?

The research finds “that a 10% increase in salary yields a 3.32% decrease in the probability of quitting”. The authors claim the “fact these results are not zero suggests that these workers have some degree of mobility and that lower paid workers are able to relocate to other employment”. However, I would also note that “This is slightly smaller than previous results in the literature that study other groups of workers”.

The fact there is labour mobility in temporary migration programs is important and positive. This means market forces are still having an effect on employees moving firms despite the more restrictive regulations governing a temporary migrant’s relationship to their employer. I know in the 457 visa program, there are always more business nominations than visa grants. The difference in these two numbers is likely to be a good proxy for a degree of mobility in Australia. Unfortunately this information is not currently publicly available.

However the authors conclusions are heavily caveated when they say “outside the Great Recession”. This indicates in periods of weak labour markets, lower mobility may be more consequential for temporary migrants than domestic workers.

When the unemployment rate is 4%, then a 10% increase in the wage is associated with a 14% decrease in the quit rate. When unemployment is 6% the 10% increase in wage corresponds to a 6% decrease in quits. At an unemployment rate of 8% we see that the quit elasticity is not significantly different from zero.

The difference in ‘quit rates’ (a proxy for labour mobility) between 4% and 6% is quite large while at 8% there is no mobility. This points to how temporary migration programs should be understood very differently in ‘full-employment’ labour markets and weak labour markets. Of course, this is similar to how other employees operate in labour markets. However the key difference is that unemployment benefits and the right to stay in Australia have a radically different impact on domestic workers than migrant workers. The worst situation for governments who manage temporary migration programs is one where there is no labour mobility, with many employees taking below market wages, increasing worker exploitation and undercutting wages and conditions. This is particularly true for progressive governments such as Democrats and the Australian Labor Party.

There are lessons here for Australia. With the unemployment rate going from under 5% to over 6% in recent years, we should assume the labour mobility rate of temporary migrants has decreased (but remained positive), at least for Indian IT workers but probably more generally. Policy makers should be aware of this when considering changes to regulations, especially those which would further restrict mobility.

Further, it is positive to see research which finds the lowest paid temporary migrant Indian IT workers have the highest propensity to quit their jobs. This counters some of the claims around the severity of labour mobility restrictions however does not remove the concerns completely. In addition, the authors find lower paid workers are more likely to return to their origin countries. We can debate whether a ‘return home’ option is preferable to the option to stay in either the U.S. or Australia, but what this does indicate is how temporary migrants react to changes in the business cycle and these programs are dynamic. Understanding temporary migration programs as simply as a static, one-size-fits-all concept, underestimates the various migrant incentives which can be accentuated under different labour market conditions.