They thought banning migrant farm labour would boost wages for native-born farm workers. They were wrong. And New Zealand may be getting ready to repeat their mistake.

On 31 December 1964, the United States ended the bracero agreements between the US and Mexico, after two years of tightened restrictions. The agreements, which began in 1942, regulated the movement of lower-skilled migrant labour – particularly for seasonal agricultural work. By the early 1960s, about half a million Mexican farm workers migrated to American farms for seasonal agricultural work on contracts lasting from six weeks to six months.

The Kennedy administration believed that the bracero agreements reduced American farm labour wages. It also did not help that the senior commissioner in the Department of Labour investigation of the bracero programme was a eugenicist who believed Mexicans were genetically inferior.

After first requiring higher wage rates for bracero workers in 1962, in an attempt to encourage hiring American workers rather than Mexican workers, the American government ended the agreements.

Half a million Mexican workers stopped being able to help harvest American crops. The work had made the migrant workers better off. It also had made their employers better off. But it ended, in hopes of forcing wages up for American farm workers.

It all sounds rather familiar to Kiwis who have been paying attention to debates about New Zealand’s Recognised Seasonal Employer (RSE) scheme.

Our scheme, which began in 2007, regulates the flow of migrant seasonal workers from our island neighbours to New Zealand farms. The R in the RSE scheme matters. Prior to the scheme, there were worries about illicit migrant labour, worker overstays, and worker exploitation. Only recognised employers have access to RSE labour. And evaluations of the programme have noted that RSE workers enjoy stronger protections than other migrant workers.

The RSE scheme expanded over the subsequent decade. 4,486 workers arrived under the scheme in its first year; pre-Covid, some 14,000 workers could arrive to assist on New Zealand farms during the seasons when they are needed. Workers spend part of the year here earning money to bring back home. Surveys of those workers show that their earnings go to things like paying tuition fees for their school-aged children, starting small businesses, and putting new iron roofs on homes that sit in the path of cyclones. It has been considered a gold-standard in foreign development assistance.

Employers also appreciate the scheme. RSE workers arrive strongly motivated and prove highly productive. Horticultural workers are often paid piece-rates that vary with the quantity of fruit picked. Work by Waikato University’s Professor John Gibson, joint with the World Bank’s David McKenzie, showed that median earnings for a RSE worker with three years’ experience were 21.5% higher than median earnings for a local worker with three years’ experience.

But the scheme also has its detractors. A report released earlier this month by NZIER asserted that, in the absence of RSE workers, seasonal employers would be forced to invest in machines and that wage increases would follow. A report by the Productivity Commission, released last week, agreed with NZIER.

But neither report provided any particular evidence that banning foreign workers would lift locals’ wages. And while it sounds intuitively appealing that banning foreign workers might have that effect, things are a bit more complicated.

If we go beyond Econ 101 basic supply and demand, we need to consider whether migrant workers are really substitutes for local workers, or whether they might instead be complements.

What does that mean? Let’s take a very simple example. Suppose someone eats a lot of carrot sticks and onion dip and that you consider their diet to be your business. You want them to eat more carrots. Would banning onion dip force them to sate their hunger with more carrots?

If you think banning onion dip encourage more carrot-eating, you’re claiming that the two are substitutes for each other. If the price of onion dip dropped by a lot, the person might just eat spoonfuls of onion dip; if the price of onion dip went up instead, our friend would only eat carrots.

If you think instead that a ban on onion dip would end carrot-eating, you’re claiming that the two complement each other. They are better together, and neither is all that great on its own. Making either harder to get reduces the consumption of both.

And so we are faced with an empirical question. If New Zealand accepted fewer RSE workers, would agricultural employers pay a lot more to attract more local workers and to mechanise parts of their operation?

International evidence suggests foreign workers are more likely to be complements to local workers than substitutes for them. Meta-analysis by Giovanni Peri showed migrants typically have either no effect, or small positive effects, on locals’ wages. Large positive effects on locals’ wages are more likely than any negative effect.

Local evaluations suggest that RSE workers are more likely to be complements to local workers rather than substitutes for them. RSE evaluation work commissioned by Immigration New Zealand suggests the scheme has enabled companies to make substantial investments in new fruit processing facilities. Availability of RSE workers meant the whole sector could expand, hiring more Kiwi workers in parts of the process that follow along after the fruit is picked.

But the local evaluation work is more qualitative and anecdotal than empirical. So we turn back to the rather rigorous empirical work on America’s bracero restrictions in the 1960s.

The bracero programme shared a lot of similarities with New Zealand’s RSE scheme. It allowed for a lot of migrant seasonal agricultural labour to assist on farms. It was killed by President Johnson because of a pervasive belief there, like here, that migrant labour depresses local wages.

Did banning bracero workers prove a boon for American agricultural workers? Work published in the American Economic Review by Michael Clemens and co-authors in 2018 showed what happened. Because there were substantial state-by-state differences in reliance on bracero workers prior to the ban, the effects of the ban were easier to tease out.

Farms shifted to mechanisation where possible; production declined for crops for which machines could not easily take the place of workers. But agricultural wages were unchanged, and employment of local workers did not increase. The bracero workers were not a substitute for local workers and forced mechanisation did not improve worker wages.

Clemens quotes agricultural economist William Martin who, in the aftermath of the ban, described those advocating the ban as having been “obviously … extremely naïve”.

Banning RSE workers here, or substantially reducing their numbers, is not likely to improve wages for local agricultural workers. It is instead likely to result in shifts in agricultural production, and some mechanisation in areas where such mechanisation is possible. In areas where mechanisation is harder, we should expect decreased employment of locals in the processing plants that rely on a successful harvest.

It would be helpful if policy analysts here in New Zealand recommending substantial changes to the RSE programme had some passing familiarity with core, relevant research results published only two years ago in the world’s foremost economics journal.

It might help us avoid repeating others’ mistakes.