VoxEU Column Development Migration

What are the development impacts of “best practice” seasonal migration schemes?

The debate over immigration often highlights the effects on the native population. This column instead looks at the country that immigrants leave behind. It presents a multi-year evaluation of New Zealand’s seasonal migration programme. In its first two years, the main developing countries supplying workers to the programme have seen a large improvement in household wellbeing.

The debate over immigration often revolves around the effects on the native population – but such focus risks neglecting a key benefit of migration.

Almost all OECD countries have temporary worker migration programmes, with seasonal workers the largest single category, totalling 576,000 workers in 2006 (OECD, 2008). While these programmes offer the possibility of a “triple-win” – benefitting the migrant, sending country, and receiving country – they remain controversial with some critics arguing that workers will over-stay and push down the wages of poor native workers (e.g. Borjas 2007), while others worry about the possible exploitation of migrants and whether they can earn enough to make it worthwhile if the duration of work is short (Ruhs 2006). One reason for controversy may be that there is a dearth of rigorous evidence on development impacts of seasonal migration, making it difficult for public debate to have firm foundations.

Evidence from a new seasonal migration scheme in New Zealand

In a recent working paper (Gibson and McKenzie 2010) we provide robust evidence on the development impact of a best practice seasonal migration programme. Our paper reports on a prospective multi-year evaluation of the effect on incomes, consumption, and other welfare indicators for households in two Pacific Island countries who supplied workers to New Zealand’s new Recognised Seasonal Employer scheme. This scheme has many features that take account of lessons from previous seasonal worker programmes.

The scheme began in 2007 and initially allowed up to 5,000 migrant workers per year to be recruited for seasonal work in New Zealand’s horticulture and viticulture industries, with the cap raised to 8,000 in October 2008. Migrants work for an accredited employer, for up to seven months per 11 month period, and may return if recruited again. The employers are obliged to:

  • Show that no New Zealanders are available for the work,
  • contribute one-half the cost of airfares for their migrant workers,
  • pay market wages and provide a minimum quantity of work,
  • and be responsible for costs of repatriating any migrant workers who overstay their visa.
  • The scheme pays special attention to development impacts in the Pacific Island countries that provide over three-quarters of the workers, with initial support from New Zealand’s foreign aid and labour ministries to aid effective recruitment of workers from the Pacific.

Between 2007 and 2010 we conducted four waves of surveys in the two countries providing 70% of the Pacific Island workers in the scheme– Tonga and Vanuatu. In each country we surveyed approximately 450 households drawn from about 50 communities, including households supplying workers, households with applicants who were not (yet) recruited and non-applicant households. Our baseline survey was before workers left to work in New Zealand in the first season, and then re-interviews were 6, 12 and 24 months later. Using these rich baseline data and institutional knowledge of how recruitment for the programme occurred, we use propensity-score matching to identify an appropriate set of households to act as a comparison group for the households participating in the scheme, and then use panel difference-in-differences and fixed effects estimation to assess the impacts of the scheme on household incomes, consumption, durable assets, and subjective well-being.

What are the consequences of participating in seasonal migration for these households? Figure 1 summarises key results for monetary welfare measures, contrasting households who ever had an Recognised Seasonal Employer worker with matched households whose applicants were not recruited (following the recommendation of Crump et al. 2009) we drop observations with propensity scores outside the 0.1-0.9 range, and in the highlighted results we also trim observations above the 99th percentile to ensure robustness against outliers. In our full working paper we use several different definitions of counterfactual groups and get very similar results to those highlighted here:

  • The per capita income of participating households is 36%-39% higher than for matched households who wanted to participate in the scheme but did not have their workers recruited, using the difference-in-differences specification.
  • When household fixed effects are used the range of income gains widens, to 25%-42%, with the decline to 25% mainly because of volatility in measured household incomes between the waves of the panel survey in Vanuatu.
  • The per capita expenditures of participating households go up by rather less than do incomes (an increase of only about 10%), suggesting that much of the extra income from participating in the scheme is saved, at least in the short-run covered by our surveys.

Figure 1. Impact of the Recognised Seasonal Employer scheme on income and expenditure

These effects in Figure 1 are an average effect of participating in the scheme. But households varied substantially in their exposure to Recognised Seasonal Employer work, with some participating only once and others having workers who were recruited multiple times. There is variation in the length of the employment contracts, with the median spell being only five months rather than the seven month maximum allowed by the programme. When we re-estimate the difference-in-differences and fixed effects regressions with our measure of treatment effects being the number of months of participation for the household, we find that each additional month that a worker was in New Zealand raises household per capita income by between 4% and 5%.

In addition to these effects on monetary welfare there are several other benefits to households from participating in the Recognised Seasonal Employer scheme. The subjective economic welfare of participating households is increased by approximately one-half of a standard deviation, which is slightly more than the proportionate increase in per capita income. Households that supplied Recognised Seasonal Employer workers were also more likely to make dwelling improvements, to open bank accounts, and to make major durable goods purchases. Furthermore, in Tonga, we find substantial increases in secondary school attendance for 15- to 18-year-olds in households participating in the programme.


The vast wage differences across countries are a sizeable economic distortion, and offer the possibility of large gains through international migration. From a development perspective, a key challenge is to increase the opportunities for poor, relatively less skilled, individuals to participate in migration. Temporary worker programmes offer perhaps the best such opportunity. Our research provides the first rigorous evaluation of the impact of a seasonal migration policy on households in the sending country, and finds gains in household wellbeing which greatly exceed those measured for other popular development interventions like microfinance and conditional cash transfers. Coupled with analysis which shows very low rates of overstaying and modest impacts on the native labour force (New Zealand Department of Labour 2010), these results suggest more countries should give seasonal worker programmes a chance.


Borjas, George (2007), “A Lemon in the Senate: The immigration deal is a travesty of a mockery of a sham”, National Review Online, 17 May.

Crump, Richard, V Joseph Hotz, Guido Imbens, and Oscar Mitnik (2009), “Dealing with limited overlap in estimation of average treatment effects”, Biometrika, 96(1):187-199.

Gibson, John and David McKenzie (2010), “The development impact of a best practice seasonal worker policy”, BREAD Working Paper 286.

New Zealand Department of Labour (2010), “Final Evaluation Report of the Recognised Seasonal Employer Policy (2007-2009)”, Department of Labour, Wellington..

OECD (2008), International Migration Outlook – SOPEMI -2008, OECD: Paris.

Ruhs, Martin (2006), “The potential of temporary migration programmes in future international migration policy”, International Labour Review, 145(1-2):7-35.

105 Reads