Maximizing the Development Impacts from Temporary Migration Recommendations for Australia’s Seasonal Worker Programme Maximizing the Development Impacts from Temporary Migration: Recommendations for Australia’s Seasonal Worker Programme Abstract The Seasonal Worker Programme (SWP) was formally introduced in 2012 following a four- year pilot scheme. The SWP maintains the dual objectives of: (i) contributing to the eco- nomic development of nine participating Pacific Island countries and Timor-Leste; and (ii) filling labor shortages in the Australian agriculture, accommodation (in selected locations) and tourism sectors (the Northern Australia tourism pilot). This paper assesses the first of these objectives, evaluating the impact of the SWP on workers, their households, and com- munities. In doing so, it aims to build on the evidence already collected on the development impacts of the Pacific Seasonal Worker Pilot Scheme (PSWPS) (Gibson and McKenzie 2011) and shed further light on how the programme can be improved to increase the benefits flowing through to the Pacific region. i © 2017 International Bank for Reconstruction and Development/The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of Consultants working for and the staff at The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the le- gal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons At- tribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution – Please cite the work as follows: World Bank. 2018. Maximizing the Devel- opment Impacts from Temporary Migration: Recommendations for Australia’s Seasonal Worker Programme. Washington, DC: World Bank. doi: - License: Creative Commons Attribution CC BY 3.0 IGO Translations – If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations – If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. ii Third-party content – The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting form such infringe- ment rests solely with you. If you wish to reuse a component of the work, it is your respon- sibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group. 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org Cover photo: Vedrana Music Cover design: Mai Chanthaboury iii Table of Contents Acknowledgements ......................................................................................................................................................... viii Abbreviations ........................................................................................................................................................................... ix Executive Summary ....................................................................................................................................................... x 1. Introduction ....................................................................................................................................................................... 1 2. Literature Review ............................................................................................................................................. 5 2.1 Global Literature ............................................................................................................................................................ 6 2.2 Pacific-focused Literature .................................................................................................................................... 9 3. Description of the Seasonal Worker Programme ................................................ 13 3.1 Origins .................................................................................................................................................................................... 14 3.2 Pacific Seasonal Worker Pilot Scheme ...................................................................................................... 15 3.2.1 PSWPS Evaluation Findings .............................................................................................................. 16 3.3 Seasonal Worker Programme ........................................................................................................................... 18 4. Evaluation Design and Methodology ........................................................................................ 21 4.1 Description of Evaluation Methodology .................................................................................................... 22 4.2 Sampling Strategy and Data Collection ................................................................................................... 23 4.2.1 Household Surveys ............................................................................................................................................ 23 4.2.2 Community Surveys ......................................................................................................................................... 24 4.2.3 On-site Worker Surveys ............................................................................................................................... 24 5. Results ..................................................................................................................................................................................... 27 5.1 Individual Worker Results .................................................................................................................................... 28 5.1.1 Determinants of Participation ................................................................................................................. 28 5.1.2 Seasonal Worker Characteristics ......................................................................................................... 29 5.1.3 Selection into the programme ............................................................................................................... 30 5.1.4 Labor Market Activity Prior to Departure ...................................................................................... 32 iv 5.1.5 Earnings, Costs, Remittances, and Savings .............................................................................. 33 5.1.6 Communication ................................................................................................................................................... 39 5.1.7 Skills Transfer ........................................................................................................................................................ 40 5.1.8 Plans Upon Return ............................................................................................................................................. 41 5.1.9 Workers’ Evaluation of the Programme ......................................................................................... 43 5.2 Household level Results.............................................................................................................................................. 45 5.2.1 Methodology .......................................................................................................................................................... 45 5.2.2 Income and Savings ......................................................................................................................................... 45 5.2.3 Impact on Household Expenditure ...................................................................................................... 46 5.2.4 Impact on Human Capital ........................................................................................................................... 47 5.2.5 Impact on Household Labor Supply .................................................................................................. 47 5.2.6 Impact on Home Improvements and the Accumulation of Assets ............................ 48 5.2.7 Investment and Impact on Business Ownership ..................................................................... 49 5.3 Community-level Results ....................................................................................................................................... 50 5.3.1 Community Role in Worker Selection ............................................................................................... 50 5.3.2 Participating Community Characteristics .................................................................................. 51 5.3.3 Community Contributions ......................................................................................................................... 52 5.4 Programme-level Results ....................................................................................................................................... 53 5.4.1 Predeparture Briefing ................................................................................................................................... 53 5.4.2 Add-on Skills Training (AOST) ................................................................................................................ 54 5.4.3 Superannuation .................................................................................................................................................. 55 6. Aggregrate Development Impact ................................................................................................. 56 7. Increasing the Development Impact of the SWP ................................................... 59 8. Conclusion .......................................................................................................................................................................... 65 References ..................................................................................................................................................................................... 67 Boxes Box 3.1 PSWPS Evaluation Core Findings ................................................................................................................. 17 v List of Figures Figure 3.1 Annual Arrivals under the Pacific Seasonal Worker Pilot Scheme (FY09 – FY12) .......................................................................................................................................................... 15 Figure 3.2 Annual Arrivals under the Seasonal Worker Programme (FY13 – FY17) .............. 18 Figure 5.1 Higher earnings remain the key determinant for worker participation............... 29 Figure 5.2 Those being selected into the programme have strong connections in Australia ........................................................................................................................................................................ 31 Figure 5.3 Employment rates vary across participating countries .................................................... 32 Figure 5.4 Employment status varies for those employed prior to departure .......................... 33 Figure 5.5 Net SWP earnings mostly meet workers’ expectations .................................................... 34 Figure 5.6 The vast majority of workers remit money while in Australia ..................................... 37 Figure 5.7 The frequency with which workers remit money is highly variable across countries ....................................................................................................................................................................... 37 Figure 5.8 The majority of seasonal workers communicate with home at least monthly ......................................................................................................................................................... 40 Figure 5.9 The most common skills acquired in Australia are those picked up on the job ..................................................................................................................................................................... 40 Figure 5.10 The skills acquired are deemed to improve employment prospects upon return ........................................................................................................................................................................... 41 Figure 5.11 Earnings are sufficient to sustain workers’ livelihoods during the return period ............................................................................................................................................... 42 Figure 5.12 Most participating workers are in the programme for long-term 
 employment ............................................................................................................................................................ 43 Figure 5.13 The majority of seasonal workers are satisfied with their experience in Australia ............................................................................................................................................................ 44 Figure 5.14 Given their satisfaction, workers are willing to recommend the SWP to others .............................................................................................................................................. 44 Figure 5.15 Maturity is the most important determinant of community endorsement .... 51 Figure 5.16 Most of the community contributions by go towards the church ......................... 53 Figure 5.17 Only in select participating countries are workers missing pre-departure briefings ....................................................................................................................................................................... 54 Figure 5.18 Not all participating countries benefit from Add-on Skills Training ...................... 55 Figure 5.19 Many workers continue to struggle to access their superannuation ................. 55 vi List of Tables Table 4.1 On-site worker survey sample ................................................................................................................... 25 Table 5.1 Seasonal worker characteristics .............................................................................................................. 30 Table 5.2 Weekly Earnings Net of Taxes ..................................................................................................................... 34 Table 5.3 Earnings, costs, remittances, and savings ....................................................................................... 35 Table 5.4 The average male seasonal worker stands to earn slightly more through the SWP .................................................................................................................................................................................... 36 Table 5.5 Seasonal workers are predominantly using Western Union to remit ........................... 38 Table 5.6 Calling Costs to the Pacific from Australia ....................................................................................... 39 Table 5.7 Household-level impacts on income and savings ....................................................................... 46 Table 5.8 Household-level impacts on expenditure .......................................................................................... 46 Table 5.9 Household-level impacts on education and health outcomes .......................................... 47 Table 5.10 Household-level impacts on labor supply ..................................................................................... 48 Table 5.11 Household-level impacts on home improvements and the accumulation of assets ............................................................................................................................................................................... 49 Table 5.12 Participating communities are better equipped in some areas, but not all ........ 52 Table 5.13 Contributions to the community vary widely across participating countries . 52 Table 6.1 Aggregate development impact across participating countries .................................. 57 vii Acknowledgements This report has been prepared by Jesse Doyle (Social Protection Economist, World Bank) and Manohar Sharma (Senior Economist, World Bank). They have been supported by a team consisting of Felix Leung (Consultant, World Bank), Ronald Duncan (Emeritus Profes- sor, Australian National University), Anuja Utz (Senior Operations Officer, World Bank) and Priya Chattier (Consultant, World Bank). It has been prepared under the initial oversight and management of Jehan Arulpragasam (former Practice Manager for Social Protection and Jobs, East Asia and Pacific Region, World Bank) and Franz Drees-Gross (former Coun- try Director, Pacific Islands and Timor-Leste, World Bank) and more recently under the guidance of Philip O’Keefe (Practice Manager, Social Protection and Jobs, East Asia and Pacific region, World Bank) and Michel Kerf (Country Director, Pacific Islands and Timor- Leste, World Bank). The peer reviewers were Stephen Howes (Professor of Economics and Director, Development Policy Centre, Australian National University) and John Gibson (Pro- fessor of Economics, the University of Waikato). The team would like to acknowledge a number of people who have provided inputs and guidance during the course of report preparation. Steven Kaleb, Anna McNicol, Frank Thompson, Aedan Whyatt, Tanya Morjanoff, Cameron Reid, and Zoe Kidd from the Austral- ian Department of Foreign Affairs and Trade (DFAT) provided guidance in both compiling the survey instruments and extracting the most relevant findings. Mark Roddam, Alison Durbin, Susannah Smith, Sheridan Langford, James Elton, Brian Weiss, Jessica Main, Lisa Calabria and Douglas Hay from the Australian Department of Employment (DoE, former- ly Department of Education, Employment and Workplace Relations) provided support in linking up with the seasonal employment units in sending countries and coordinating with Approved Employers (AEs) in both the horticulture and accommodation sectors. Halahingano Rohorua, ‘Ata’ata Finau, Viliami Fifita and the team from the Tonga Depart- ment of Statistics undertook the household and community surveys in Tonga, while the team from the Overseas Employment Division under the Ministry of Internal Affairs in Ton- ga provided timely data on departing seasonal workers. Simil Johnson, Harry Nalau, Benuel Lengue, Rara Soro and the team from the Vanuatu National Statistics Office provided sup- port in undertaking the household and community surveys throughout Vanuatu, despite the challenges posed by the onset of Tropical Cyclone Pam. Lionel Kaluat, Julie Rereman and the team from the Seasonal Employment Unit in Vanuatu assisted with the data on departing Ni-Vanuatu seasonal workers. Michael Lokshin and Zurab Sajaia from Survey Solutions provided technical support in what was the first trial of this software in the Pacific region. Aretitea Teeta, Claire Cronin, Detaviana Freitas, Geoffrey Kausei, Kale Gore, Niuatui Niuatui, Pyone Myat Thu, Rochelle- Lee Bailey, Sione Faleafa, and Tutua Teeba undertook the on-site worker surveys. Tom Perry, Kara Mouyis, Benjamin Brighouse and Vedrana Music managed the external com- munications and dissemination of the report. Christopher Stewart edited the report, and Mai Chanthaboury undertook the graphic design and formatting. viii Abbreviations ABARES Australian Bureau of Agricultural and Resource Economics and Sciences AE Approved Employer AOST Add-on Skills Training ATET Average treatment effects on the treated DFAT Department of Foreign Affairs and Trade DIBP Department of Immigration and Border Protection DoE Department of Employment FWO Fair Work Ombudsman FY Financial Year GP Global Practice LMAP Labor Mobility Assistance Program MTO Money Transfer Operator PIC Pacific Island Country PSWPS Pacific Seasonal Worker Pilot Scheme PSM Propensity Score Matching RSE Recognised Seasonal Employer Scheme SIDS Small Island Developing States SWP Seasonal Worker Programme TOP Tongan Pa’anga VUV Vanuatu Vatu ix Executive Summary The Seasonal Worker Programme (SWP) was initially introduced in 2012 following a four- year pilot scheme, which commenced in August 2008. It started off with a 12,000-worker cap for the initial four years of implementation, but this was removed in 2015, and a host of additional reforms aimed at addressing the demand-side constraints of the programme were introduced. Since these reforms, the programme has expanded rapidly–with 6,166 workers arriving in FY17. The SWP is one of only a handful of migration schemes globally that maintains the explicit objective of contributing to the economic development of labor- sending countries. This paper builds on the earlier evaluation of the Pacific Seasonal Worker Pilot Scheme (PSWPS) (Gibson and McKenzie 2011) and assesses the extent to which the programme has met this objective, since it moved from a pilot into a fully-fledged scheme. Given the limited data availability and relatively small size of the SWP, it is not possible to measure its economic development impact in terms of Gross Domestic Product.1 As a result, this report primarily focuses on measuring the development impact of the programme in terms of its income effects, but also examines a number of other areas. The results are presented at the individual, household, community and aggregate level. What impact does the programme have for individual workers? Some participating workers are employed prior to departure, but the majority (65 per- cent) are not, and for these workers the scheme represents an opportunity to fulfil their productive potential. The income gains vary by country but, in all cases, represent a sig- nificant increase on their earning potential back home. The Pacific-wide factor increase on income is 4.3, while it stands as high as 5.6 in the case of Tonga. Over a six-month employ- ment stint, the average Pacific seasonal worker is remitting approximately A$2,200 while in Australia and transferring A$6,650 in savings home at the end of their stay. While the vast majority (86 percent) of Pacific seasonal workers are remitting money home, a large number are using costly money transfer operators, which capture a higher percentage of their earnings in fees. More women are benefiting from the SWP but, proportionally, their representation remains low (14.4 percent) and the same as when the programme was introduced. Female workers in the SWP are earning slightly less than men, but remitting more. This is despite them having a higher mean level of education than the male workers participating in the SWP. Only 42 percent of Pacific seasonal workers have the opportunity to take part in the formal Add-on Skills Training, however, a much larger proportion learn skills on the job. The 1 Remittances are also not factored into GDP. x vast majority (91 percent) felt the skills they had learned in Australia would improve their employment prospects upon return. Given the substantial wage gains and skills obtained in Australia, Pacific seasonal workers reported being very satisfied with their experience in Australia. What does the programme mean for participating households? At the household level, the programme has a significant impact on savings, but not in- come (excluding earnings from the SWP). In Tonga, SWP households had 169 percent more savings per capita than nonparticipating households. The lack of a significant impact on per capita income is likely a result of the relatively low level of investment due to the lack of sound business opportunities. The impacts on per capita expenditure, on the other hand, were positive and statistically significant in the case of Tonga (37 percent more than other households). This extended to both cash expenditure and the value of own-produced food that was being consumed. The cash expenditure was primarily being channeled into home improvements and the accumulation of durable assets. Participating households in Tonga and Vanuatu were 14-16 percent more likely to have made improvements to their dwellings. The additional income also permitted the purchase of farm equipment, cars and pickup trucks. The SWP has notable impacts on certain human development outcomes for participat- ing households. The proportion of school-age children enrolled and attending classes was 7.7 percent higher in Tonga for those involved with the programme. Meanwhile, the impact of participation on health outcomes was negligible. In terms of the incentives (or disincen- tives) the programme provides for remaining household members to work, there was no notable impact on household labor supply. What are the community-level benefits from the SWP? Participating communities are also clearly benefiting from the significant contributions of returning workers. A larger portion of these contributions are, however, going to the church rather than to improving community infrastructure or investment in local schools. As a result, the improvements in public infrastructure are less pronounced than could be expected for the level of contributions from returned workers. It is plausible that the church is providing needed community services with these funds, but there is little information available and a need for greater transparency. What have been the aggregate development impacts? At the aggregate level, the SWP has employed 17,320 Pacific Islanders since 2012 and delivered approximately A$144 million in net income gains to the region. The programme is clearly delivering on its core objective of contributing to the economic development of participating countries, as measured in terms of income. The aggregate development impact to date is most significant for Tonga, Vanuatu, Samoa, and Timor-Leste. The A$99.4 million in net income that Tonga has gained through the programme since its inception is more than double the annual bilateral aid budget from the Australian Govern- ment (DFAT 2017). The A$26.2 million earned in FY17 also represents more than double the xi A$12.4 million generated through Tonga’s exports. For Vanuatu, the A$31.5 million gener- ated amounts to approximately 45 percent of Australia’s annual bilateral aid budget (DFAT 2017). The A$5.8 million in income gains for Samoa and A$5.5 million for Timor-Leste are equally remarkable, given they were later entrants to the programme. How can the development impacts from the programme be enhanced? The programme is clearly delivering significant development gains for most participat- ing countries, but there are further improvements that could be made to the programme to maximize the potential development impacts. The report sets out 11 recommendations that could help further the development impacts of the scheme: 1.) Enhance opportunities for countries with lower rates of participation. Put in place provisions that will help countries with low participation (Nauru, Kiribati, Papua New Guinea, the Solomon Islands, Tuvalu) to gain a foothold in the scheme. Invest additional resources in strengthening sending capacity in these countries through the soon to be launched Pacific Labor Facility. 2.) Ensure poorer areas of participating countries are benefiting from the programme. Ensure that poorer individuals and communities have equal access to information and support in applying for the SWP. Another measure would be to link registries of poorer households with the work-ready pools that sending countries maintain. 3.) Lower barriers to participation for more remote areas. Where feasible, provide decentralized access to passport and visa services, as well as health checks that ensure prospective workers from the outer islands face similar cost structures for participation. Sending countries could also invite Approved Employers (AEs) or their recruiting representatives to visit the outer islands and more remote areas of par- ticipating countries. 4.) Increase the female participation rate in the programme. Given the SWP is a demand-driven scheme, efforts to boost the female participation rate should be concentrated towards altering the perceptions of AEs in Australia and looking at expanding into other industries that would support a higher participation rate amongst women. 5.) Focus recruitment efforts on unemployed labor. For certain sending countries, most seasonal workers being sent through the programme are in formal sector jobs prior to departure. The opportunity cost for this cohort is substantially higher, given the forgone wage earnings and temporary exodus of skills. Sending countries should focus recruitment efforts on the unemployed who constitute a large percentage of the working-age population. xii 6.) Examine the scope for reducing predeparture costs. The cost of flights incurred by workers could be reduced if the incentives were put in place for AEs to book flights in advance. Furthermore, a case could be made for subsidizing the A$280 visa fee and potentially setting up clothing exchanges for departing workers. 7.) Encourage the use of lower-cost Money Transfer Operators (MTOs)/ banks for remitting. Almost all Pacific seasonal workers are remitting using Western Union which remains one of the most expensive options. Further efforts should be made to promote the Send Money Pacific website which compares the costs of different MTOs/ banks, as well as ensuring that basic financial literacy is covered in predepar- ture briefings, which would allow departing workers to make these calculations. 8.) Enhance portability and ease of accessing superannuation. Many Pacific seasonal workers remain unable to access their accrued superannuation in Australia. There is an estimated A$11.4 million in superannuation contributions that workers have not been able to access. The preferred option would be to establish an arrange- ment whereby superannuation earnings in Australia can be easily transferred across the Pacific Provident Fund accounts. Where this is not feasible, workers should be assisted to submit their withdrawal applications. 9.) Provide tailored financial advice and savings options for Pacific seasonal workers upon return. Offer return workers tailored financial advice and options to tie up their savings in investments with higher returns. Ultimately, the money derived from the programme are private earnings, so there should be no requirement that they save or invest, but ensuring that they are fully informed and aware of all options at their disposal would provide a net benefit. 10.) Provide greater transparency around the use of community contributions made through the programme. Put in place measures to ensure that there is a degree of transparency around the expenditure of community contributions from the pro- gramme. This is particularly important for the church which receives a large share of earnings through the programme. 11.) Address the demand-side constraints to increase the number of arrivals. Most importantly, this involves providing an equal footing with the Working Holiday Maker Visa programme. Eliminating or revising the second-year visa extension for Work- ing Holiday (subclass 417) visa holders would remove the incentive in place for the 36,264 backpackers in rural areas, who predominantly work in horticulture. This is not to suggest backpackers should be entirely replaced with Pacific workers (they are also an important source of labor for the horticulture sector), but simply that levelling the playing field would be a step in the right direction. Successfully implementing these recommendations will require a significant commit- ment from the Australian Government agencies responsible for managing the SWP, AEs, as well as the full set of countries participating in the programme. Nevertheless, the additional development gains from their realization are expected to be significant. xiii 1. Introduction Introduction Maximizing the Development Impacts from Temporary Migration 1. Introduction Australia’s Seasonal Worker Programme (SWP) is one of only a handful of migration schemes globally that maintains the explicit objective of contributing to the economic development of labor-sending countries. The link between migration and development has long been recognized and evaluated, but only recently has it been formalized in the global development agenda. The 2030 Agenda for Sustainable Development consists of 17 goals and 169 targets aimed at eliminating extreme poverty and ensuring sustainable devel- opment. Several of these mention migrants or migration (IOM 2017). Seasonal migration programs have been widely adopted globally since the introduction of the Bracero program in the United States in the 1940s and the Gastarbeiter program in the Federal Republic of Germany in the 1950s. While their aim has been primarily focused on plugging seasonal labor shortages, a new wave of programs (including Australia’s SWP) have broadened the focus to incorporate a development objective. The SWP was introduced in 2012 following a four-year pilot scheme, which commenced in August 2008. It started off with a 12,000-worker cap for the initial four years of implementation, but this was removed in 2015 and a host of additional reforms were intro- duced that aimed to address the demand-side constraints of the programme. Since these reforms, the programme has expanded rapidly with 6,166 workers arriving in FY17. This compares to 1,473 workers in FY13 during the initial year of operation. The dual objectives of the SWP are to: (i) contribute to economic development in partner countries by providing employment, remittances and opportunities for upskilling; and (ii) provide benefits to the Australian economy and to Australian employers who can dem- onstrate that they cannot source suitable Australian labor (DoE 2017). This paper builds on the earlier evaluation of the Pacific Seasonal Worker Pilot Scheme (PSWPS) (Gibson and McKenzie 2011) and assesses the extent to which the programme has met the first of these objectives, since it moved from a pilot into a fully-fledged scheme. The programme is open to Fiji, Kiribati, Nauru, Papua New Guinea, Samoa, the Solomon Islands, Timor-Leste, Tonga, Tuvalu and Vanuatu. This evaluation picks up data on all participating countries, except Papua New Guinea, due to the difficulties experienced in ac- cessing workers. The in-country household and community surveys were confined to Ton- ga and Vanuatu, given they represent the vast majority of workers (78 percent of arrivals in FY17). The SWP covers the Australian agriculture, accommodation (in selected locations) and tourism sectors (the Northern Australia tourism pilot). The evaluation is primarily fo- cused on workers in the agriculture sector, given this is where the programme is concen- trated, however, some workers in the accommodation sector have also been sampled. 2 Maximizing the Development Impacts from Temporary Migration The balance of this report is structured into seven sections: • Section Two positions this evaluation within both the global and Pacific-focused lit- erature on migration and remittances, examining the existing evidence on the impacts migration can have on income, expenditure, savings, investment, human capital forma- tion, labor supply, economic growth and inequality. • Section Three details the origins of the programme, the key findings from the PSWPS evaluation, and the evolution of the SWP since its inception in 2012. • Section Four lays out the design of the evaluation and research methodology adopted, including the sampling strategy for the household, community, and on-site worker sur- veys. • Section Five then examines the full set of results from the evaluation, breaking them down into individual worker results (determinants of participation; selection into the programme; labor market activity prior to departure; earnings, costs, and remittances; communication; skills transfer; plans upon return; and workers’ own evaluation of the programme); household level results (impacts on income and savings, expenditure, labor supply, human capital, investment and the accumulation of assets); community-level results (the community’s role in worker selection; participating community character- istics; community contributions) and programme-level results (predeparture training; add-on skills training–AOST–and superannuation). • Section Six explores the aggregate development impact of the programme. • Section Seven looks at several measures that could enhance the development impact. • Section Eight draws some general conclusions. Maximizing the Development Impacts from Temporary Migration 3 2. Literature Review 2. Literature Review 2.1 Global Literature There is a large and growing body of global evidence around the impact of migration on economic and social development (Ratha, Mohapatra, and Scheja 2011). While the focus of this research has been predominantly on the impacts of migration flows between developing and developed countries, there is an increasing focus on South-South migration. Migration has been found to have a stabilizing effect on income, directly for the migrants themselves and indirectly at the macroeconomic and household level through remittances (World Bank 2006b; Chami et al. 2009). The income gains for the migrants themselves inevitably outweigh remittances (particularly for nonseasonal workers), although this gen- erally receives less attention in the literature.2 The effects transmitted through remittances are also significant–a vast body of evi- dence from across the developing world suggests that ‘remittances reduce the depth and severity of poverty, as well as indirectly stimulate economic activity’ (Ratha, Mohapa- tra, and Scheja 2011). This effect is largely transmitted through a direct increase in levels of consumption. Adams (1991) noted that international remittances have a small, but posi- tive effect on poverty. In this study, the number of ‘poor households declines by 9.8 percent when predicted per capita household income includes international remittances’ (Adams 1991, 73). Gupta, Patillo and Wagh (2007) find that, in Sub-Saharan Africa, remittances have a direct poverty-mitigating effect and promote financial development. Fajnzylber and Lopez (2008) note that, in Latin America and the Caribbean, remittances have had an impact on poverty, although this has been relatively modest–an increase of one percentage point in the ratio of remittances to GDP leads to a decline in poverty of about 0.4 percent. Anyanwu and Erhijakpor (2010) determine that, after instrumenting for possible endoge- neity, a 10 percent increase in official international remittances as a share of GDP leads to a 2.9 percent decline in the poverty headcount ratio. The share of remittances spent on investment tends to vary by income level. The relative focus on consumption versus investment goods changes ‘if per capita incomes rise suf- ficiently with the receipt of remittances’ (Ahsan et al. 2014, 67). In terms of investments in entrepreneurial activity, the literature suggests that this is heavily influenced by the business environment in sending countries and the potential for remittances to influence the reservation wage (Ahsan et al. 2014). Woodruff and Zenteno (2007) find that remit- tances result in an increased investment in small businesses in Mexico. In other studies, 2 Income gains dwarf remittances by a factor of up to 12.5 in one recent study (World Bank 2016). 6 Maximizing the Development Impacts from Temporary Migration the evidence is mixed–Amuedo-Dorantes and Pozo (2006) find that remittance-receiving households do not necessarily invest more heavily in family-owned businesses. Meanwhile, there is no evidence that seasonal migration to New Zealand has generated new businesses or higher levels of self-employment in Tonga or Vanuatu (Gibson & McKenzie 2010). While most literature points to migration and remittances having a negative impact on labor supply in sending countries, there is some evidence to the contrary. Kim (2007) finds that, in the case of Jamaica, households with a remittance income have a higher reservation wage and reduce labor supply by moving out of the labor force. Jadotte et al. (2015) determine the effect on both labor participation and hours worked to be nega- tive in Haiti, although the magnitude of these effects may be more severe in urban than rural areas. Cox-Edwards and Rodriguez-Oreggia (2009) separate sporadic from persis- tent remittances and find a neutral effect on labor force participation in Mexico. Meanwhile, Posso (2012) uses aggregate level data for a panel of 66 developing countries and finds an inverted-U relationship, which suggests remaining household members may increase their labor supply in the initial instance to bear the expenses of migration-related costs. Strong evidence is available on the role that migration and remittances play in human capital formation, both for the migrants themselves and their sending households. The existing literature suggests the effects of migration and remittances on education may differ–the former potentially leading to family disruption that impacts schooling, the latter reducing constraints that hinder human capital investments (Ahsan et al. 2014). Notable studies in Mexico find that the combination of lack of oversight due to absent parents and low expected returns to migration have a negative impact (McKenzie 2006; McKen- zie and Rapoport 2006), while other studies show the remittance effect outweighs this, with households spending more on education at the margin, with notable positive impacts on enrolment and attendance (Cox-Edwards and Ureta 2003; Mansuri 2007; Adams and Cuecuecha 2010). A study examining a sample of 18 countries in Sub-Saharan Africa noted that the improvements in education outcomes could be as high as 4.2 percent at the pri- mary level and 8.8 percent at the secondary level, for a 10 percent increase in remittances (Amakon and Iheoma 2014). Unfortunately, the empirical evidence examining the impact of migration and remittanc- es on health outcomes is more limited. The earlier studies drawing negative links between migration and health outcomes find one of two effects: (i) marginal increases in infant mor- tality in the initial instance; or (ii) greater importation and dissemination of certain dis- eases, such as tuberculosis (Perez-Stable et al. 1986; Kanaiaupuni and Donato 1999). More recent studies (Frank and Hummer 2002; Hildebrandt and McKenzie 2005) provide evidence to the contrary, suggesting that migration from Mexico is correlated with low- er infant mortality rates and higher birth weights of children in sending households. In Indonesia, Lu (2012) also finds that adults in emigrant households were significantly less susceptible to being underweight than those in nonmigrant households. In terms of the health impacts on migrants themselves, Stillman, McKenzie, and Gibson (2007) find that the mental health of Tongan migrants improves with migration. Maximizing the Development Impacts from Temporary Migration 7 Another channel for potential development impacts is at the point of return, where migrants transfer knowledge and technology to their sending communities and coun- tries. The theory holds that return migrants have ‘spillover benefits through transfers of innovative ideas, skills, and knowledge to others in the home country’ (World Bank 2014, 77). The magnitude of these impacts evidently hinges on the skill level of the migrant, as well as the type and nature of work engaged in abroad. Gibson and McKenzie (2010) find that there is considerable knowledge flow from return skilled migrants and that they are more likely to engage in trade deals, invest in business start-ups, and advise companies and the government than they otherwise would have been. The benefits for low-skilled migrants are less clear. A study conducted in Vietnam (ILSSA 2010) amongst low-skilled migrants found that, despite them gaining valuable skills abroad (both professional and language), these were rarely used upon return. While there is firm evidence that migration and remittances have solid first-order effects on income, consumption, human capital formation and, therefore, poverty, there is little evidence that substantial GDP growth in sending countries has been caused by remittances (Clemens and McKenzie 2014). Numerous empirical studies have failed to detect a positive impact on growth resulting from increased remittance inflows. Despite the beneficial impacts that remittances have on poverty, schooling, health, and housing, considerable recent research points to workers’ remittances making little to no contri- bution to economic growth in developing countries. Barajas et al. (2009, 1) undertake an econometric analysis of this issue and conclude that ‘at best, workers’ remittances have no impact on economic growth’. After reviewing the various impacts that remittances have had in developing countries, Chami and Fullenkamp (2013, 50) remark that ‘Perhaps most disappointing is the lack of a remittances-growth success story … given that in some coun- tries remittances have exceeded 10 percent of GDP for long periods of time, one would have hoped to find at least one example of remittances serving as a catalyst for significant economic development’. Multiple hypotheses have been put forward to explain why this may be the case. One commonly cited reason is due to the difficulty of separating the cause from the effect– ‘if remittances react counter-cyclically to growth, then the negative relationship between the two is a result of reverse causality running from growth to remittances, not vice versa’ (Ratha, Mohapatra, and Scheja 2011, 8). Putting this aside, other studies indicate that it could be because: (i) increases in remittances are often due to changes in measure- ment, rather than actual increases; (ii) cross-country regressions have too little power to detect their effects on growth; and (iii) the opportunity cost of losing workers has a negative offsetting impact in sending countries (Clemens and McKenzie 2014). Jongwanich (2007) suggests that, rather than driving growth, the merit of remittances is from their ability to raise income levels for the poor. Other possibilities are that the business and regulatory environment hinders investment by inexperienced investors–which would describe most migrant households–or that the cultural environment makes individual entrepreneurship very difficult. 8 Maximizing the Development Impacts from Temporary Migration As far as the impact of migration on inequality, the evidence varies by country. Lipton (1980) found that not only was inequality a push factor for migration, but it also tended to increase interpersonal and inter-household inequality within and between villages. Since then, a number of regional and country-level studies have found evidence of migration both increasing and decreasing inequality (Black et al. 2005). In many cases, migration raises inequality initially, as only relatively well-off households have the resources to send work- ers abroad. This effect dissipates over time as a diaspora is established and the barriers to, and costs of, migration are reduced (Taylor et al. 2005; Koechlin and Leon 2007; Shen et al. 2010). In summary, remittances are found to have unambiguously positive impacts on poverty through household consumption increases. Other household impacts such as on school- ing, health, housing, labor supply, skills transfer, and business investment are, however, not as clear-cut. Despite the huge expansion in global remittances, researchers have been unable to identify a positive linkage at the macroeconomic level between remittances and economic growth–as measured in terms of GDP. 2.2 Pacific-focused Literature Within the global literature on migration, the Pacific region is relatively well represented and remittances and the role they play in the small-island economies have been well- documented. Bertram and Watters (1985) characterized many Pacific Island Countries (PICs) by their two defining stock-flow relationships: migration-remittances and aid-bu- reaucracy, terming them ‘MIRAB’ states. In these economies, the remittances generated from the diaspora and the bureaucracy sustained–by aid flows–are fundamental to their very existence. Since this conceptualization, numerous early studies examined the role of remittances in the development process. Remittances in the Pacific, as in many parts of the world, have tended to go towards con- sumption (including house construction). Research conducted by Tongamoa (1987) sug- gested that as much as 70 percent of remittances to Tonga was spent on imported tinned and preserved foods, beverages, and tobacco. Hayes (1993) noted a similar trend in Papua New Guinea, with remittance recipients spending 88 percent of all remittance income on food in the small island of Ware. Remittance recipients in the region appear to be increasingly directing these flows towards investment (Connell and Brown 2005). Several studies have noted their use for investment purposes. Brown and Walker (1995) found that ‘a significant proportion of remittances received by Samoan and Tongan households were used for business and farm investments’. In Papua New Guinea, Boyd (1990) noted the shift in the use of remittances from consumption to investment as consumption needs were increasingly satisfied. Maximizing the Development Impacts from Temporary Migration 9 Other important uses of remittances in the Pacific include for community and various social purposes. Shankman (1976) noted the significant sums that have been ‘remitted directly or indirectly for church construction and maintenance, pastors’ incomes, and related activities’. These findings are substantiated by surveys conducted by Brown (1995) that suggested that ’41 percent and 18 percent of Samoan and Tongan migrants’ remit- tances respectively were sent to nonhousehold institutions run by the church, including primary and secondary schools. Social uses primarily consist of ‘meeting the costs of wed- dings, funerals and other ceremonies, but also for the costs of education that may lead to long-term economic gains’ (Connell and Brown 2005). In certain PICs, such as Samoa, the use of remittances to enhance their social status through trying to obtain chiefly titles or parliamentary seats was also noted (So’o 1998). Feeny et al. (2014) took a close look at the impact of remittances on economic growth in Small Island Developing States (SIDS)–that is, countries of small economic and popu- lation size that have a relatively larger ratio of remittances to GDP. Contrary to most recent research, they found that there is a positive growth impact in SIDS; but that this result only holds for SIDS in Sub-Saharan Africa and the Pacific–not in Latin America and the Caribbean. One reason for this different outcome that the researchers briefly explore is that remittances reduce the level of macroeconomic volatility experienced by SIDS. Their results suggest that remittances are associated with lower growth volatility in SIDS, which may mean that remittances create a more stable environment for investors. This result, however, leaves unexplained why the results on economic growth are different for SIDS in Sub-Saharan Africa and the Pacific from outcomes in Latin America and the Caribbean. With the introduction of the seasonal worker schemes in Australia and New Zealand, the region has benefited from two rigorous studies assessing the development impacts of circular migration (Gibson and McKenzie 2010; Gibson and McKenzie 2011). The multi- year evaluation of New Zealand’s Recognised Seasonal Employer (RSE) Scheme in 2010 determined that it had overwhelmingly achieved its goal of promoting development in the Pacific Islands through: (i) raising incomes; (ii) allowing households to accumulate more as- sets; (iii) improving subjective standards of living; (iv) improving school attendance; and (v) delivering monetary benefits at the community level (Gibson and McKenzie 2010). Mean- while, the 2011 evaluation of the Australian PSWPS (the predecessor to the SWP), deter- mined the income gains for participating Tongan households to be approximately A$2,600 (Gibson and McKenzie 2011). While the aggregate development impact was deemed to be small at that time, largely due to the small number of workers (only 215 individuals par- ticipated in the first two years of the PSWPS), they noted the significant potential for the programme in the case that it managed to scale up (as it subsequently has). These results are explored in greater detail in Section 3.2. 10 Maximizing the Development Impacts from Temporary Migration More recently, The World Bank (Curtain et al. 2016) has examined the potential econom- ic benefits through to 2040 if both labor-sending and labor-receiving countries around the Pacific Rim were to significantly remove barriers to labor migration. This research, conducted as a part of the broader Pacific Possible research series, determined that these reforms would deliver significant welfare gains–with migrants’ incomes increasing by around US$13 billion (approximately one-third of Gross National Income) and remittances to the Pacific Islands by around US$900 million. In light of this report’s focus on the programme objective of ‘contributing to the economic development of labor-sending countries’, the question has to be asked: what is meant by ‘economic development’? If economic development is synonymous with economic growth, and economic growth is measured by GDP, the global evidence so far points to the lack of a positive relationship. Alternatively, Gibson and McKenzie (2010, 2011) and The World Bank (2016) see the income effect of migration and remittances as welfare-improving and that this is a measure of economic development. Given the limited time and data availability and relatively small size of the SWP, it is not possible to measure its economic development impact in terms of GDP. This report is, therefore, is largely restricted to measuring the eco- nomic development impact of the programme in terms of its income effects. Maximizing the Development Impacts from Temporary Migration 11 3. Description of the Seasonal Worker Programme 3. Description of the Seasonal Worker Programme 3.1 Origins A preferential migration scheme for the Pacific had been considered long before the SWP was eventually introduced in 2012. Several decades earlier, the Australian Government commissioned the landmark Jackson review (1984) of the aid program. One recommenda- tion that emerged clearly from this exercise was that Australia should consider a special migration program tailored towards the South Pacific (Senate Foreign Affairs Committee 2003). Every subsequent aid review has recommended that Australia provide greater la- bor markets’ access to PICs. Concerns around the impact this may have on the domestic unemployment rate and Australia’s nondiscriminatory aid policy were, however, consid- ered barriers to progress (the Simons Review 2003). Two related trends underpinned a shift in outlook towards a preferential migration pro- gramme for the Pacific: (i) Australia’s unemployment rate; and (ii) worsening shortag- es of labor in the agricultural sector. Between 1996 and 2008 Australia’s uninterrupted period of economic growth led to a downward trend in the domestic unemployment rate, falling over four percentage points (8.6 percent to 4.1 percent). Over that same period, the millennium drought–the country’s worst since European settlement–led to major and permanent shifts away from agricultural employment. The drought was estimated to have wiped out 100,000 jobs in the agricultural sector (Houston 2004; Treasury 2004). Sub- sequent labor shortages were felt across the entire sector, but were particularly acute in horticulture, which is more labor intensive. In 2005, the National Farmers’ Federation (NFF) released its labor shortage action plan. The objective of this plan was to ‘identify realistic solutions that can be pursued…to assist the agriculture industry in resolving labor shortages’ (NFF 2005). The plan recommend- ed the consideration of a guest worker visa targeted towards Pacific Rim countries and those receiving aid from Australia. In December 2005, the government launched an inquiry labelled ‘Perspectives on the Future of the Harvest Labor Force’, which specifically exam- ined the potential of a seasonal worker program (Senate Employment Committee 2006). Despite numerous submissions claiming labor shortages across the industry, the Commit- tee was not prepared to recommend that such a scheme should proceed. This conclusion was drawn on the basis that many growers purported to be advocating for such a channel to simply reduce labor costs, rather than to plug labor shortages. 14 Maximizing the Development Impacts from Temporary Migration The World Bank released its ‘Pacific Islands At Home and Away’ report shortly thereaf- ter (World Bank 2006a) which underlined the economic case for a guest worker scheme for the Pacific. Meanwhile, in New Zealand, momentum had built for the introduction of their own equivalent scheme. In late 2006, the New Zealand Cabinet agreed to the initia- tion of what was labelled the RSE work policy (McKenzie et al. 2008). The policy was set out to specifically address labor shortages in the horticulture and viticulture industries. For the initial year, a cap of 5,000 workers was set on the scheme. This was subsequently lifted to 8,000 workers. The early successes of the RSE, coupled with increasing pressure from Pacific Island governments, resulted in the newly elected Rudd Government initiating a pilot scheme in 2008. 3.2 Pacific Seasonal Worker Pilot Scheme In the lead-up to the Pacific Islands Forum in August 2008, the PSWPS was announced (Doyle and Howes 2015). The pilot was spread over a three-year period from February 2009 through to June 2012 and focused on the horticulture industry. The PSWPS was split into two distinct phases: (i) the first covered the remainder of FY09 and was capped at 100 workers; and (ii) the second was from FY10-12 with a cap of 2,400 workers. Kiribati, Tonga and Vanuatu were the first three countries to sign onto the scheme in November 2008 (TNS Social Research 2011). Tonga was by far the largest beneficiary of the PSWPS, capturing 81 percent of the plac- es offered under the pilot scheme (Figure 3.1). The initial stage of the pilot was confined to the Swan Hill-Robinvale area of Victoria and Griffith in New South Wales, two regions with an existing Tongan diaspora. The first group of 50 Tongan workers provided the country with a first-mover advantage, which it sustained throughout the remainder of the pilot. Figure 3.1 Annual Arrivals under the Pacific Seasonal Worker Pilot Scheme (FY09 – FY12) 1200 V nu tu son l Work rs 1000 Tuv lu 800 Ton 600 Timor-L st 400 00 Solomon Isl nds 200 V nu tu S 00 S mo 0 Tuv lu 00 FY09 FY10 FY11 FY12 P pu N w Guin Ton 00 Fin nci l Y r N uru Timor-L st 00 V nu tu Solomon Isl nds 00 Tuv lu S mo 0 FY09 FY10 FY11 Ton FY12 P pu N w Guin Fin nci l Y r Timor-L st N uru Source: DIBP 2017a. Solomon Isl nds S mo FY10 FY11 FY12 P pu N w Guin Fin nci l Y r N uru Maximizing the Development Impacts from Temporary Migration 15 Vanuatu was the second largest source country for the pilot scheme, but accounted for a mere 7 percent of total arrivals, with severe institutional capacity constraints prevent- ing more Vanuatu workers from participating. Kiribati, the other early signatory, sent no workers in the first year and only 52 over the duration of the pilot. It was inhibited largely by the cost of transport to Australia, meaning growers were required to pay more upfront costs for Kiribati workers than Tongans or Ni-Vanuatu. Papua New Guinea was invited to join the pilot from the outset, but did not officially sign on until FY11 (Luthria and Malaulau 2013). It sent a total of 82 workers over the final two years of the scheme. The other Pacific Island countries to participate in the pilot were Nauru, Samoa, the Solomon Islands, and Tuvalu. Collectively, these four countries sent just 38 workers between them, accounting for 2 percent of total arrivals. The 1,589 Pacific workers who ultimately took part in the PSWPS represented 64 per- cent of the 2,500-worker cap set by the Australian Government (Figure 3.1). The inability of the pilot scheme to reach the cap was at odds with the horticultural labor shortages reported by the NFF (estimated at 22,000 workers). While earnings and remittances for individual workers under the PSWPS were significant (see Box 3.1), the low numbers meant that the pilot translated into minimal aggregate development impact for Pacific Island countries. An employer survey carried out by Hay and Howes (2012) found that the key reason for this shortfall was the lack of an aggregate labor shortage in the horticulture industry. Working holiday makers, primarily from a group of high-income countries in Europe and Asia, were readily available and provided with incentives by the Australian Government to work in horticulture. Moreover, the lack of information about the scheme, and perceived level of cost and risk created additional barriers. Despite these early challenges, the Australian Government did make several adjustments to the parameters over the course of the pilot, aimed at increasing industry demand. The final evaluation of the PSWPS was completed in 2011 and suggested that the pilot has ‘demonstrated it can meet the needs of the horticulture industry for seasonal labor’ (TNS Social Research 2011). The introduction of a fully fledged programme was deemed preferable to extending the pilot. It was thought that this would generate greater industry confidence in the permanency of the scheme and, therefore, the returns for growers from investing in it. 3.2.1 PSWPS Evaluation Findings From 2009–2011, The World Bank conducted a series of on-site surveys of Tongan, Kiri- bati and Vanuatu workers, as well as additional in-country household surveys in Tonga and Kiribati. The subsequent evaluation (Gibson and McKenzie 2010) (Box 3.1) adopted a two-prong approach to assess the development impacts of the pilot scheme: (i) first examining the selection process of workers into the PSWPS; and (ii) subsequently indi- rectly estimating the impact on participating households by using the data collected on incomes earned abroad, costs borne by workers, as well as remittances and savings. The core findings from this earlier evaluation are summarized in Box 3.1. 16 Maximizing the Development Impacts from Temporary Migration Box 3.1 PSWPS Evaluation Core Findings Selection of workers. For participating workers, their households were not differen- tially richer or poorer than other households in their same villages, but were coming from poorer villages. These same households were slightly larger than nonparticipating households. They were also much more likely to have previously gone to Australia, and were significantly more likely to have previously worked for wages. They also rated themselves as more likely to take risks than nonparticipating workers. Earnings, costs, and Remittances. Mean weekly earnings were found to be in the range of A$468-515 for both groups, leading to total net income over six months averaging A$12,000-13,000 (Table B3.1.1). Table B3.1.1 PSWPS Earning, Costs and Remittances (A$) Indicator (AUD) Participating country Tonga Kiribati Average weekly earnings $468 $515 Total earnings net of taxes in 6 months $12,174 $13,385 Average Initial expenses $1,063 $1,644 Average weekly expenses $200 $280 Net earnings less total expenses $5,911 $4,461 Remittances to household members $4,628 - Remittances to non-household members $413 - Net and aggregate development impact. When factoring in the opportunity cost of forgone income at home, the net development impact was calculated to be approxi- mately A$2,600 per participating household, or A$456 on a per capita basis. This amounted to an aggregate income gain of A$343,000 for Tonga, A$28,600 for Kiribati and A$26,000 for Vanuatu (Table B3.1.2). These were considered to be relatively small, with the programme deemed to have yet to meet its full potential. Table B3.1.2 PSWPS Aggregate Development Impacts (A$) Net income gain from first two years Participating country Tonga Kiribati Vanuatu Pacific Seasonal Worker Pilot Scheme $343,200 $28,600 $26,000 Recognized Seasonal Employer Scheme $4,336,200 $189,200 $7,898,000 Source: Gibson and McKenzie 2011. Maximizing the Development Impacts from Temporary Migration 17 3.3 Seasonal Worker Programme The rollout of the full SWP was officially announced by the Australian Government in December 2011 (FMFA 2011). While the programme remained the same in essence, there were changes to both the structure and scale of the SWP. The government opted to trial three new sectors, which were perceived to be suffering from seasonal labor shortages: aquaculture, cotton and cane. Meanwhile, the cap on the number of workers was expanded substantially to 12,000 workers over the four-year period spanning from FY13–FY16 (Doyle and Howes 2015). Within this broader cap, there were also annual caps imposed–with the majority (80 percent) of places allocated to the horticulture sector and the remainder (20 percent) set aside for the trial sectors. The government made a concerted effort to ensure the scheme was more employ- er friendly than the PSWPS by both easing the reporting requirements for Approved Employers (AEs) and providing a single point of contact in government. The uptake, how- ever, was still slow with 1,473 workers arriving in FY13 (74 percent of the annual cap3) and 2,014 workers arriving in FY14 (81 percent of the annual cap) (Figure 3.2). Doyle and How- es (2015) identified six core demand-side constraints to the programme through a wide- spread survey of employers in the horticulture industry. These included: (i) the lack of an aggregate labor shortage; (ii) the additional costs; (iii) excessive risk assumed by AEs; (iv) a lack of awareness amongst industry; (v) the reputation of Pacific seasonal workers; and (vi) the burdensome set of administrative requirements. Figure 3.2 Annual Arrivals under the Seasonal Worker Programme (FY13 – FY17) 7,000 Uncapping of the programme V nu tu 6,000 Tuv lu 5,000 Ton 4,000 Timor-L st Solomon Isl nds 3,000 S mo 2,000 P pu N w Guin 1,000 N uru 0 Kirib ti FY13 FY14 FY15 FY16 FY17 Source: DIBP 2017a. 3 During the initial years of the program the annual cap increased each year. 18 Maximizing the Development Impacts from Temporary Migration The government announced a major set of reforms in 2015, which helped address some of these constraints. The most important of these was the uncapping of the programme. This means there is no longer a constraint on the number of workers who can participate in a given year. Other important reforms included: (i) the removal of minimum hiring periods (from 14 weeks); (ii) the elimination of some of the existing administrative requirements for AEs; and (iii) the introduction of the tourism sector alongside accommodation (only in selected areas). The Labor Mobility Assistance Program (LMAP) was announced just prior to these reforms with the explicit aim of: (i) increasing the number of workers; (ii) ensuring sending communities benefit; (iii) improving the employability of returned workers; and (iv) enhancing partner government capacity to manage worker outflows. These reforms have collectively paved the way for substantial growth under the SWP (Figure 3.2). In FY16, there were 4,490 arrivals and in FY17 this expanded to 6,166. This represents more employment opportunities than are available across the entire formal sector of some participating countries (for example, Nauru and Tuvalu). It is, however, still substantially below the 36,264 working holiday makers working in rural areas (predomi- nantly in the agriculture sector) annually and the number of workers in New Zealand’s equivalent RSE Scheme (close to the 10,500 worker cap for FY17) (DIBP 2017b). Maximizing the Development Impacts from Temporary Migration 19 4. Evaluation Design and Methodology 4. Evaluation Design and Methodology 4.1 Description of Evaluation Methodology The evaluation consisted of three core components. Component One comprised of household surveys in participating countries covering key impacts at the household level. This component was limited to Tonga and Vanuatu, given that they are the two largest participating countries (accounting for 78 percent of arrivals in FY17) and the only two that could provide an adequate sample. The primary target group was households with first-time workers, however, given the limited number of first-time workers across the SWP, returning worker households were also surveyed. The household surveys were carried out in 2015 in Tonga and 2016-17 in Vanuatu. The initial intention was for the household surveys to be carried out simultaneously, however, Tropical Cyclone Pam in Vanuatu delayed their implementation. The household surveys contained both a ‘treatment group’ of participating worker households and a ‘comparison group’ of nonparticipating worker households with similar observable characteristics. Par- ticipating households were located with assistance from the Overseas Employment Divi- sion in Tonga and the Seasonal Employment Unit in Vanuatu. Nonparticipating households were sampled using a segmentation method (detailed in Section 4.24.2 Sampling Strategy and Data Collection). Component Two involved a survey of community leaders that assessed the community- level impacts of the scheme. The community surveys were carried out in parallel with the household surveys and also limited to Tonga and Vanuatu. Component Three involved on-site worker surveys that covered participating workers at the end of their stay in Australia. The entire group of Tongan and Vanuatu seasonal work- ers from the household surveys were covered. In addition, a sample of seasonal workers from the remaining PICs (Fiji, Kiribati, Nauru, Samoa, the Solomon Islands, and Tuvalu) and Timor-Leste were also sampled. The seasonal workers from Papua New Guinea were not available to be surveyed. These surveys were conducted largely in Australia using a team of bilingual enumerators. 22 Maximizing the Development Impacts from Temporary Migration 4.2 Sampling Strategy and Data Collection The main sample for the evaluation consisted of the household surveys in Tonga and Vanuatu. In addition, three community leaders were targeted in each participating and nonparticipating community where the household surveys were taking place. The target sample size for Tonga and Vanuatu was 150 community leaders, although this ultimately depended on the number of villages that could send workers to participate in the SWP. For the on-site worker surveys in Australia, all workers picked up in the household surveys were targeted for Tonga and Vanuatu. For the remaining sending countries, all workers available were surveyed from those countries that sent less than 50 workers in FY15 and only one-half were targeted for those countries that sent more than 50 workers in FY15. The ‘treatment group’ in the SWP development evaluation was not randomly selected, as seasonal workers self-select to apply and are recruited into the scheme. Recruitment is determined on a rolling basis in line with peak employment needs across the Australian horticulture industry. Forward notice of upcoming recruitments can range between two to 10 weeks. Recruitment is also done in small groups, given the incentive employers have in screening workers. Because recruitment is staggered over several months, selecting a random sample of workers from the complete list of workers would have meant waiting until the recruitment was completed. This was not practical for the purposes of the evalu- ation. As such, the treatment group was instead selected out of the list of SWP workers provided by the DoE that were set to depart over the evaluation period 2015-2017. Initially the evaluation team had planned to conduct all household surveys in 2015, however, Tropi- cal Cyclone Pam in Vanuatu delayed the process. 4.2.1 Household Surveys The ‘treatment group’ for the household surveys was formed from returning worker households and some who were participating for the first time. These were located with the assistance of the Overseas Employment Division in Tonga and the Seasonal Employ- ment Unit in Vanuatu. While surveying first-time workers would have been preferable in order to form a true baseline, this was not feasible given the high return rates in the pro- gramme. The island groupings surveyed were Tongatapu, ‘Eua, Ha’apai and Vava’u in Ton- ga. For Vanuatu, the surveys were concentrated on the five most heavily populated islands with seasonal workers – Ambrym, Efate, Malekula, Santo and Ambae. Given the significant costs involved with travelling internally in Vanuatu, it was not possible to cover all islands that are participating in the SWP. The selection of nonparticipating households was more complicated as updated house- hold listings were not readily available for Tonga nor Vanuatu. Furthermore, the high field costs of carrying out complete household listings in these countries combined with the budget constraint made this option unworkable. As an alternative, a segmentation method was adopted for the selection of nonparticipating households. Maximizing the Development Impacts from Temporary Migration 23 It should be noted that nonparticipating households were not selected from the same village to avoid spill-over effects. Instead, a neighboring village without participants in the SWP was randomly selected and a segmentation method adopted. Once the village had been chosen, households were separated into geographic clusters using the satellite feature on Google Maps. A limited number of clusters were then randomly selected and a group of households selected from within these clusters. A higher number of treatment group households than comparison group households were sampled in both Tonga and Vanuatu. A total of 506 household surveys were conducted in Tonga and 244 in Vanuatu. 4.2.2 Community Surveys The sampling strategy for the community surveys was the same across both participat- ing and nonparticipating communities. In each instance, three community leaders were surveyed from the same communities captured in the household surveys. The village chief was surveyed in each community. Where the head of the village was not available, his spouse or counsellor was surveyed along with one school headmaster or teacher and one religious or health worker. A total of 178 community surveys were conducted in Tonga and 139 in Vanuatu. 4.2.3 On-site Worker Surveys For the on-site worker surveys, a large proportion of the Tongan and Vanuatu season- al workers from the household surveys were sampled (Table 4.1). In addition,, a sample of seasonal workers from the remaining PICs (Fiji, Kiribati, Nauru, Samoa, the Solomon Islands, Tuvalu) and Timor-Leste were also sampled. For those countries that sent less than 50 workers in FY15, all workers were surveyed. For those countries that sent more than 50 workers in FY15, only one-half were sampled. It should be noted that the sample of workers from Fiji, Kiribati, Nauru, the Solomon Islands, and Tuvalu are extremely small, and reflective of the numbers these countries were sending under the programme during the survey period.4 Given the small sample size and associated issues with drawing country- level conclusions, these countries have been pooled together for the subsequent analysis as the ‘other PICs’. 4 Fiji and the Solomon Islands subsequently experienced an uptick in participation in FY17 (following the sur- vey period), while participation either flatlined or declined in the case of Kiribati, Nauru, and Tuvalu. 24 Maximizing the Development Impacts from Temporary Migration Table 4.1 On-site worker survey sample Country Sample Fiji 6 Kiribati 7 Nauru 10 Samoa 57 Solomon Islands 8 Timor-Leste 43 Tonga 181 Tuvalu 5 Vanuatu 72 Total 389 Source: World Bank staff calculations. Maximizing the Development Impacts from Temporary Migration 25 5. Results 5. Results 5.1 Individual Worker Results The results reported in this section were obtained from the on-site worker surveys. They cover the impacts for individual workers and examine: (i) the factors driving participation in the programme; (ii) the demographic characteristics of seasonal workers; (iii) how work- ers were selected into the programme; (iv) their labor market activity in sending coun- tries prior to departure; (v) their earnings, costs, remittances, and savings in Australia; (vi) their communication with their families while in Australia; (vii) the skills they have acquired through the programme; (viii) their plans upon return; and (ix) their own evaluation of the scheme. 5.1.1 Determinants of Participation While there are a complex combination of microeconomic and social motivations for migration, the decision to migrate is ultimately a function of the perceived differences in living standards between two geographical locations (World Bank 2006c). The pull factors for the SWP, which drive Pacific Islanders decision to participate are primarily eco- nomic. The most important determinant of participation was the wage gap between Aus- tralia and the Pacific. Approximately 86 percent of Pacific workers cited ‘earning a higher wage’ as ‘very important’ in their decision to participate (Figure 5.1). The potential uses of these earnings were also an equally important consideration, with home improvements, starting a business, and paying for school fees the three primary expenditure categories mentioned (Figure 5.1). The potential to acquire additional skills in Australia and improve English language pro- ficiency were critical factors, but not across all participating countries. For those coun- tries where workers had a higher propensity to have dropped out of school at the primary level, such as Vanuatu, these were important reasons. For those countries where workers were likely to have either completed or partially completed high school, they did not factor this in. Despite the fact that the SWP does not offer a pathway to permanent residency, workers from several participating countries mentioned this as an important factor. It was, for example, cited as ‘very important’ for 71 percent of Samoans. 28 Maximizing the Development Impacts from Temporary Migration Figure 5.1 Higher earnings remain the key determinant for worker participation Relative importance of various factors in determining workers’ participation Source: World Bank staff calculations. 5.1.2 Seasonal Worker Characteristics The observable characteristics that define those participating in the programme are partly determined by sending country screening and selection processes, and partly by employer preferences in Australia. Table 5.1 provides an overview of the average worker from each participating country. Across the programme, men are far more likely to be selected to participate than women. Since the inception of the scheme, 2,378 women have participated or 13.7 percent of total arrivals over the period from 2012–2017. Amongst those sampled, only 9.8 percent (or 38 workers) were women. In terms of seasonal work- ers’ age, some participating countries have had issues with younger seasonal workers and, therefore, explicitly try to recruit older workers as a proxy for maturity. The average age across all participating male workers was 32 years, while it was marginally higher for par- ticipating female workers (34 years). Given that the work in the SWP is primarily low skilled and concentrated in the agricul- ture sector, there is no minimum education requirement. Most workers have completed some secondary schooling, but dropped out prior to Form Seven. The exceptions are Vanu- atu and some of the other PICs (for example, the Solomon Islanders) where many workers have only completed primary school or, in extreme cases, no schooling. This is reflective of significantly lower secondary enrolment rates in these countries. For Timor-Leste, on the other hand, the average seasonal worker possesses a university degree. Given that Timor- ese workers are concentrated in the accommodation sector, this could reflect a prefer- Maximizing the Development Impacts from Temporary Migration 29 ence towards recruiting higher skilled workers with at least some level of English language proficiency. Female workers had also generally completed higher levels of education than male workers. Some 23.6 percent of female workers had completed a university degree compared to 6.3 percent amongst male workers. Table 5.1 Seasonal worker characteristics Participating Gender Age Height Weight Education Marital Country (years) (cm) (kg) Status Samoa Male 31 173.6 91.5 Secondary Married Timor-Leste Male 30 165.3 56.9 Tertiary Single Tonga Male 33 177.7 102.1 Secondary Married Vanuatu Male 33 155.8 68.4 Primary De Facto Other PICs Male 34 168.5 85.1 Secondary Married Pacific average Male 32 168.7 79.6 Secondary Married Source: World Bank staff calculations. In terms of marital status, 62 percent of sampled participating workers are married. They form the majority of workers for most participating countries. Workers from Timor- Leste are less likely to be leaving partners behind (Table 5.1). Meanwhile, a large portion of workers from Vanuatu reported being in de facto partnerships. Marriage rates were also lower amongst female workers with only 47.3 percent reporting to have married compared to 63.8 percent amongst male workers. 5.1.3 Selection into the Programme Most participating workers first heard about the SWP through their connections: either friends (45 percent), family (24 percent) or community leaders (4 percent). The remainder found out about the scheme through the media (22 percent) and a small minority through other sources. As a result, those hailing from nonparticipating communities or communi- ties with fewer workers, where awareness of the scheme is lower, have a disadvantage. In terms of the selection process itself, many workers (32 percent) reported being selected into the programme through friends. This corresponds with evidence that AEs are relying on the team leader from a group of Pacific Islander workers to recruit new workers from their country for the following season. Some 28 percent of seasonal workers were recruited directly by employers, while the remainder (40 percent) were either selected by their family members or community leaders. In the more geographically dispersed countries, location was an important factor in the selection process. In Tuvalu, 60 percent of participating workers suggested that migrat- ing internally was a critical factor in improving their chances of being selected. Many of those workers surveyed had migrated from the outer islands to Funafuti specifically to seek selection into the programme. For Kiribati and Vanuatu, 43 percent and 26 percent of workers respectively felt that internal migration was important for selection. For the remaining countries, this was considered a negligible factor in the selection process. 30 Maximizing the Development Impacts from Temporary Migration While connections abroad are not specifically factored into the selection process, it is clear that those with friends and family in Australia have a far higher chance of being selected (Figure 5.2). For Tongan workers, the majority have at least some connections in Australia. Even for those participating countries with relatively small diasporas, such as Vanuatu, many workers are receiving assistance from either friends or family in Australia. On average 43 percent of all seasonal workers received some help from friends or family in Australia, while a much lower proportion of the overall population in sending countries would have such connections. Figure 5.2 Those being selected into the programme have strong connections in Australia Percentage of seasonal workers who received help from either friends or family in Australia Pacific Average - 43% Source: World Bank staff calculations. Maximizing the Development Impacts from Temporary Migration 31 5.1.4 Labor Market Activity Prior to Departure The development impact for individual workers is partly determined by their labor mar- ket activity prior to departure. For workers who are employed, the forgone wage earn- ings are more significant, meaning the opportunity cost of participation is higher.5 Across different participating countries, employment rates are highly variable (Figure 5.3). In the two largest sending countries, Tonga and Vanuatu, the vast majority of workers are unem- ployed prior to departure. For others, such as Samoa, the other PICs, and Timor-Leste, it is more common for workers to be in employment. Figure 5.3 Employment rates vary across participating countries Percentage of workers employed in the 6 months prior to departure Source: World Bank staff calculations. For those in employment, their status also varies significantly across countries. In the Pacific, rates of informality in the labor market are generally high. In Vanuatu this figure is 90 percent (World Bank 2014b). Seasonal workers are far more likely to be in formal sec- tor employment than national averages. Some 59 percent of seasonal workers who were employed prior to departure were in formal sector jobs. Most employed seasonal workers from Timor-Leste and the ‘other PICs’ were in ‘paid employment’ (as opposed to informal employment) prior to departure. Departing workers from Vanuatu, on the other hand, were mostly self-employed or unpaid family workers (Figure 5.4). Although the majority of the work available through the SWP is concentrated in the agriculture sector and low-skilled, most of the seasonal workers who were employed prior to departure came from the ser- vices sector. This holds particularly true for Timor-Leste (74 percent), and Vanuatu (70 percent). 5 It should be noted that employment rates are generally difficult to calculate across the Pacific, due to the high rates of informality and employment in the subsistence agriculture sector. 32 Maximizing the Development Impacts from Temporary Migration Figure 5.4 Employment status varies for those employed prior to departure Percentage of workers according to their employment status in the 6 months prior to de- parture Source: World Bank staff calculations. 5.1.5 Earnings, Costs, Remittances, and Savings The opportunity to earn higher wages in Australia was the driving motivation for participa- tion across all PICs and Timor-Leste. Seasonal workers are covered by the minimum enti- tlements under the National Employment Standards (FWO 2017). The majority of workers in the programme fall under the horticultural Award, which is A$18.29 per hour. A signifi- cant number of those working in the horticultural sector are also paid using a piecework rate, which is paid per unit picked, packed or sorted, rather than per hour. These workers must earn at least 15 percent more per hour than the minimum hourly rate prescribed in this award (that is, A$21.03 per hour). There is some variability in the Australian net weekly earnings across participating coun- tries–ranging from A$549 in earnings (‘Other PICs’) to A$767 for workers from Tonga (Ta- ble 5.2). This is a function of their employer, the type of work undertaken, whether they were being paid an hourly rate or a piece rate, whether they were a return or first-time worker, and a number of other exogenous factors, such as weather patterns that influ- enced their ability to work without extended breaks. Given that the sample of workers surveyed for Tonga and Vanuatu is significantly larger, the net weekly earnings for these two countries provide a more accurate reflection of what seasonal workers make. For all participating countries, the earnings in Australia under the SWP represent a significant increase on what they are able to make in their home country (Table 5.2). Maximizing the Development Impacts from Temporary Migration 33 Table 5.2 Weekly Earnings Net of Taxes Participating Country At home (A$ equivalent)6 In Australia (A$) Factor increase Tonga $137 $767 5.6 Samoa $133 $727 5.5 Timor-Leste $127 $672 5.3 Other PICs $182 $549 3.0 Vanuatu $252 $743 2.9 Pacific Average $163 $702 4.3 Source: World Bank staff calculations. The expectations of net weekly earnings for workers from most participating countries were in line with the weekly Award rate for horticulture of A$694.90. For many countries where net weekly earnings fell below this mark, workers were dissatisfied with their take- home pay (Figure 5.5). This is particularly true for the ‘other PICs’. Figure 5.5 Net SWP earnings mostly meet workers’ expectations Percentage of workers who felt the net earnings from Australia met their expectations Pacific Average - 78% Source: World Bank staff calculations. Given the disparity in net weekly earnings between seasonal workers from participat- ing countries, the total earnings for a six-month period also differ significantly. With the high rate of returning workers for Tonga and Vanuatu (and, therefore, higher productiv- ity), the average net earnings for workers from these countries for a six-month period was A$19,000-A$20,000. For the ‘other PICs’, the total earnings are substantially lower at approximately A$14,000. The average across the programme was slightly over A$18,000 (Table 5.3). 6 The earnings ‘at home’ reflect the net weekly earnings seasonal workers who were employed prior to depar- ture reported. It should also be noted that the factor increase would be significantly larger for those that were not in formal employment. 34 Maximizing the Development Impacts from Temporary Migration The predeparture expenses for seasonal workers vary across participating countries. The cost of a passport is similar across countries with the exception of Vanuatu. Internal travel is significantly more expensive for countries that are geographically dispersed, such as Tonga and Vanuatu. For countries where workers are predominantly coming from the main island or near the international airport, the internal travel costs are negligible. The cost of the Temporary Work Visa (subclass 403) is A$280 and applicable to all participat- ing countries, although this is subsidized by the government in certain PICs.7 The Approved Employers cover A$500 of the international airfare, with the remainder being picked up by the seasonal worker either upfront or through payroll deductions. For workers that are travelling from more remote locations, this cost can prove significant. Table 5.3 Earnings, costs, remittances, and savings8 Samoa Timor- Tonga Vanuatu Other Pacific Leste PICs Average Total earnings net of $18,902 $17,474 $19,942 $19,318 $14,274 $18,252 taxes in 6 months (per worker) Predeparture expenses Passport $42 $59 $38 $94 $56 $67 Internal travel $54 $51 $119 $320 $23 $152 Police clearance $26 $1 $86 $31 $38 $27 Medical $33 $50 $3 $84 $22 $54 Visa $169 $240 $119 $139 $188 $167 Int. Airfare $754 $1,284 $951 $769 $1,228 $892 Work clothing $108 $87 $122 $60 $47 $80 Other costs $13 $50 — $45 — $35 Total predeparture $1,199 $1,822 $1,438 $1,542 $1,602 $1,474 Ongoing weekly expenses in Australia Accommodation $121 $121 $121 $121 $121 $121 Food $43 $43 $43 $43 $43 $43 Transport $33 $33 $33 $33 $33 $33 Telephone $24 $20 $24 $22 $31 $23 Health insurance $52 $52 $52 $52 $52 $52 Airfare deduction $29 $46 $36 $34 $47 $34 Total expenses for 6 months $7,826 $8,190 $8,034 $7,930 $8,502 $7,956 Net earnings less total expenses $9,851 $7,462 $10,470 $9,846 $4,170 $8,822 Remittances and Savings HH members $3,997 $2,665 $1,459 $683 $2,419 $2,029 Non-HH members 0 0 $924 $141 $76 $141 Savings $5,854 $4,797 $8,087 $9,022 $1,675 $6,653 Source: World Bank staff calculations. 7 While this is a standard expense of $280, the amounts reported in Table 5.3 are simply what workers have reported paying and may factor in government subsidies for the visa cost. 8 Certain expenses have been equalized (accommodation, food, and transport) given the high variance across participating countries, despite similar work locations in Australia. The airfare deduction has been calcu- lated by total reported international airfare divided by 26 weeks. Remittances to household members and nonhousehold members have been calculated using the total reported remittances to those groups, divided by total weeks spent in Australia and multiplied by 26 to standardize across a six-month period. Maximizing the Development Impacts from Temporary Migration 35 In terms of their ongoing expenses, seasonal workers are primarily incurring expenses for their accommodation, weekly food needs, transport to and from the farm or hotel, and for the calls made home. In addition, deductions are made by the AEs for their interna- tional airfare and health insurance premiums. These costs are more or less constant across countries, with a couple of exceptions–telephone bills and deductions for airfares, which are slightly higher for workers from more remote participating countries. In deducting the total expenses from net earnings, we derive an estimate of how much workers could stand to make from six months’ employment in Australia (Table 5.3). Table 5.4 The average male seasonal worker stands to earn slightly more through the SWP Net earnings, expenses, remittances and savings for male and female seasonal workers Earnings, costs, Gender remittances and savings Male seasonal Female seasonal workers (A$) workers (A$) Total earnings net of taxes in 6 months $18,918 $17,479 Predeparture expenses $1,205 $1,536 Total expenses in Australia for 6 months $8,827 $8,735 Net earnings less total expenses $8,886 $7,208 Remittances and Savings HH members $2,082 $2,136 Non-HH members $142 $144 Savings $6,662 $4,928 Source: World Bank staff calculations. The average male worker in the programme stands to earn slightly more than the aver- age female worker (Table 5.4). It is likely that these differences arise because of the divi- sion of labor while in Australia. For those working in the agriculture sector, men are often given jobs picking (which are paid using a piece rate), while women are often allocated jobs sorting fruit and vegetables (which are paid an hourly rate). The predeparture expenses incurred by females are slightly higher, but these are primarily due to higher internal travel costs. The females surveyed were largely from Vanuatu, so this reflects country-specific differences rather than gender differences. The ongoing expenses for female workers in Australia were approximately the same as for males, except for the weekly amount spent on communication. Female workers spent an average of A$28 per week communicating with their families compared to A$23 for men. Female workers were, on average, likely to remit more, but save less than their male counterparts. 36 Maximizing the Development Impacts from Temporary Migration Figure 5.6 The vast majority of workers remit money while in Australia Percentage of workers who remitted money home while in Australia Pacific Average - 86% Source: World Bank staff calculations. The vast majority of workers remit money while in Australia, although the frequency with which workers’ remit varies across countries. Across the programme, 50 percent of workers remit money home more than once a month. For Samoa, Tonga, and the ‘Other PICs’, the majority of workers remit more than once a month. Whereas in Vanuatu, a large proportion of seasonal workers only remit once during their entire employment stint in Australia (Figure 5.7). Figure 5.7 The frequency with which workers remit money is highly variable across countries Percentage of workers that remit at different intervals while in Australia Source: World Bank staff calculations. Maximizing the Development Impacts from Temporary Migration 37 There appears to be a relatively weak correlation between the cost structure of sending remittances and the frequency with which seasonal workers send money home to their families (Table 5.5). Nearly all participating workers in the SWP (97.8 percent) reported using Western Union to remit money to their families. While Western Union is the most widespread and well-recognized Money Transfer Operator (MTO) in the Pacific, it is by no means the most cost effective. The difference between the total cost of remitting A$500 using Western Union and the lowest cost option (KlickEx) in Tonga, for example, is 11.5 percentage points. Meanwhile, in certain countries Western Union is the most expensive option. Table 5.5 Seasonal workers are predominantly using Western Union to remit9 Fiji Kiribati Samoa Solomon Tonga Tuvalu Vanuatu Highest Cost NAB Western Westpac Islands Westpac Western Xpress Option (13%) Union (10%) (18.5%) NAB (18.1%) (17.4%) Union (10%) Money (19.9%) Lowest Cost XendPay ANZ KlickEx Compass KlickEx Money Gram Compass Option (2.2%) (1.8%) (0.7%) (6.5%) (0.4%) (7%) (6.5%) Seasonal Western Western Western Western Western Western Western workers’ choice Union (7.1%) Union (10%) Union (8.8%) Union (13.2%) Union (11.9%) Union (10%) Union (10.7%) Source: Send Money Pacific 2017. The Australian and New Zealand Governments support an initiative called Send Money Pacific, which provides a cost comparison for sending money from Australia, New Zea- land and the United States to eight PICs. Unfortunately, despite the existing efforts to promote the service and the user-friendliness of the platform, only 9 percent of all workers surveyed had heard of the service. Of those who had heard of it, only 6 percent had used it to determine the lowest cost remittance option. That is two workers out of a total of 389 surveyed. The lack of awareness and use of this service can go a long way in explaining why Western Union remains the preferred option for remitting amongst Pacific seasonal work- ers. There are also potentially other reasons, which could be driving the use of Western Union, including the accessibility of their network and brand trust. 9 Send Money Pacific does not contain data for Nauru. 38 Maximizing the Development Impacts from Temporary Migration 5.1.6 Communication Given the design of the SWP, workers are separated from their families for periods of up to six months for most participating countries (Fiji, Papua New Guinea, Samoa, the Solo- mon Islands, Timor-Leste, Tonga, Vanuatu) and up to nine months for others (Kiribati, Nauru and Tuvalu). With these extended absences, communication is critical to mitigating the potential downside impacts on the family unit. This is partly hindered by the costs of calling the Pacific and Timor-Leste from Australia–costs which are exorbitantly high by global standards. Timor-Leste is the most expensive country in the world to call from Aus- tralia at A$2.85 per minute for a fixed line (Table 5.6). By comparison, the only two coun- tries that compare are Haiti (A$2.60) and Iraq (A$2.10). While Timor-Leste is an outlier, the cost per minute for Vanuatu is also exceptionally high at A$1.95 per minute, which is nearly an order of magnitude higher than those from New Zealand (A$0.21). Table 5.6 Calling Costs to the Pacific from Australia Participating Country Per minute block call rates (Fixed, A$) Samoa $0.90 Timor-Leste $2.85 Tonga $0.90 Vanuatu $1.95 Other PICs $1.58 Pacific Average $1.61 Source: Telstra 2017. Despite the costs incurred, most seasonal workers are communicating with their families at least monthly (Figure 5.8). Moreover, there appears to be minimal correlation between calling costs and the regularity of communication. For example, 91 percent of workers from Timor-Leste are communicating with their family either ‘once a week or more’ or ‘several times a month’. The only countries that appear to have an issue with communication is Vanuatu where some workers are only communicating with their families ‘once a month’ or a couple of times during their entire stay in Australia. Some 96 percent of seasonal workers used ‘telephones’ to communicate with their fami- lies, while Skype (3 percent) and e-mail (4 percent) were far less frequent forms of com- munication. The main reason cited for using telephones as a form of communication was the cost, but the lack of access to Internet and slow speeds throughout the Pacific are also likely to have been an influencing factor. Only 1 percent of seasonal workers did not have access to a telephone in Australia. Moreover, although many were located in remote regions of Australia, telephone reception did not appear to be an issue–with 98 percent of workers suggesting that telephone reception on the property was strong. Maximizing the Development Impacts from Temporary Migration 39 Figure 5.8 The majority of seasonal workers communicate with home at least monthly Intervals with which workers communicate with their families (percentage) Source: World Bank staff calculations. 5.1.7 Skills Transfer A critical aspect of the SWP is centered on the development and transfer of skills to par- ticipating countries. This was the driving motivation behind the introduction of the add-on skills training (AOST) component, which offers first-time workers the opportunity to re- ceive training in English literacy, numeracy, information technology and first aid (covered in Section 5.4.2). Most workers (74 percent) felt that they had acquired useful skills while in Australia. The skills learned were primarily picked up on the job, including picking, packing, sorting and pruning skills. A large proportion of participating workers were also afforded opportunities to learn other transferrable skills, such as the ability to operate a forklift or tractor. Many workers also reported having improved their English language proficiency because of their participation in the programme. Financial literacy, ICT, and numeracy skills were mostly confined to those who had the opportunity to take part in the AOST. Figure 5.9 The most common skills acquired in Australia are those picked up on the job Percentage of workers reporting the acquisition of a certain skill Source: World Bank staff calculations. 40 Maximizing the Development Impacts from Temporary Migration Across the programme, most workers (91 percent) felt as if the skills acquired in Aus- tralia could be put to good use in their home countries. This held particularly true for Sa- moa, Vanuatu, and Timor-Leste where the vast majority of workers suggested the skills obtained would help improve their employment prospects upon return (Figure 5.10). Figure 5.10 The skills acquired are deemed to improve employment prospects upon return Percentage of workers who feel the skills acquired in Australia will improve employment prospects Pacific Average - 91% Source: World Bank staff calculations. 5.1.8 Plans Upon Return An important aspect of the programme revolves around both: (i) what seasonal workers intend to do during the six-month return period; and (ii) whether there is an exit strategy once workers retire from the programme. In terms of workers’ plans during the return period, many use their earnings from the programme to sustain their livelihoods (Figure 5.11). This is to be expected, given both the large income gains and lack of employment opportunities in many of the participating countries. This is of particular note in Samoa and Tonga, where the majority of workers are not working during the return period. It is important to note that, while these workers are not engaged in formal employment, many of them are engaged in productive activities, such as rebuilding their dwellings or subsist- ence agriculture. Maximizing the Development Impacts from Temporary Migration 41 Almost all workers who were employed in the six months prior to departure intended to either continue with their previous occupation or to commence a new job, which would only require their services for six months of the year. This is a positive result, which dem- onstrates that the programme is not fundamentally altering Pacific Islanders’ motivations to work upon return. Some seasonal workers were exchanging roles with members of their extended family. With this arrangement, one would depart for Australia for six months and their remaining household member would cover their responsibilities at home. On their return, that extended family member would depart for Australia and the seasonal worker would cover their responsibilities upon return. This type of arrangement was particularly common for Samoan seasonal workers. A smaller cohort of workers intended to engage in further study during the six-month return period. This was particularly important for Timor-Leste where 23 percent of workers reported this as their intention. Figure 5.11 Earnings are sufficient to sustain workers’ livelihoods during the return period Workers intentions for the six-month return period (percentage) Source: World Bank staff calculations. 42 Maximizing the Development Impacts from Temporary Migration Many participating workers are in the programme for long-term employment (Figure 5.12). For all participating countries, apart from Vanuatu, the majority of participating workers intended to remain in the programme for as many years as possible. This is re- flected in the high rates of returning workers. Figure 5.12 Most participating workers are in the programme for long-term employment Workers intentions to stay on in the SWP for varying periods (percentage) Source: World Bank staff calculations. For the more established participating countries, such as Tonga and Samoa, the percent- age of returning workers is substantially higher than for those countries that are rela- tively new entrants and are trying to gain a foothold in the programme. For those coun- tries with return workers, I-Kiribati have on average returned for the highest number of seasons (3.5), compared with 3.47 for Vanuatu and 3.14 for Tonga. The development gains from the programme are also more widely spread in countries where new workers have an opportunity to participate, rather than having the same workers return every season. 5.1.9 Workers’ Evaluation of the Programme Given the substantial wage gains and skills obtained in Australia, the vast majority of Pacific seasonal workers reported being very satisfied with their experience in Australia. On a scale of 1 to 10 where 1 is ‘not at all satisfied’ and 10 is ‘extremely satisfied’, the aver- age was 8.6 across all participating countries (Figure 5.13). For Tonga, the largest sending country in the programme, the average was 9.9. Maximizing the Development Impacts from Temporary Migration 43 Figure 5.13 The majority of seasonal workers are satisfied with their experience in Australia Workers reported level of satisfaction with their experience on a scale of 1 to 10 Pacific Average - 8.6 Source: World Bank staff calculations. As a result of their positive experience in Australia and high reported levels of satisfac- tion, participating workers are willing to recommend the SWP to others in their village (Figure 5.14). This endorsement of the programme is near universal across participating countries with only 9 percent of participating workers stating that they would not suggest it to others from their community. Figure 5.14 Given their satisfaction, workers are willing to recommend the SWP to others Percentage of workers who would recommend the programme to others in their village Pacific Average - 98% Source: World Bank staff calculations. 44 Maximizing the Development Impacts from Temporary Migration 5.2 Household-level Results 5.2.1 Methodology The impacts of the programme at the household level were estimated using the Propen- sity Score Matching (PSM) technique. This takes explicit account of the potential endoge- neity of SWP workers’ status, as well as some of the limitations imposed by the data. While the original intention was to use baseline and one-year-after end line surveys to compute difference-in-difference (DID) measures of impact, this approach was rejected and impact was assessed using the Propensity Score Matching (PSM) technique. The principal motiva- tion for the change in methodology was the infeasibility of collecting a ‘true baseline’, due to the high return rate of seasonal workers. The average number of seasons worked at the time of the household survey in Vanuatu was 3.47, while in Tonga it was 3.14. If the change in relevant outcomes were being measured between seasons two and three (rather than before joining the scheme and after), then it is likely that the overall impacts of the scheme would have been significantly underestimated. For this reason, household level impacts were estimated by comparing outcomes of par- ticipating SWP households with a ‘comparison’ group of nonparticipating households that were selected based on PSM. The average values of different outcomes were then compared between the two groups to compute the “average treatment effect on the treat- ed” (ATET). This constitutes the measured impact. A limitation of this approach is that it is only able to match on observable characteristics and, therefore, can produce biased estimates of programme-level impacts (in certain circumstances). Nevertheless, given first time worker households were not available to be surveyed ex ante, this was considered the optimal approach. 5.2.2 Income and Savings Household income (excluding earnings from SWP) was computable only in the case of Tonga. The impact of SWP participation on household income is not statistically significant in Tonga (Table 5.7). SWP households in both Tonga and Vanuatu do, however, have higher levels of annual savings than nonSWP households. In the case of Tonga, SWP households had, on average, 169 percent savings more per capita than nonSWP households while, in the case of Vanuatu, the level of savings was 64.9 percent higher (although not statisti- cally significant). Participation in the SWP clearly has a positive impact on savings. Maximizing the Development Impacts from Temporary Migration 45 Table 5.7 Household-level impacts on income and savings Average treatment effects on the treated households Outcome Tonga Vanuatu Impact Statistical Impact Statistical (ATET) significance (ATET) significance Per capita household income -0.142 Not significant - - Per capita household savings 1.69*** Significant 0.649 Not significant (99% level) * p<0.1, ** p<0.05, *** p<0.01 Source: World Bank staff calculations. 5.2.3 Impact on Household Expenditure Participation in the SWP had a statistically significant impact on overall household expenditure for some participating countries (Table 5.8). Average annual per capita cash expenditure amongst SWP household members was higher than among non-SWP house- hold members in both Tonga (43.6 percent) and Vanuatu (55.8 percent, although not statistically significant). With regards to the value of own-produced food consumed, the impact of the SWP was sizable and statistically significant in the case of Tonga, imply- ing that SWP households were able to grow and harvest more food. This impact was not, however, statistically significant in the case of Vanuatu. The impacts on total per capita household expenditure were large and statistically significant in Tonga, where SWP house- holds spent on average 37.2 percent more than non-SWP households. Table 5.8 Household-level impacts on expenditure Average treatment effects on the treated households Outcome Tonga Vanuatu Impact Statistical Impact Statistical (ATET) significance (ATET) significance Per capita cash expenditure 0.436*** Significant 0.558 Not significant (99% level) Value of own-produced food 0.263* Significant 0.283 Not significant consumed (90% level) Total per capita household 0.372*** Significant 1.07 Not significant expenditure (99% level) * p<0.1, ** p<0.05, *** p<0.01 Source: World Bank staff calculations. 46 Maximizing the Development Impacts from Temporary Migration 5.2.4 Impact on Human Capital The household survey provided two measures of human development outcomes: (i) the proportion of school-age children actually attending school; and (ii) the proportion of family members reporting to be in very good health (self-reported). Table 5.9 shows the difference in the mean levels for these outcome variables for treatment and comparison group households. The programme did not have statistically significant impacts on health outcomes in either Tonga or Vanuatu. This is consistent with the existing literature, which does not provide strong evidence that migration significantly impacts health outcomes. In terms of education outcomes, the proportion of school-age children in school was 7.7 percent higher amongst participating households in Tonga and statistically significant at the 90 percent level, however, in the case of Vanuatu, the impact on education outcomes was not significant. Table 5.9 Household-level impacts on education and health outcomes Average treatment effects on the treated households Outcome Tonga Vanuatu Impact Statistical Impact Statistical (ATET) significance (ATET) significance Proportion of school-aged 0.077* Significant 0.034 Not significant children in school (90% level) Proportion of family members 0.046 Not significant -0.052 Not significant in very good health * p<0.1, ** p<0.05, *** p<0.01 Source: World Bank staff calculations. 5.2.5 Impact on Household Labor Supply A key issue for highly remittance-dependent countries is the way in which these inflows affect the willingness to work for nonmigrating household members. In this context, the question is whether SWP households are likely to reduce labor supply due to higher remit- tance inflows. Table 5.10 shows that, while there are some differences between SWP and non-SWP households with regards to the proportion of family members working outside the home in both Tonga and Vanuatu, these differences are not statistically significant. This would indicate that SWP earnings are not negatively impacting household labor sup- ply. Maximizing the Development Impacts from Temporary Migration 47 Table 5.10 Household-level impacts on labor supply Average treatment effects on the treated households Outcome Tonga Vanuatu Impact Statistical Impact Statistical (ATET) significance (ATET) significance Proportion of family members 0.037 Not significant -0.012 Not significant working outside the home * p<0.1, ** p<0.05, *** p<0.01 Source: World Bank staff calculations. 5.2.6 Impact on Home Improvements and the Accumulation of Assets Across both countries, participating households were far more likely to have made im- provements to their dwellings with 16 percent more households upgrading in Tonga and 14 percent in Vanuatu. These impacts were assessed in terms of households’ propensity to improve their dwellings. These impacts were statistically significant at the 90 percent level in both Tonga and Vanuatu. Participating households also had a higher propensity towards both buying and selling major assets.10 As for types of assets purchased, these were predominantly farm equipment and TV sets in the case of Vanuatu, and cars and pickup trucks in the case of Tonga (Table 5.11). Ownership of cattle, pigs, and poultry were not statistically different between participating and nonparticipating households in either Tonga or Vanuatu. 10 For the purposes of the survey, major assets were considered those with a marked value of more than A$200. 48 Maximizing the Development Impacts from Temporary Migration Table 5.11 Household-level impacts on home improvements and the accumulation of assets Average treatment effects on the treated households Outcome Tonga Vanuatu Impact Statistical Impact Statistical (ATET) significance (ATET) significance Major improvements to dwelling 0.159** Significant 0.140* Significant (95% level) (90% level) Bought or sold a major asset in the last 12 months 0.240** Significant 0.275** Significant (95% level) (95% level) Per capita liquid assets 0.210 Not significant 0.483 Not significant Ownership of TV sets 0.041 Not significant 0.139** Significant (95% level) Ownership of generators 0.009 Not significant 0.012 Not significant Ownership of cars and pickup trucks 0.165** Significant 0.024 Not significant (95% level) Ownership of farm equipment -0.005 Not significant 0.203** Significant (95% level) Number of pigs owned 0.279 Not significant 0.242 Not significant Number of chickens owned 0.404 Not significant -2.29 Not significant Number of cows owned -0.252 Not significant 0.230 Not significant * p<0.1, ** p<0.05, *** p<0.01 Source: World Bank staff calculations. 5.2.7 Investment and Impact on Business Ownership Earning money to start a business was one of the key determinants for participation in the SWP for the majority of seasonal workers (see Section 5.1.1 Determinants of Participa- tion). Some 100 percent of participating Tongan workers cited it as ‘very important’, while 54 percent of seasonal Vanuatu workers mentioned it as being either ‘very important’ or ‘important’. The data collected through this evaluation suggests that, while not all partici- pating workers are realizing this objective, some are–7.1 percent of Tongan households and 30.4 percent of households in Vanuatu reported that the money from the SWP had been used to start a business. For those households that had workers intending to start a busi- ness but eventually did not, the two main constraints cited were ‘not enough remittances’ and ‘no business opportunities’. For these households, it is clear that the expected rate of return on investing remittances in business ventures–allowing for the expected risk pre- mium–is lower than other areas (for example, expenditures on consumption, schooling, or home improvements). Maximizing the Development Impacts from Temporary Migration 49 5.3 Community-level Results While individual workers and their households are the primary beneficiaries of the SWP, the community has a role to play both in the selection of workers and the distribution of benefits provided through the programme. As communal-based societies, a strong ethos of resource sharing is common throughout the Pacific (World Bank 2016). There are few limitations on what resources can be shared, but three methods of resource sharing are particularly important: (i) specialized exchange (whereby individuals or households exchange goods or services of similar value with each other); (ii) generalized reciprocity (when individuals or households provide resources to one another when in need, with some expectation that resources will be returned in the future); and (iii) community collection (whereby contributions are made for ceremonial events or resources collected for commu- nity-wide needs). The latter is particularly important in the case of earnings from the SWP. 5.3.1 Community Role in Worker Selection Through the on-site worker surveys, 16 percent of participating workers reported hav- ing been selected into the programme by community leaders. Through the community surveys in Tonga and Vanuatu, community leaders were asked whether there was a clear selection process in place at the community level. Only 23 percent of communities in Vanuatu reported having one in place, while in Tonga 34 percent of communities reported the same. For the most part, workers in both these sending countries are either recruited through a ‘work ready pool’, directly hired by AEs or taken on through a private sector recruitment agent. In those communities where leaders did have a role in the selection process, certain worker characteristics were prioritized. Given some of the isolated incidents of behavioral issues and workers absconding during the initial years of the programme, maturity was the most critical single factor (Figure 5.16). Other factors at play were income level, age, edu- cation and English language proficiency. Gender was only a minor consideration in Tonga. 50 Maximizing the Development Impacts from Temporary Migration Figure 5.15 Maturity is the most important determinant of community endorsement Characteristics community leaders seek in selecting workers (percentage) Source: World Bank staff calculations. Some community leaders felt that there was room for improving the community’s role in the selection process. Specific concerns cited were that it discriminates against younger workers and people living with a disability. Despite the fact that gender did not emerge as an important consideration in the community selection process, some community leaders also felt that women were being discriminated against. This was more of a concern for community leaders in Vanuatu than Tonga. 5.3.2 Participating Community Characteristics SWP communities do not necessarily have better infrastructure or access to services than nonparticipating communities. In terms of access to public transport, the differences between participating and nonparticipating communities were negligible (Table 5.12). The same held for the quality of roads leading into the village and access to the piped water network. The one area that participating communities were better serviced in was access to infor- mation. Remote communities are still relying largely on newspapers and the mail service as their primary source of information. In both Tonga and Vanuatu, participating com- munities received mail on a more regular basis and were also significantly more likely to be able to buy a daily newspaper. This suggests that one of the barriers to participation for more remote communities could be access to relevant information about the programme –73 percent of community leaders from nonparticipating communities in Vanuatu sug- gested the main reason members of their community had not applied, was because that they were unaware of the requirements for the programme. Maximizing the Development Impacts from Temporary Migration 51 Table 5.12 Participating communities are better equipped in some areas, but not all Community characteristics for both participating and nonparticipating communities (percentage) Community characteristic Vanuatu Tonga Participating Nonparticipating Participating Nonparticipating community community community community Regular bus connection 69% 75% 74% 71% Mail delivered on a daily basis 12% 0% 2% 0% Mail delivered on a weekly basis 12% 0% 11% 0% Village store selling canned goods 93% 100% 97% 100% Access to a daily newspaper 23% 3% 78% 29% Connection to piped water network 73% 56% 91% 100% Village roads sealed with bitumen 28% 9% 85% 100% Source: World Bank staff calculations. 5.3.3 Community Contributions Community leaders in Tonga and Vanuatu clearly conveyed an expectation that season- al workers would contribute to the community in some way upon their return. The aver- age monetary amounts community leaders expected workers to contribute were relatively modest compared to net earnings through the programme–only A$171 for Vanuatu and A$234 for Tonga. By comparison, the actual amounts workers contributed were signifi- cantly higher in the case of Tonga and marginally lower in the case of Vanuatu (Table 5.13). Most community leaders suggested that the earnings and remittances had been spent within their community, however, very few indicated that they had control over how this money was spent–only 6 percent in Vanuatu and 20 percent in Tonga. Table 5.13 Contributions to the community vary widely across participating countries Average contributions to the community from SWP earnings Participating Country Local Currency Units A$ Tonga Community expectation TOP 394 $234 Actual contribution TOP 1,498 $890 Vanuatu Community expectation VUV 14,298 $171 Actual contribution VUV 11,544 $137 Source: World Bank staff calculations. 52 Maximizing the Development Impacts from Temporary Migration Equally as important as the amount being contributed to the community is what these contributions are being used for. In the case of both Tonga and Vanuatu, the majority of households reported ‘donations to local churches’ as being the most important use of the money contributed to the community (Figure 5.16). Other contributions went towards local schools and upgrading community infrastructure in the case of Tonga. Figure 5.16 Most of the community contributions go towards the church Percentage of seasonal workers reporting different community uses for their SWP earn- ings Source: World Bank staff calculations. 5.4 Programme-level Results The evaluation covered a host of programme-specific areas that could help to improve worker’s preparedness for their time in Australia, the skills acquired while in-country, and their ability to access their retirement savings upon return. 5.4.1 Predeparture Briefing As a part of the implementation arrangements set out by the Australian Government for the SWP, sending countries are required to ensure that seasonal workers have access to a predeparture briefing.11 With the exception of some of the ‘Other PICs’ (the Solomon Islands, in particular), it appears that almost all workers in the programme benefited from some form of predeparture briefing (Figure 5.17). 11 The predeparture briefings are expected to address a host of different topics ranging from wages and con- ditions offered to seasonal workers to the role of the Fair Work Ombudsman and how to raise a concern with AEs. Maximizing the Development Impacts from Temporary Migration 53 Figure 5.17 Only in select participating countries are workers missing predeparture briefings Percentage of seasonal workers who took part in a predeparture briefing 100% 100% 98% 95% Pacific Average - 97% 86% Samoa 100% Vanuatu 100% Tonga 98% Timor-Leste 95% Other PICs 86% Source: World Bank staff calculations. The content and quality of the predeparture briefings varied across participating coun- tries. While information on wages and pay deductions was provided for all departing work- ers, coverage of potential health issues and gender roles were less common. The recom- mended timeframe for covering the content set out by the DoE was two days but, in some participating countries, the duration had been cut to 0.5 days due to budgetary issues. This severely impacted the ability of the labor-sending units or nominated training providers to cover all required content. 5.4.2 Add-on Skills Training (AOST) The AOST component was incorporated into the programme to provide seasonal workers with an opportunity to gain additional skills in the areas of basic English literacy, numer- acy, information technology and first aid. First-time workers are allowed to undertake basic training in these areas up to the value of A$825 per seasonal worker. The training is expected to take place during quiet periods of work, after-hours or on weekends, depend- ing on the arrangements between the training provider and the employer. There is considerable variation among SWP participants on whether they had the opportunity to undertake AOST (Figure 5.18). Across the programme, 42 percent of work- ers were able to take part in the AOST. While almost all Samoan workers indicated that they benefited from such training, far fewer workers from Tonga, Vanuatu, and the ‘Other PICs’ were able to take advantage of this opportunity. This was largely due to constraints in setting aside time to undertake training. Many participating workers in the agriculture sector only had Sundays off and would reserve this time for attending church services or calling family members. 54 Maximizing the Development Impacts from Temporary Migration Figure 5.18 Not all participating countries benefit from Add-on Skills Training Percentage of workers who had the opportunity to participate in AOST while in Australia Pacific Average - 42% Source: World Bank staff calculations. 5.4.3 Superannuation AEs under the programme are required to pay superannuation for their workers in accordance with Australian laws. The current super guarantee constitutes 9.5 percent of gross earnings.12 While this may seem marginal, over the course of a six-months’ stint in Australia, it can account for a significant sum. For example, workers from Timor-Leste reported expected superannuation contributions of A$2,300 for six months’ work under the SWP. Despite the good intent of storing away additional savings for Pacific workers’ retirement, many workers struggle to access their superannuation. This was an issue par- ticularly for those countries where participating workers had a lower mean level of educa- tion, such as Vanuatu (Figure 5.19). Figure 5.19 Many workers continue to struggle to access their superannuation Percentage of workers able to access their superannuation upon return Pacific Average - 67% Source: World Bank staff calculations. 12 The flat tax rate of 15 percent also is applied to superannuation contributions, along with an additional 38 percent upon departure. Maximizing the Development Impacts from Temporary Migration 55 6. Aggregrate Development Impact The net development impact of the programme, measured in terms of income, is a func- tion of both the amount remitted and the savings transported home at the end of the workers’ employment stint in Australia. These figures are reported for each participating country in Table 6.1. Another important consideration is the opportunity cost of participa- tion in the programme, which is the forgone wage earnings in sending countries (Gibson and McKenzie 2011). Given the lack of readily available data on formal sector wages in the Pacific, there are two potential ways of calculating this: (i) using the average wages report- ed by workers surveyed through the evaluation; or (ii) using GDP per capita as a proxy for income, as was the methodology adopted for The World Bank’s Pacific Possible projections (Curtain et al. 2016). The former is adopted for the purpose of these calculations. The employment rates prior to departure (detailed in Figure 5.3) are applied to these figures, given for the unemployed, the opportunity cost in terms of wage income is zero. Moreover, the amount earned through household farming production per capita is added to forgone earnings using estimates derived from Household Income and Expenditure Survey data.13 13 These figures have been calculated by dividing total annual household income from home production by the average number of household members and dividing this by two (to provide the equivalent for six months). 56 Maximizing the Development Impacts from Temporary Migration Table 6.1 Aggregate development impact across participating countries Net annual income gains (A$) per participating country since the programme’s inception Samoa Timor-Leste Tonga Vanuatu Other PICs14 Programme wide Number of SWP workers annually FY13 22 21 1,199 119 112 1,473 FY14 162 74 1,497 212 69 2,014 FY15 185 168 2,179 567 78 3,177 FY16 140 224 2,624 1,198 304 4,490 FY17 309 477 2,690 2,150 540 6,166 Total workers 818 964 10,189 4,246 1,103 17,320 Net earnings less total expenses in 6 months (A$) $9,851 $7,462 $10,470 $9,846 $4,170 n.a Forgone wages in sending countries in 6 months (A$) $3,458 $3,302 $3,562 $6,552 $3,198 n.a Forgone earnings from household farming production in 6 months (A$) $171 $96 $243 $439 $142 n.a Employment rate prior to departure 75% 53% 14% 28% 81% n.a Net gain per worker (A$) after accounting for employment/ opportunity cost $7,113 $5,692 $9,759 $7,425 $1,437 n.a Net annual income gain to country (A$, thousands) FY13 $156 $120 $11,701 $884 $161 $13,022 FY14 $1,152 $421 $14,610 $1,574 $99 $17,856 FY15 $1,315 $956 $21,266 $4,210 $112 $27,859 FY16 $995 $1,275 $25,608 $8,896 $437 $37,211 FY17 $2,197 $2,715 $26,253 $15,965 $776 $47,906 Total gains $5,815 $5,487 $99,438 $31,529 $1,585 $143,854 Source: World Bank staff calculations. The forgone wage earnings are the highest in Vanuatu where a large proportion of work- ers were engaged in formal employment prior to departure. These are higher than expect- ed, given GDP per capita is approximately A$3,580, however, this reflects the fact that many workers who were employed prior to departure in Vanuatu were employers them- selves. In general, the net gains per worker are higher for those countries that are drawing on their pool of unemployed labor for the SWP, such as Tonga and Vanuatu. 14 For the purposes of providing a rounded picture of the net annual income gains, the gains for workers arriv- ing from PNG and Kiribati (where data was lacking) have been calculated by assuming the same ‘net gain per worker’ as the ‘Other PICs’ and included under this category in Table 6.1. Maximizing the Development Impacts from Temporary Migration 57 The aggregate development impact to date is significant for Tonga, Vanuatu, Samoa, and Timor-Leste. The A$99.4 million that Tonga has gained through the programme since its inception is more than double the annual bilateral aid budget from Australian Govern- ment (DFAT 2017). The A$26.2 million earned in FY17 also represents more than the A$12.4 million generated through exports in Tonga. For Vanuatu, the A$31.5 million generated amounts to approximately 45 percent of Australia’s annual bilateral budget (DFAT 2017). The A$5.8 million in income gains for Samoa and A$5.5 million for Timor-Leste are equally remarkable, given they were later entrants to the programme. For the ‘Other PICs’ the net gains are negligible, as a result of both the lower net earnings and numbers participating in the programme. For this set of countries, the net annual in- come gains could be expected to increase due to the productivity effect of workers return- ing as the programme continues, and specific measures being put in place to benefit those countries with relatively lower rates of participation (for example, the flexibility offered to Tuvaluan, Nauruan and Kiribati workers to stay in Australia for periods of up to nine months). While it was not possible to calculate these figures for PNG and Kiribati (given the lack of data), it is reasonable to assume that they would approximate those for the ‘Other PICs’. PNG sent a total of 268 workers under the SWP through to FY17, while Kiribati has sent 203 workers. With these assumptions, the total gains for PNG would be A$385,000, while for Kiribati they would be in the order of A$292,000. In total, the gains accrue to just under A$144 million across the Pacific and Timor-Leste since the programme began. 58 Maximizing the Development Impacts from Temporary Migration 7. Increasing the Development Impact of the SWP The evaluation has so far examined the development impact of the SWP for participat- ing workers, households, communities and countries. The programme is clearly delivering significant gains for each, but there are further improvements that could be made to the programme to maximize the potential development impacts. The following lessons (and corresponding recommendations) have emerged from the data collected through this eval- uation and subsequent discussions with both the Australian Government and the relevant ministries covering labor migration throughout the Pacific. Recommendation One: Enhance opportunities for countries with lower rates of participa- tion. The evidence collected through this evaluation and earlier studies (Gibson and McKen- zie 2011; Doyle and Howes 2015) has clearly demonstrated that some participating coun- tries have been able to capitalize on the opportunities presented through the programme, while others have struggled. With a large diaspora in Australia and first mover advantage through the PSWPS, Tonga now accounts for 61 percent of annual arrivals through the SWP. The net income gains for Tonga dwarf those for the second-largest sending country, Vanuatu, by a factor of three and for all other participating countries by almost an order of magnitude. Tonga’s early successes are to be commended, although further efforts need to be made to increase participation from those sending countries that have failed to benefit from the programme to a similar extent. Kiribati, Nauru, Papua New Guinea, the Solomon Islands and Tuvalu are chief amongst these. Since the annual cap on workers was removed in 2015, the SWP no longer represents a zero-sum game for the Pacific. One country’s gain does not signify another’s loss. The LMAP was established in 2014 and is being supported through funding from the Australian aid programme. One of the specific objectives set out was to help improve demand for Pa- cific seasonal workers, with a particular focus on those countries with lower rates of par- ticipation. The activities supported include: (i) developing marketing strategies; (ii) building relationships between AEs and sending countries; and (iii) matching areas of demand with participating countries’ comparative advantages. Unfortunately, while these efforts have not yet resulted in a significant increase in uptake amongst struggling sending countries, they are ultimately expected to bear fruit. The Australian Government has made specif- ic efforts to provide an advantage to the micro-states. Kiribati, Nauruan, and Tuvaluan workers are permitted to work in Australia for periods of up to nine months, for example, rather than the six months allocated for workers from other countries. These special pro- visions will help to make access more equitable. Further incentives could be provided to AEs to hire workers from these countries by lowering their AS500 contribution to workers’ airfares, for example, although this may raise objections from other countries. Ultimately, the aim is to improve design of the programme where possible without going so far as to impose country-specific quotas. The SWP is a demand-driven programme that relies upon commitment from AEs, so putting in place quotas would be a step in the wrong direction. Maximizing the Development Impacts from Temporary Migration 59 Recommendation Two: Ensure poorer areas of participating countries are benefiting from the programme. While the SWP is open to all Pacific Islanders, it is clear that those communities with superior access to information and more effective leaders have a higher probability of being selected. As Gibson and McKenzie (2011) noted from the PSWPS, there are trade-offs between spatial equity and efficiency but, at present, the balance appears to be firmly tilted towards the latter. Given that 58 percent of community leaders in Vanuatu and 85 percent in Tonga suggested the majority of SWP earnings are being spent within their immediate surrounds, it is clear that nonparticipating communities are also missing out. As the SWP is a demand-driven programme, imposing any kind of geography-related or income-related quotes would be the wrong approach and ultimately discourage employ- ers. The main priority should be that poorer individuals and communities have equal access to information about the programme and how to apply. As literacy rates are lower in poor communities, assistance could also be provided through the labor-sending units to help poorer individuals fill out and submit their applications for the SWP. Another step would be to link registries of poorer households with the work-ready pools that sending countries maintain. Tonga, for example, is developing a registry of poorer households for a condi- tional cash transfer program the country is introducing in 2018. The Government of Tonga is planning to use this registry to provide a direct path for poorer households to enter the work-ready pool for the SWP and RSE. Recommendation Three: Lower the barriers to participation from more remote areas. Two of the main issues for remote communities in Tonga and Vanuatu covered through the evaluation are that they are either: (i) unaware that the programme exists; or (ii) mis- informed about the parameters of the scheme and how to apply. Access to information about the SWP is clearly an issue for these communities and an area that labor-sending units should be actively working to address. The Government of Tonga is considering roll- ing out an information campaign to increase awareness of the scheme in the outer islands. Other participating countries should consider similar actions. In some countries, workers are also migrating internally to be selected into the programme. This is both costly and inefficient. Selection processes in these countries (Kiribati, Vanuatu and Tuvalu) should be reviewed to ensure workers from the outer islands have an equal opportunity to be recruit- ed without having to relocate. The other issue for these communities is that when workers can access the programme, they face higher cost structures (and, therefore, lower net in- come gains) than those from the capital city or main island. In Tonga, for example, work- ers from the Outer Islands are required to undergo their health checks in Tongatapu (the main island) prior to departure. This is often difficult to coordinate with their flights out to Australia and so the majority do this two to three weeks before leaving and are individually responsible for covering the associated internal travel and accommodation costs. Where possible, sending countries should be aiming to minimize these types of inefficiencies and provide decentralized access to passport, visa and health services. This would help lower the costs for workers from remote areas. Sending countries could also invite or sponsor employer representatives to visit remoter areas and connect them with community lead- ers who could assist with the initial selection. 60 Maximizing the Development Impacts from Temporary Migration Recommendation Four: Increase the female participation rate in the programme. Addi- tional efforts need to be made to ensure that women are granted equal opportunities to benefit through the programme. The female participation rate in FY17 of 14.4 percent is exactly the same as the female participation rate during the first year of the programme. The scheme has grown substantially, so many more women have benefited, but there has been no convergence in male and female participation rates. This is not for lack of de- mand–many sending countries maintain work-ready pools with substantial numbers of women who are willing and keen to participate. Moreover, the surveys revealed that the bi- ases of community leaders in the selection process is only a minor factor. The much larger issue appears to be the hiring preferences of AEs in Australia and their distinct preference for taking on male workers. As the SWP is heavily concentrated in the agriculture sec- tor, this conforms with gendered stereotypes of suitable employment for men and women. Given the SWP is a demand-driven scheme, efforts to boost the female participation rate should be concentrated towards altering the perceptions of AEs in Australia and looking at expanding into other industries that would support a higher participation rate amongst women. Recommendation Five: Focus recruitment efforts on unemployed labor. For some coun- tries, most seasonal workers being sent through the programme are in formal sector jobs prior to departure (Figure 5.3). In certain instances, workers are coming from skilled posi- tions to undertake low-skilled work in Australia at a higher wage rate. The opportunity cost for this cohort is substantially higher, given forgone wage earnings in the sending countries but, more importantly, it is not clear whether there is a sufficient pool of skilled labor in the Pacific to cover them during their six-month absence (or nine-month absence in the case of Kiribati, Nauru and Tuvalu). Across the Pacific, unemployment remains a major issue. Employment rates are estimated to be less than 50 percent of the working- age population in most countries. Over the past decade, most participating countries in the SWP have seen employment decline at the same time as working-age populations increase (World Bank 2016). Many of the unemployed are less educated, but equally capable of suc- cessfully participating in the programme. If sending countries were able to harness their potential, and ensure that selection processes give them an equal opportunity, this would go a long way in increasing the development impact of the programme. Recommendation Six: Examine the scope for reducing predeparture costs. Many of the costs incurred upfront by workers are fixed by sending-country governments (for example, passport fees, police clearances, and medical check-up fees). There is minimal justifica- tion for subsidizing these fees for Pacific seasonal workers, given that these fees largely cover the administrative cost for providing these services and seasonal worker households are better off than those who do not have the opportunity to participate. The internal and international travel costs are, however, often higher than necessary due to travel arrange- ments being made just prior to departure. AEs also are in charge of arranging international flights and have little incentive to cost save, given their share is fixed (A$500) and the re- mainder is deducted through payroll. In one instance, the AE booked business class flights due to economy class tickets being booked out and still deducted the full amount from workers’ wages. Giving seasonal workers forward notice of their impending travel and put- ting them in charge of selecting flights would help avoid these issues. A case could be made Maximizing the Development Impacts from Temporary Migration 61 for subsidizing the A$280 cost for the Temporary Work (Subclass 403) visa, given Pacific workers are also given a preferential tax rate by the Australian Government. Moreover, the seasonal employment units could look at setting up clothing exchanges for departing workers (many of the clothes purchased for work in Australia are only used during their stint in country). Recommendation Seven: Encourage the use of lower-cost MTOs/ banks for remitting. Al- most all (98 percent) of Pacific seasonal workers are using Western Union to remit money from Australia. With a cash-to-cash rate of 11.9 percent per A$500 transfer, Western Union is the 5th most expensive out of 22 options. The cheapest option in Tonga, by com- parison, is KlickEx which charges 0.4 percent per A$500 transfer. If the sum of Western Union fees and commissions are applied to the aggregate remittances sent through the programme since 2012, this would account for A$16.35 million, making Western Union the third largest beneficiary of the SWP, lagging behind only Tonga and Vanuatu. The Austral- ian Aid program has long supported the Send Money Pacific website, which compares the cost of sending remittances from Australia and New Zealand to a host of Pacific Island labor-sending countries. The website details the fees, total cost (as a percentage) and the speed of transfers for all MTOs and banks. It is a public good for a region, which is heav- ily reliant on remittances. Unfortunately, it is underutilized by both the general public and Pacific seasonal workers. Only 9 percent of Pacific seasonal workers had heard of Send Money Pacific and, out of 389 workers surveyed, only two had actually used it. If the lowest cost option for remittances had been used, this would have generated A$13.56 million for Pacific households since 2012. Send Money Pacific is an excellent service, but it should be promoted both during predeparture briefings and while workers are on-site. Recommendation Eight: Enhance portability and ease of accessing superannuation. Su- perannuation contributions for Pacific seasonal workers are mandatory under Australian law for AEs. These constitute 9.5 percent of total earnings. An estimated total of A$31.6 million in superannuation has been earned since the programme was introduced in 2012, of which A$11.4 million remains inaccessible by Pacific workers. This represents a significant lost opportunity for additional earnings to flow through to the Pacific, which could be rela- tively easily accounted for. Given many Pacific seasonal workers have previously worked in the formal sector in their home countries, they are likely to be members of a national Provi- dent Fund. The preferred option would be to establish an arrangement whereby superan- nuation earnings in Australia could be easily transferred to their Provident Fund accounts. Where this is not feasible, workers should be assisted with submitting their withdrawal applications during reorientation briefings. This is particularly critical for participating countries where levels of literacy (and financial literacy) are lower, such as Vanuatu and the Solomon Islands. Moreover, the additional 38 percent tax on accumulated super upon departure could be reconsidered. 62 Maximizing the Development Impacts from Temporary Migration Recommendation Nine: Provide tailored financial advice and savings options for Pacific seasonal workers upon return. Many workers in the programme are faced with both pres- sures from their family and community to share earnings upon return. In addition, returns on term deposits offered through commercial banks are extremely low throughout the Pa- cific by international standards. As a result, the majority of earnings are used to finance current consumption rather than being allocated towards savings or investment. If Pacific seasonal workers were offered tailored financial advice upon return and options to tie up their savings in investments with higher returns, then this would help generate additional benefits for participating workers and their households in the long-term. Ultimately, the money derived from the programme are private earnings, so there should be no require- ment that they save or invest, but ensuring that they are fully informed and aware of all options at their disposal would provide a benefit. Recommendation Ten: Provide greater transparency around the use of community con- tributions made through the programme. The church is the primary beneficiary of com- munity contributions made through the SWP. Direct contributions to other institutions, such as schools or to other community projects are much less common. While workers’ contributions to community are encouraging, given their significantly higher levels of earn- ings, this would translate into much more support for SWP workers and their left-behind families if community members could observe both the level of contributions workers make and the community activities their contributions fund. For this reason, it would do well for community-level recipients of worker contributions, including the churches, to make both the receipt of contributions as well as their uses, as transparent as possible. This would also have a positive impact on future levels of contributions to the community. The use of coercion in extracting contributions (as reported by some workers) must be discouraged where possible. Recommendation Eleven: Address the demand-side constraints to increase the number of arrivals. All of the above recommendations will mean little if the number of arrivals un- der the programme does not increase. Reducing predeparture costs and shifting to lower cost remittance services amount to improvements at the margin, where additional work- ers have the scope to fundamentally increase the aggregate development impacts of the programme. Doyle and Howes (2015) identified six core demand-side constraints to the programme, including: (i) the lack of an aggregate labor shortage; (ii) the additional costs; (iii) excessive risk assumed by AEs; (iv) a lack of awareness amongst industry; (v) the repu- tation of Pacific seasonal workers; and (vi) the burdensome set of administrative require- ments. Since the programme’s inception in 2012, the Australian Government has made extensive efforts in all of these areas apart from the first. Backpackers still outnumber Pacific seasonal workers in the agriculture industry by almost six to one. The government provides an explicit incentive for backpackers to work in the agriculture industry by offer- ing them a second-year visa extension conditional on them undertaking work in rural areas. Eliminating the second-year visa extension for Working Holiday (subclass 417) visa holders would remove the incentive in place for the 36,264 backpackers in rural areas, who pre- dominantly work in horticulture. A softer option could be to adopt the New Zealand practice of providing only a three-month extension or broadening it to incorporate all sectors (Doyle and Howes 2015). Based on the current participation rate from countries (Table 6.1), if all Maximizing the Development Impacts from Temporary Migration 63 these positions were to go to Pacific Islanders, this would represent an additional A$282 million in net annual income gains for the Pacific. This is equivalent to approximately 26 percent of the Australian Government’s entire aid budget for the region. The government is currently undertaking a comprehensive review of the Working Holiday Maker Programme (DAWR 2016). The current arrangements for the second-year visa extension could be re- considered in the context of the SWP as a part of this review process. This is not to suggest backpackers should be entirely replaced with Pacific workers (they are also an important source of labor for the horticulture sector), but simply that levelling the playing field would be a step in the right direction. 64 Maximizing the Development Impacts from Temporary Migration 8. Conclusion In the space of a decade, the SWP has gone from a pilot of fewer than 100 arrivals an- nually to a programme attracting over 6,000 workers from the Pacific and Timor-Leste each year. In doing so, it has made the leap from being a scheme that was, in many ways, peripheral for industry and sending countries alike, to one that is core to both filling labor shortages in Australia and improving living standards throughout the region. The earlier evaluation of the PSWPS noted that, with only a small number of participating workers, the programme had not yet reached its potential (Gibson and McKenzie 2011). Several re- forms addressing the demand-side constraints of the SWP, including uncapping the pro- gramme and streamlining administrative requirements, have ensured steady growth in arrivals. This growth had underpinned significant development gains for the region. With over 6,000 seasonal jobs available annually, the SWP offers more employment opportuni- ties than the entire formal sectors of several of its participating countries. Some participating workers are employed prior to departure, but the majority are not, and for these workers the scheme represents an opportunity to fulfil their productive potential. The income gains vary by country, but in all cases, represent a significant factor increase on their earning potential back home. These workers are remitting approximately A$2,200, on average, and up to A$4,000 per season in Australia for some sending coun- tries. Moreover, the are remitting regularly. For many participating countries, the savings accrued are substantially higher than the amounts being remitted and average to A$6,650 across the programme. Communication with family occurs, on average, several times a month and this is clearly mitigating some of the potential negative consequences of family separation. Many workers report picking up both job-related skills and soft skills while in Australia – with 91 percent suggesting these skills would improve their employment pros- pects upon return. Employment generation has to take place, however, in the Pacific Island countries for these aspirations to be realized. At the household level, the programme has demonstrated that it can have positive im- pacts on food consumption and total household expenditure. The impacts on per capita household savings are also significant in the case of Tonga, while the effect on longer-term household income is marginal. The programme clearly has an impact on the propensity of households to make improvements to their dwelling and also purchase major assets (greater than A$200 in value). The effects on human development outcomes are signifi- cant for school enrolment in the case of Tonga, but insignificant for health improvements for both Tonga and Vanuatu. Meanwhile, despite widespread concerns that participation in the programme would diminish the motivation to work for remaining family members, the programme is not impacting household labor supply. Maximizing the Development Impacts from Temporary Migration 65 The impact on investment in business is not large, especially in Tonga. While seasonal workers assign a high level of importance to investing in business, the fact that they do not invest upon return should not surprise for two main reasons. First and foremost, business opportunities in the Pacific are often poor. Second, it should be assumed that households make decisions that are in their best interests. For seasonal worker households in the Pa- cific, it is clear that the expected rate of return on investing earnings in business ven- tures–allowing for the expected risk premium–is lower than expenditures on consumption, schooling, or home improvements. Participating communities are also clearly benefiting with significant contributions from returning workers. A larger proportion of these contributions are, however, going to the church rather than to improving community infrastructure or investment in local schools. As a result, the improvements in public infrastructure are less pronounced than could be expected for the level of contributions from returned workers. It is plausible that the church is providing needed community services with these funds, but there is little in- formation available and a need for greater transparency. At the aggregate level, the SWP has employed 17,320 Pacific Islanders since 2012 and delivered approximately A$144 million in net income gains to the region. The programme is clearly delivering on its core objective of contributing to the economic development of participating countries, as measured in terms of income. There are a host of interventions that could help the Pacific extract additional value out of the programme; chief amongst these is increasing the numbers working in Australia annually. This will require reforms that level the playing field between the SWP and other sources of agricultural labor, such as the Working Holiday Maker Programme. The SWP has proven its ability to deliver devel- opment impacts to the region, but now that the programme has been uncapped, there is scope for it to provide further benefits still to the Pacific and Timor-Leste. 66 Maximizing the Development Impacts from Temporary Migration References Adams, R.H. Jnr. 1991. “The effects of international remittances on poverty, inequality, and development in rural Egypt.” Research Report 86. Washington, DC: International Food Policy Research Institute. Adams, R.H. Jnr., and A. Cuecuecha. 2010. “Remittances, Household Expenditure and In- vestment in Guatemala.” World Development 38 (11): 1626-41. Ahsan, A., M. Abella, A. Beath, Y. Huang, M. Luthria, and T. Van Nguyen. 2014. “Internation- al Migration and Development in East Asia and the Pacific.” Washington, DC: World Bank. Amakon, U., and C.G. Iheoma. 2014. “Impact of Migrant Remittances on Health and Edu- cation Outcomes in Sub-Saharan Africa.” IOSR Journal of Humanities and Social Science 19(8), Ver. I., 33-44. Amuedo-Dorantes, C. and S. Pozo. 2006. “Remittances as insurance: evidence from Mexi- can immigrants.” Journal of Population Economics 19(2): 227-254. Anyanwu, J.C. and A.E.O. Erhijakpor. 2010. “Do International Remittances Affect Poverty in Africa?” African Development Review 22(1): 51-91. Barajas, R. Chani, C. Fullenkamp, M. Gapen and P Montiel. 2009. “Do Workers’ Remittances Promote Economic Growth?” IMF Working Paper WP/09/153. Washington, DC: IMF. Bertram, G. and R.F. Watters. 1985. “The MIRAB economy in South Pacific microstates.” Pacific Viewpoint 26(2): 497-519. Black, R., C. Natalai, and J. Skinner. 2005. “Migration and inequality.” Background Paper to the Equity and Development World Development Report. Development Research Centre on Migration, Globalisation and Poverty. Brighton: University of Sussex. Boyd, D.J. 1990. “New wealth and old power: circulation, remittances and the control of inequality in an Eastern Highlands community, Papua New Guinea.” In Migration and De- velopment in the South Pacific, edited by J. Connell. Canberra: NCDS Monograph. No. 24, 97-106. Brown, R.P.C. 1995. “Hidden Foreign Exchange Flows: Estimating Unofficial Remittances to Tonga and Western Samoa.” Asia Pacific Migration Journal 4(1): 35-54. Maximizing the Development Impacts from Temporary Migration 67 Brown, R. and A. Walker. 1995. “Migrants and their remittances. Results of a household survey of Tongans and Western Samoans in Sydney.” Sydney: University of New South Wales, Centre for South Pacific Studies. Chami, R., D. Hakura, and P. Montiel. 2009. “Remittances: An Automatic Output Stabi- lizer?” IMF Working Paper WP/09/91. Chami, R. and C. Fullenkamp. 2013. “Beyond the Household.” IMF Working Paper WP/50/03. Washington, DC: IMF. Clemens, M.A. and D. McKenzie. 2014. “Why Don’t Remittances Appear to Affect Growth?” Policy Research Working Paper 6856. Washington, DC: World Bank. Connell, J. and R.P.C. Brown. 2005. “Remittances in the Pacific: An Overview.” Manila: Asian Development Bank. Cox-Edwards, A., and M. Ureta. 2003. “International Migration, Remittances and School- ing: Evidence from El Salvador.” Journal of Development Economics 72(2): 429-61. Cox-Edwards, A., and E. Rodriguez-Oreggia. 2009. “Remittances and Labor Force Partici- pation in Mexico: An Analysis Using Propensity Score Matching.” World Development 37(5): 1004-14. Curtain, R., M. Dornan, J. Doyle, and S. Howes. 2016. “Pacific Possible - Labor mobility: The ten-billion-dollar prize.” Washington, DC: World Bank. Department of Agriculture and Water Resources (DAWR). 2016. “A review of the Work- ing Holiday Maker Visa.” http://www.agriculture.gov.au/ag-farm-food/working-holiday- maker-review. Department of Employment (DoE). 2017. “Seasonal Worker Programme Implementation Arrangements.” Canberra: DoE. Department of Foreign Affairs and Trade (DFAT). 2017. “2017-18 Total Australia ODA to the Pacific.” Canberra: DFAT. Department of Immigration and Border Protection (DIBP). 2017a. “Seasonal Worker Pro- gramme annual visas granted data.” Canberra: DIBP. ———. 2017b. “Working Holiday Maker (subclass 462) visa programme report.” Canberra: DIBP. Doyle, J. and S. Howes. 2015. “Australia’s Seasonal Worker Program: Demand-side Con- straints and Suggested Reforms.” Washington, DC: World Bank. 68 Maximizing the Development Impacts from Temporary Migration Fair Work Ombudsman (FWO). 2017. “Minimum workplace entitlements.” Canberra: FWO. Fajnzylber, P. and J. Humberto Lopez. 2008. “Remittances and Development: Lessons from Latin America.” Washington DC: World Bank. Feeny, S., S. Iamsiraroj, and M. McGillivray. 2014. “Remittances and Economic Growth: Larger Impacts in Smaller Countries?” The Journal of Development Studies 50(8): 1055-66. Former Minister for Foreign Affairs (FMFA). 2011. “Pacific and East Timor workers help- ing Australian farmers and tourism industry.” https://foreignminister.gov.au/releases/ Pages/2011/kr_mr_111218.aspx?w=GYLX0mNSz4nLQKYuPOSgLQ%3D%3D. Frank, R. and R.A. Hummer. 2002. “The Other Side of the Paradox: The Risk of Low Birth Weight among Infants of Migrant and Nonmigrant Households within Mexico.” Interna- tional Migration Review 36(3): 746-65. Gibson, J., and D. McKenzie. 2010. “The Development Impact of a Best Practice Seasonal Worker Policy.” Policy Research Working Paper 5488. Washington, DC: World Bank. ———. 2011. “Australia’s Pacific Seasonal Worker Pilot Scheme (PSWPS): Development Im- pacts in the First Two Years.” Working Paper in Economics 09/11. Hamilton, New Zealand: University of Waikato. Gupta, S., C. Patillo, and S. Wagh. 2007. “Impact of Remittances on Poverty and Financial Development in Sub-Saharan Africa.” IMF Working Paper, WP/07/38. Washington, DC: IMF. Hay, D., and S. Howes. 2012. “Australia’s Pacific Seasonal Worker Pilot Scheme: why has take-up been so low?” Development Policy Centre, Discussion Paper 17. Canberra: Austral- ian National University. Hayes, G.R. 1993. “The Demographic Situation in Papua New Guinea and Its Policy Im- plications: An International Perspective.” Proceedings of the 19th Waigani Seminar. Port Moresby: University of Papua New Guinea. Hildebrandt, N. and D. McKenzie. 2005. “The Effects of Migration on Child Health in Mexi- co.” Policy Research Working Paper No. 3573. Washington, DC: World Bank. Houston, B. 2004. “Drought costs 100,000 agricultural jobs.” Farm Weekly. Institute of Labour Science and Social Affairs (ILSSA). 2010. “Survey on the Situation of Returning Migrants in Vietnam.” Hanoi: Ministry of Labour, Invalids and Social Affairs. International Organization for Migration (IOM). 2017. “Follow-up and Review of Migration in the Sustainable Development Goals.” International Dialogue on Migration No. 26. Ge- neva: IOM. Maximizing the Development Impacts from Temporary Migration 69 Jadotte, E. and X. Ramos. 2015. “The Effect of Remittances on Labour Supply in the Repub- lic of Haiti.” Discussion Paper 9541. Bonn: Institute for the Study of Labor (IZA). Jongwanich, J. 2007. “Workers’ Remittances, Economic Growth and Poverty in Developing Asia and the Pacific Countries”. UNESCAP Working Paper, WP/07/01. Kanaiaupuni, S.M. and K.M Donato. 1999. “Migradollars and Mortality: The Effects of Mi- gration on Infant Survival in Mexico.” Demography 36(3): 339-53. Kim, N. 2007. “The Impact of Remittances on Labor Supply: The Case of Jamaica.” Policy Research Working Paper 4120. Washington, DC: World Bank. Koechlin, V. and G. Leon. 2007. “International Remittances and Income Inequality: An Em- pirical Investigation.” Journal of Economic Policy Reform 10(2): 123-41. Lipton, M. 1980. “Migration from Rural Areas of Poor Countries: the Impact of Rural Pro- ductivity and Income Distribution.” World Development 8(1): 1-24. Lu, Y. 2012. “Household Migration, Remittances and Their Impact on Health in Indonesia.” International Migration. 51(1): 202-15. Luthria, M. and M. Malaulau. 2013. “Bilateral Labor Agreements in the Pacific: A Devel- opment-Friendly Case Study.” In Let Workers Move: Using Bilateral Labor Agreements to Increase Trade in Services, edited by S. Sáez. Washington, DC: World Bank, 129-47. Mansuri, G. 2007. “Does Work Migration Spur Investment in Origin Communities? Entre- preneurship, Schooling, and Child Health in Rural Pakistan.” In International Migration, Eco- nomic Development and Policy, edited by C. Özden and M. Schiff. Washington, DC: World Bank, 99-140. McKenzie, D. 2006. “Beyond Remittances: The Effects of Migration on Mexican House- holds.” In International Migration, Remittances & the Brain Drain, edited by C. Özden and M. Schiff. Washington, DC: World Bank, 123-47. McKenzie, D. and H. Rapoport, 2006. “Can Migration Reduce Educational Attainment? Evi- dence from Mexico.” Policy Research Working Paper 3952. Washington, DC: World Bank. McKenzie, D., P.G. Martinez, and L.A. Winters. 2008. “Who is Coming from Vanuatu to New Zealand under the New Recognized Seasonal Employer Program?” Policy Research Work- ing Paper 4699. Washington, DC: World Bank. National Farmers’ Federation (NFF). 2005. “Labor Shortage Action Plan.” Canberra: NFF. Perez-Stable, E.J., G. Slutkin, E.A. Paz, and P.C. Hopewell. 1986. “Tuberculin reactivity in United States and foreign-born Latinos: results of a community-based screening pro- 70 Maximizing the Development Impacts from Temporary Migration gram.” American Journal of Public Health 76(6): 643-6. Posso, A. 2012. “Remittances and aggregate labor supply: Evidence from sixty-six develop- ing nations.” The Developing Economies 50(1): 25-9. Ratha, D., S. Mohapatra, and E. Scheja. 2011. “Impact of migration on economic and social development: a review of evidence and emerging issues.” Policy Research Working Paper No. WPS 5558. Washington, DC: World Bank. Senate Employment Committee. 2006. “Perspective on the future of the harvest labor force.” Canberra: Parliament of Australia. Senate Foreign Affairs Committee. 2003. “A Pacific engaged: Australia’s relations with Papua New Guinea and the islands of the south west Pacific.” Canberra: Parliament of Australia. Shankman, P. 1976. “Migration and underdevelopment: the case of Western Samoa.” Boul- der, Colorado: Westview Press. Shen, I-L., F. Docquier, and H. Rapoport. 2010. “Remittances and inequality: a dynamic mi- gration model.” The Journal of Economic Inequality 8(2): 197-220. So’o, A. 1998. “The establishment and operation of Samoa’s Political Party System.” In Political Parties in the Pacific Islands, edited by R. Rich, L. Hambly, and M.G. Morgan. Can- berra: ANU Press, 185-206. Stillman, S., D. McKenzie, and J. Gibson. 2007. “Migration and Mental Health: Evidence from a Natural Experiment.” Policy Research Working Paper No. 4138. Washington, DC: World Bank. Taylor, J., J. Mora, R. Adams, and A. Lopez-Feldman. 2005. “Remittances, Inequality and Poverty: Evidence from Rural Mexico.” Working Paper No. 05-003. Department of Agricul- tural and Resource Economics. Davis: University of California. Telstra. 2017. “International calling.” https://www.telstra.com.au/home-phone/interna- tional-calling#/v. The Simons Review. 2003. “One Clear Objective: Poverty Reduction Through Sustainable Development.” Canberra: Australian Parliamentary Committee to Review Australia’s Overseas Aid Program. TNS Social Research. 2011. “Final Evaluation of the Pacific Seasonal Worker Pilot Scheme.” Melbourne: TNS Social Research. Maximizing the Development Impacts from Temporary Migration 71 Tongamoa. T. 1987. “Migration, Remittances and Development: A Tongan Perspective.” Unpub M.A. thesis. Sydney: University of Sydney. Treasury. 2004. “The impact of the 2002-03 drought on the economy and agricultural employment.” Economic Roundup Autumn 2004. Canberra: Treasury. Woodruff, C. and R. Zenteno. 2007. “Migration networks and microenterprises in Mexico.” Journal of Development Economics 82: 509-28. World Bank. 2006a. “At Home & Away: Expanding Job Opportunities for Pacific Islanders Through Labor Mobility.” Sydney: World Bank. ———. 2006b. “Global Economic Prospects: Economic Implications of Remittances and Mi- gration.” Washington, DC: World Bank. ———. 2006c. “Determinants of Migration.” (http://siteresources.worldbank.org/INTECA/ Resources/257896-1167856389505/Migration_Chapter3.pdf). ———. 2014. “Well-being from Work in the Pacific Island Countries.” Washington, DC: World Bank. ———. 2016. “Systematic Country Diagnostic for Eight Small Pacific Island Countries: Pri- orities for Ending Poverty and Boosting Shared Prosperity.” Washington DC: World Bank. 72 Maximizing the Development Impacts from Temporary Migration The World Bank The World Bank PNG and Pacific Islands 1818 H Street, NW Level 19, 14 Martin Place Washington, D.C. 20433, USA Sydney NSW 2000 Tel: (202) 4731000 Australia Fax: (202) 4776391 www.worldbank.org/en/country/pacificislands www.worldbank.org