Are cash-for-work programmes good for local economic growth? The case of donor-funded public works for refugees and nationals in Jordan

ABSTRACT This article investigates whether public works / cash for work (CfW) programmes contribute to economic growth locally, beyond benefits paid to participants, especially in contexts of flight and migration. Based on quantitative and qualitative research conducted in Jordan, it affirms that CfW substantially promotes economic growth through multiplier effects since CfW participants spend most income locally. Some programmes in Jordan improve also the employability of their participants, which, however, does not transform into higher employment rates because the Jordanian labour market is extremely tight. Finally, the programmes empower women by easing labour-market access and – though not irrevocably – changing gender roles.


Introduction
Public works programmes, also called cash-for-work (CfW), or labour intensive infrastructure programmes, have gained immense interest over recent years because evidence is growing that they have multiple positive effects: In countries such as India, Ethiopia, Indonesia and Peru, they have proven to provide employment, income and social protection to vulnerable households in addition to dearly needed infrastructure in poor communities and sometimes also to improve the skills of participants.International donors have therefore launched CfW schemes also in contexts of post-conflict reconstruction and migration to test whether they unfold similarly positive impacts in less favourable settings.
While evidence on the many positive direct effects of CfW schemes is growing, only little is known so far on their more indirect effects, especially in contexts of crisis and migration.From a theoretical perspective, it is plausible to believe, first, that CfW can have positive effects on social cohesion (Zintl and Loewe 2022)unless it is badly targeted, whereby the effects might even turn negative.Second, we can also hypothesise that economic growth on a local level benefits from CfW through spill-over effects such as multiplier effects.
This article contributes to closing the gap in evidence for the second hypothesis.It presents findings of a multi-method research project on the effects of donor-funded CfW programmes in Jordan conducted during a three-month mission in 2019.These programmes mushroomed after 2016 as a means of support for Syrian refugees and vulnerable Jordanians.
The article shows that CfW can have an economic impact far beyond its direct effects on participantseven in the context of flight and migration.In particular, CfW schemes in Jordan have multiplier effects benefitting entire local communities.Participants spend most of their income locally, and the programmes source part of their building materials locally, though it is difficult to assess the magnitude of these effects.Theoretically, CfW could also generate capacity effects if participants invest parts of their wages into physical or human capital (e.g.land, machines, training).In Jordan, however, these effects have been almost negligible because the existing CfW schemes typically employ their participants for only 3-6 months.The wages are thus just enough to allow participants the repayment of debts but not to make further investments.Likewise, CfW programmes in Jordan have hardly any labour market effects.They help participants to upgrade their soft skills such as commitment, patience or time management, and some also provide technical skillsbut, except partially for female first-time labour market entrants, this does not translate into improved prospects on the primary labour market after CfW employment.Of course, this is at least partly due to the overall lack of employment opportunities on the Jordanian labour market.Still, CfW seems to reduce the in Jordan quite pervasive so-called "shame culture," making work in sectors deemed dishonourable and unacceptable, such as the waste sector, more reputableeven for women.
The article proceeds as follows: section "CfW and economic growth on a local level" discusses key results of the existing theoretical and empirical literature on the effects of CfW on economic growth.Section "Refugees in Jordan" describes the Jordanian context including the situation of refugees.Section "CfW programmes in Jordan" informs about CfW programmes there.Section "Research methodology" lays out the methodology of our empirical research on CfW in Jordan.Then we present our "findings" and "Conclusions and policy implications".

CfW and economic growth on a local level
CfW can be defined as public action providing jobs to poor households at rather low wages, most often for the creation of infrastructure.Its aim is to achieve a "double" or even "triple dividend" by reducing poverty and vulnerability through (i) paid employment, (ii) the creation of public goods (streets, water pipes, sanitation, irrigation, dams and other) in underserved regions andat least in some instances -(iii) the provision of skills training (Gehrke and Hartwig 2018).CfW is thus simultaneously an instrument of labour market policy, social protection and pro-poor infrastructure investment.From a social protection perspective, CfW is a conditional cash transfer tool with the conditionality of hard physical work in the public sphere that contributes effectively to the reduction of poverty and vulnerability (Loewe and Schüring 2021).
CfW is applied in quite different settings: as (i) short-term relief in contexts of conflict and crisis; (ii) stabiliser with a medium-term focus during economic recessions or after natural desasters; and, (iii) a long-term employment guarantee (Roelen, Longhurst, and Sabates-Wheeler 2018).
In academic readings, there is broad consent that CfW pays off only if it generates at least two of the aforementioned three dividends (Gehrke and Hartwig 2018).However, there are significant trade-offs between them: unskilled labour alone cannot build high-quality infrastructure; and short-term jobs are difficult to combine with meaningful skill training.
This constraint may disappear when CfW's more indirect effects are also considered.Once evidence confirms that CfW has substantial indirect effects on economic growth, gender roles and social cohesion, it may be regarded as a key tool of development even if it does not make noticeable contributions to local infrastructure or skills training (Sakketa and von Braun 2019, 44).
We look at economic growth at the local level and see it as a multidimensional process, where quality of life, equity, social inclusion, environmental protection and the reduction of multidimensional poverty are as important as income growth.The International Labour Organization captures it well in its definition of local economic development (LED) as "promoting participation and local dialogue, connecting people and their resources for better employment and a higher quality of life for both men and women" (ILO 2018).
In theory, CfW can contribute to economic growth in six ways: First, by creating public goods and facilitating poor households' access to education, health, water, sanitation and markets it can boost investment, production, trade and consumption.Unfortunately, we have only little empirical evidence so far to say what kind of infrastructure is particularly effective in this regard (Gehrke 2015).
Second, CfW can have positive impacts on income and poverty even beyond the group of participants through so-called multiplier effects.The original multiplier effect has been described in 1936 by John Maynard Keynes (Keynes 2007) and refers to the circulation of money within comparatively closed national economies: since CfW participants are likely to spend most of their income within their communities, they also raise the income of their neighbours.The neighbours also spend much of their additional income locally, raising again the income of other neighbours, and so on.The magnitude of the multiplier CJDS / LA REVUE m can be computed as follows: where s is the share of the income that people in the respective neighbourhood save on average and i is the share of the income that people spend on average outside the community.The significance of the effect depends thus on the share of income spent on locally purchased goods and services (Soifer 2014).
Empirical evidence for the multiplier effect is limited so far (Bhalla, Kangasniemi, and Winder 2021).Keddeman (1998) confirms that CfW participants spend most of their income locally.Tessitore (2013) finds evidence for a small multiplier effect in Somalia.Dejene and Cochrane (2022) show that CfW programmes in Ethiopia have even nation-wide effects, though these are small and short-lived.For other kinds of social transfer programmes, however, multiplier effects were found to be significant (Barrientos 2008;Robinson and Levy 2014).Research on several Sub-Saharan countries has quantified these effects between USD 1.3 and USD 2.6 for each USD 1.00 transferred (World Bank 2015).
Third, CfW can stabilise the income of both, participants and non-participants even in the long-term by encouraging investments.Participants may invest part of their CfW income into physical or human capital and thereby enlarge their future income-generation possibilities (see e.g.Barrientos and Malerba 2020).Even non-participants may increase their investments: As long as low-income populations are vulnerable to manifold risks they are usually risk-averse; they deposit any possible savings at safe places, from where they can withdraw at any time without loss once they have to deal with shocks caused by bad harvests, illness or other risks.They change their behaviour once they have at least basic social protection such as a chance to participate in CfW programmes in emergency cases (the so-called "insurance effect").Only then, people start investing into machines, new modes of production or better education for their kids, which brings about new risks but raises expected future income levels (Gehrke and Hartwig 2018).
Unfortunately, there is little evidence for these two effects, as well.One reason is that most CfW programmes provide only limited-term and low-wage employment such that participants can hardly make any savings (Gehrke and Hartwig 2018).Still, Borga and D'Ambrosio (2019) and Gehrke (2015) confirm that CfW participants and non-participants alike increased their investments when CfW programmes were launched in Ethiopia and India respectively.Similarly, Rosas Raffo and Sabarwal (2016) found that CfW participants in Sierra Leone are four times more likely to establish a new business than non-participants.Barrientos and Malerba (2020) show that CfW programmes in several countries have reduced child labour, increased investment in education and incentivised farmers to diversify into non-farm activities.
Fourth, CfW can improve the future employability of participants and thereby contribute to the reduction of under-and unemployment not only in the short but also in the longer term.Some CfW programmes include an explicit training component but also other programmes promoteintentionally or notparticipants' technical or soft skills like organizing oneself at work.Work experience and skills can be particularly important for women who suffer often from disproportionate labour markets barriers (Jones 2021).For them, participation in CfW can be an asset as suchespecially where many people think that women cannot or should not work in the public sphere (Kabeer 2011).
Empirical studies provide, however, once again, quite limited and contradictory evidence on the employability effect of CfW schemes.The main reason is probably that the utility of skills acquired in CfW schemes varies extremely in terms of nature, degree and context-specific applicability (Gehrke and Hartwig 2018).
Fifth, CfW can raise wage levels on primary labour markets; for instance, MNREGA drove up agricultural wages in India by 4.3 per cent per year (Gehrke and Hartwig 2018).Yet, the effect depends on wage levels paid on the primary labour market: In India, it did not contribute to closing the gender wage gap because available female labour did not become scarce.And in our case study, Syrian participants are banned from entering the primary labour market in many economic sectors.We thus excluded this effect from our analysis.
Sixth, CfW can add to economic development, by fostering social cohesion, e.g.intrasocietal and state-society trust, an inclusive identity and cooperation for the common good (Burchi et al. 2022).This is particularly important in contexts where social cohesion is weak or under pressure.
Evidence on this assumption is particularly weak because testing whether CfW contributes to economic growth through improving social cohesion requires two checks: first, to what extent CfW enlarges the different aspects of social cohesion 1 and, second, to what degree these different aspects of social cohesion support economic development.We therefore disregard this argument in this article.Yet, we discuss and confirm its first part, i.e. to what extent and how CfW affects social cohesion in Jordan, in more depth in another article (Zintl and Loewe 2022).

Refugees in Jordan
Jordan is an upper middle-income economy, which has achieved a comparatively high level of human development although it suffers from rather unfavourable framework conditions (limited water and natural resources, almost land-locked, unstable neighbouring countries, inflow of millions of migrants throughout history, neopatrimonial policymaking, wide-spread favouritism; Loewe, Blume, and Speer 2008).Absolute poverty rates are low but many households live just above the poverty line and are highly vulnerable to shocks.
Jordan suffers from a lack of productive employment: Even before the Covid-19 crisis the official unemployment rate rose from 11 to 23 per cent between 2010 and 2020, but there are much higher shares of youth unemployment and of people in qualitative and quantitative underemployment (Krafft, Assaad, and Keo 2018).
High population growth is partly due to the arrival of approximately 670,000 refugees from Syria, representing 7 per cent of Jordan's total population in 2020 (UNHCR 2020).They add to migrants, who had entered Jordan in previous decades, and their descendants: Palestinians, Iraqis, Yemenis, Sudanese, Egyptians and other.
The presence of the refugees constitutes a considerable burden for Jordan, in terms of infrastructure challenges and weakening social cohesion.20 per cent of Syrian refugees live in refugee camps (almost exclusively in Za'atarī and Al-Azraq) while 80 per cent live in host communities, where they compete with the locals over limited resources and public services, mainly in the areas of water, waste management, energy, transportation, housing and schooling (Schubert and Haase 2018).Many Jordanians complain that water shortages, waste management problems and high housing prices have increased since the Syrians came (NAMA and KAS 2018).In addition, they voice fears about Syrians taking away jobs and houses and overcrowding their schools (Jones et al. 2019).These fears have negatively affected social cohesion (Kuhnt et al. 2017).Open clashes broke out, though, only in a few cases (UNHCR, UNICEF and WFP 2014).
At the same time, Syrian refugees suffer from rising living costs and unemployment.They face significant barriers to entering the Jordanian labour market (Razzaz 2017) because most economic sectors are closed to non-Jordanians (World Bank 2019).Only some sectors such as food processing, handicrafts, and tailoring have been opened up to 200'000 Syrians in 2016 as a result of the so-called Jordan Compactan agreement between Jordan and the European Union, which promises Jordan increased official development aid and enhanced access to the European product market once Jordan has set up a work permit programme for Syrian refugees (Seeberg 2022).However, by 2019, only a third of all Syrians at working age got a work permit and many are still not aware of the fact that they can get one for some sectors (Jones et al. 2019).
Most Syrians therefore try to get a job on the informal labour market, where they compete with other migrants and low-skilled Jordanians.If caught without work permit, they can be sent from their current community to a formal refugee camp or even back to Syria.In 2016, only 20 per cent of Syrians at working-age living outside camps had any job, compared to 32 per cent of Jordanians (Krafft, Assaad, and Keo 2018).
As a result, more than half of Syrian refugee households depend solely on transfers paid by international donors (Jones et al. 2019).Once registered with UNHCR (United Nations High Commissioner for Refugees), they are entitled to a regular cash assistance of 65-360 EUR per month and household depending on the household size; UNICEF grants a top-up of 25 EUR for each child.Several other donors run smaller cash transfer programmes in parallel.On average, Syrian households receive 285 Jordanian Dinar (JOD) per month from wages and transfers, most of which they spend on housing (69 per cent).Health and food each account for just 11 per cent of their total spending, education for 9 per cent, and transportation for 8 per cent.17 per cent of adolescents suffer from hunger, while 35 per cent report chronic illnesses (Jones et al. 2019).

CfW programmes in Jordan
Jordan spends 12 per cent of GDP on social insurance, public health and unconditional social transfer schemes but these schemes benefit the middle class more than the poor (Loewe 2019;Loewe and Jawad 2018).Most of them cover almost exclusively Jordanians: Hardly any refugee is formally employed, which is the precondition for social insurance coverage, and only during the Covid-19 pandemic the social assistance schemes started to provide some benefits to non-Jordanians (Hagen-Zanker and Both 2021).Previously, only the Ministry of Health extended so-called "green-cards" for free treatment in public primary health care facilities not only to very poor Jordanian households but also to Syrian refugees (Hagen-Zanker, Ulrichs, and Holmes 2018).
Foreign donors have therefore set up CfW programmes after 2016 targeting Syrian refugees explicitly but also vulnerable Jordanian households.Until 2019, their total budget came to about EUR 300 million, employing some 70,000 workers during this time.Most programmes were initiated by Germany, Norway, Korea, WFP, UNHCR and UNICEF, but implemented by international or Jordanian non-government organisations, Jordanian government agencies or international organisations.
All employ Jordanians along with Syrians, and women along with men because the programmes are meant not only to (i) provide employment and income to vulnerable households; but also to (ii) strengthen social cohesion between Syrians and Jordanians; and (iii) promote the integration of women into economic life.
In some programmes, women do the same kind of work as men, either together with them or in separate teams.In others, they do slightly different kind of work (excluding hard physical work).And in some schemes, women assume totally different tasks, like cooking for the male participants who work in construction.Women account for 16 per cent of participants in ILO's Labour Intensive Infrastructure Programmes, 23 per cent in GIZ programmes and almost half in WFP programmes (Loewe, Zintl et al. 2020).
The most frequent CfW activities are in the construction, rehabilitation and cleaning of "grey" infrastructure (streets, dams, schools, health clinics) or "green" infrastructure (water reservoirs, irrigation systems, municipal parks, ecosystems), in waste collection and recycling, and intensification of agriculture.They typically take place in communities in the north of Jordan, where most Syrians live, and in the Za'atarī and Al-Azraq refugee camps.
Some design features, such as the number of participants and the duration of employment, used to vary between the projects (30-90 working days) but in mid-2019, most donors agreed on joint standard operating procedures to be applied for all CfW projects in Jordan.All CfW projects adhere to decent work conditions (including minimum wage) and thus offer rare formal employment opportunities to Syrian refugees.Some CfW programmes, such as those of the WFP, provide training for their participants.

Research methodology
In order to answer the overarching research question "What effects has CfW on economic growth on a local level?"and following the theoretical clues from the reviewed literature (see section "CfW and economic growth on a local level"), we formulated four research hypotheses: 1.The creation and maintenance of infrastructure by CfW programmes increases the quality of life and the investment, production and income in the local communities.2. The wages paid by CfW programmes have positive direct effects on the participants and indirect effects on non-participants through multiplier effects.3. The wages paid by CfW programmes have positive effects on the investment behaviour of participants' and non-participants' households.4. Work experience and skills acquired during participation in CfW lead to better employability and employment opportunities after completionespecially for women.
In order to test the hypotheses, we applied a quali-quanti mixed methodology including four main elements (see Loewe, Zintl et al 2020, for more details): (i) Extracting theoretical and empirical literature on the effects of CfW in Jordan and elsewhere.This includes, in particular, a survey by NAMA Strategic Intelligence Solutions among 572 participants of the ILO's CfW programmes (NAMA 2019) and an impact evaluation of German-funded CfW programmes by the German Institute for Development Evaluation (DEval), which includes a panel survey among 667 participants of three GIZ programmes in Jordan (Roxin et al. 2020).(ii) Qualitative analysis of open interviews with 133 experts and observers from the government, public administration, the private sector, civil society, academia and the donor community inside the communities of selected CfW sites and in the Jordanian capital Amman. 2 (iii) Qualitative and quantitative analysis of semi-structured interviews with 247 stakeholders (77 CfW participants, 97 shopkeepers in the local communities and 73 other non-participants).The interviews were conducted in nine rural sites 3 selected on the basis of four criteria: (a) the presence of a significant number of refugees, (b) a concentration of CfW activities with hypothetically large impacts on economic growth due to (c) a small community size and (d) relative isolation from other communities guaranteeing that the effects of CfW programmes do not diffuse too much into neighbouring communities and are hence still measurable.We aspired to cover an even number of Jordanians and Syrians, CfW participants and non-participants, men and women.However, as almost all shopkeepers in our research sites turned out to be male Jordanians, only 72 interviewees were Syrians (while 170 were Jordanians and 5 were Egyptians) and only 82 were women (while 155 were men).To analyse the information gathered by interviews, we applied two techniques.First, we formulated codes for all aspects of our hypotheses, all aspects of the topic and all other issues highlighted by our interlocutors.Then, we ran a systematic qualitative exploration of the content of the interview transcripts with the software programme ATLAS.ti.Second, we counted the frequency of similar responses, statements and arguments made in all interviews to see which of them represented widespread or exceptional opinions. 4For reasons of confidentiality, interview quotations are fully anonymised, only revealing interviewees' respective roles and connection to CfW programmes (participant, non-participant, shopkeeper, project staff, etc.).(iv) Econometric analysis of a census among all 1847 participants of GIZ's CfW Green Infrastructure Programme at the end of their engagement (called "GIZ Post-employment Survey" hereafter).52 per cent of the respondents were Jordanians, 48 per cent Syrians, 26 per cent women, and 74 per cent men.We ran several probit regressions to identify factors determining the perceptions and experiences of the CfW participants. 5

Findings
Our research failed to deliver clear evidence for the alleged positive effects of the public goods created or maintained by CfW programmes in Jordan (Hypothesis 1) but we found clear evidence for the positive direct and indirect effects of the wages paid by CfW programmes (Hypotheses 2-3).Our research also confirmed that some CfW programmes improve the skills and hence the long-term employability of their participants (Hypothesis 4) even though these effects do not transform into higher employment rates.In addition, CfW programmes contribute to female empowerment: They open the door for women to the labour market and contribute to overcoming societal disapproval of them working in public.
Hypothesis 1: The creation and maintenance of public goods by CfW programmes increases the quality of life and the investment, production and income in the local communities.
Some CfW programmes in Jordan produce or maintain public goods that are most likely to have positive long-term effects on economic growth.Many of them contribute to the rehabilitation of dams, water reservoirs and irrigation systems, the intensification of agriculture, and protective measures against soil erosion.However, with our methods, we have not been able to provide clear evidence for the effects of the public goods produced by CfW programmes in Jordan on growth.These effects are difficult to measure and tend to emerge only in the long term, diffuse across the country and mix with the effects of manifold other factors.Here, we have to rely on other studies showing that the infrastructure built by CfW often has significant longer-term effects on local economic development (Gehrke and Hartwig 2018).In addition, several interviewees expected that the community would benefit from the projects even on the long-term: "Even if they do not work with it [the CfW programme] themselves, […] it will help them later in watering their farms.However, not everyone sees this." The goods and services created by other CfW programmes are likely to have rather limited effects on economic growth as such.For instance, waste collection and recycling, embellishment of parks, and tree planting may contribute to the quality of life of local residents but are insignificant for economic activities.Likewise, the renovation of schools and health clinics may matter for the well-being of students and patients but presumably has no major effects on economic growth.Even the building and maintenance of village roads has probably no larger impact on economic growth because the Jordanian road network does not suffer from serious gaps.Furthermore, we did not find reliable ways to separate the effects of the public goods created or maintained by CfW programmes from the effects of public goods created through other means.
However, numerous interviewees stated that the public goods created by CfW programmes were still very important for economic activities in general and their own quality of life in particular.Most of them referred to the rehabilitation of schools, health centres, street lamps and telecommunication lines.Others appreciated cleaner village streets, upgraded municipal parks, nature reserves and tree plantations.A young man said: "The infrastructure has been improved […].It has helped the community."Another one said: "There is a health centre now in Umm Jimal.That is an improvement."And a third one claimed: "There used to be barren land, and now it is becoming a forest, this is attractive to birds and animals.And there will also be a park, people will CJDS / LA REVUE enjoy it."Similarly, in the survey conducted by Roxin et al. (2020), 97 per cent of the interviewed cash workers stated that the CfW outputs were assessed positively throughout their respective host communities.
Hypothesis 2: The wages paid by CfW programmes have positive direct effects on the participants and indirect effects on non-participants through multiplier effects.
Our research confirms that CfW programmes in Jordan have positive effects at least in the short term on the level of income of both, CfW participants and non-participants.
The income of CfW participants' households increases, on average, by almost a quarter, as Roxin et al.'s (2020) quantitative approach was able to distil.Households participating in any of the three CfW programmes of GIZ used to have a monthly income of JOD 376 on average in 2017 when they had applied for participation and a total monthly income of about JOD 460 during their participation.Participation in these programmes has thus raised the income of participating households by JOD 85 on average, or 22 per cent, even though the typical wage of CfW participants in Jordan is 220 JOD.The difference is due to substitution effects: CfW participants give up their previous jobs to participate in CfW or they receive less transfers from public social assistance programmes or other households (Roxin et al. 2020, 60).The increase in household income is particularly high for households with a female CfW participant.As women have much better chances to get a job in CfW programmes than elsewhere on the labour market, the substitution of residual income is lower than for men.
After CfW employment, its positive net effect on household incomes ends.On average, household income then is even lower than before CfW participation, especially for households with an above average household income (Roxin et al. 2020, 64).Probably, former cash workers cannot just re-enter, without friction, the jobs given up for the sake of CfW employment.
The income of non-participants in host communities increases as wellmainly because CfW participants spend most of their income locally.30 per cent of the stakeholders we interviewed told us that they spend most of their income on food, which is normally purchased in local shops (e.g. in contrast to clothes).47 per cent spend most of their income on housing rents, which also contributes to local income levels because most landlords live in the same village.22 per cent spend their CfW income mostly on debts repay, mostly to local creditors (neighbours, landlords and local shopkeepers) (see Appendix Table 1).
Of course, some fraction of CfW wages flows out of local communities like the amounts spent on health care, education, water, electricity, clothes, and building materials for the improvement of dwellings.Yet, many interviewees reported that they bought large parts of even these items within the local communitiesnot least because most of the places that we visited are so remote that it would be too expensive and time-consuming to buy elsewhere.Very few Syrian refugees reported to remit CfW earnings home to Syria.In contrast, migrant workers, who mostly arrive without family members, send the better part of their income as remittances to their countries of origin and thus create no multiplier effect (Razzaz 2017).
The results of the GIZ post-employment census are similar.Its participants were asked about the three main ways they spent their CfW wage.The largest share ticked housing rent (50 per cent), followed by food (42 per cent), open bills (41 per cent), debt repay (33 per cent), health (24 per cent), household items (21 per cent) and transportation (17 per cent) (see Appendix Table 2).With the exception of health, transportation, and possibly household items, the bulk of these expenditures benefits other people in the community such as shop owners or landlords.Interestingly, spending patterns differ with nationality, marital status, educational achievement and location, so also multiplier effects differ (see Loewe and Schüring 2021).
These findings are corroborated by two other surveys: Syrian refugees in Jordan spend 69 per cent of their total income on housing and just 11 per cent each on health and food (Jones et al.2019); participants of ILO's Labour Intensive Infrastructure Programmes mentioned most often "daily consumption" (83 per cent) and debt repay (31 per cent) as one of the three main uses of their CfW wages (NAMA & ILO 2019).
Finally, the findings are similar to those on UNHCR's unconditional cash transfer programmes in Jordan (Hagen-Zanker, Ulrichs, and Holmes 2018), where, however, the transfer is much smaller.
Local shopkeepers confirmed these multiplier effects of CfW participants' wages.We asked shopkeepers close to the sites of CfW programmes whether their sales had risen after CfW programmes had been set up and whether some of their clients had higher incomes to spend.Over half of the shopkeepers confirmed, and almost a tenth said that their sales had increased even significantly because many of their customers got employed by a CfW programme.A shopkeeper in al-Qasr said that "not only on me, the whole market benefitted from the programmes.People come here and […] buy more goods than before."People from Kafr S  awm told us that a local chicken restaurant had increased its sales significantly because of the CfW programmes: "One place […] used to sell 50 chickens, now it can sell 70."Some shopkeepers said that even if CfW participants do not spend more money, they repay at least their debts, which also means that money is channelled into the local circuit.Not only shopkeepers but also local experts affirmed that CfW participants spend the largest share of their wages within their local communities.
However, the more indirect multiplier effect of CfW on income is just as temporary as the direct effect.A shopkeeper stated: "[I]n those three months [when the programme was running], I noticed an increase in sales.But when the programme ended, the increase also stopped." Another indirect effect of CfW programmes on income is due to local procurement.Some CfW projects "try to buy all materials and intermediate products locally, which in turn benefits the local economy" (project staff member).Likewise, most CfW programmes buy the food for the lunch meals for workers from local markets: "The local market of Kafr S  awm saw development because they are providing us with the food for the programme" (other project staff member).Some shopkeepers however complained that the CfW programmes had not bought their equipment at their or another local shop.Apparently, the intention to procure materials locally is often weighed against regulations to choose the best-priced offer.
Hypothesis 3: The wages paid by CfW programmes have positive effects on the investment behaviour of participants' and non-participants' households.
Our results confirm that CfW programmes can have investment effects.In the Jordanian context, they tend to be limited however, mainly because the CfW wages are low and the employments are short.

CJDS / LA REVUE
Several observers of local development stated that CfW programmes have a positive, though diffuse effect on investment.A project staff member remarked: "There is impact on the local economic market and there is more investment in communities."Similarly, a shopkeeper said: "I have heard that the people invest more because of the programmes.It has affected our markets."However, only some CfW participants can make savings.Most need their CfW wage to cover their daily expenses or to repay debts.Only seven CfW participants told us that they were planning to invest part of their CfW wages (see Appendix Table 1) through planting crops in the garden for sale on the local market (two); buying cattle (two); keeping bees for honey; opening a bakery; and setting up a household repair shop (one each).Many participants stressed that they would invest more money if only their CfW employment was a bit longer.A participant in Deir Allah stated: "I will invest the money that I have got from the programme but the duration of my employment is too short.So, I do not get enough money for the investment." The survey conducted in 2019 among participants of ILO's Labour Intensive Infrastructure Programmes also found that only 0.9 per cent of respondents reported investing part of their CfW wages in business, 0.5 per cent bought animals as an additional source of income and 1.4 per cent planned to save part of their wages (NAMA & ILO 2019).
Likewise, just 27 per cent of the respondents to the GIZ Post-employment Survey could make savings at all and, of these, 16 per cent saved just for personal items, that is, future consumption rather than investments, 12 per cent made provisions for future health care needs, 9 per cent for education, and only 3 per cent for small business investments (see Appendix Table 2).
Yet, the CfW programmes may promote investments also through multiplier effects.33 per cent of the interviewed CfW participants had used high shares of their CfW wage to repay debts to local shops, neighbours, relatives, friends or landlords (see Loewe and Zintl 2020: Table 11).Most of these debts had been informal credits, and their repayment, on top of stable or higher income from sales, allows creditors to increase their consumption or their investments.
Hypothesis 4: Work experience and skills acquired during participation in CfW leads to better employment opportunities after completionespecially for women.
Our research confirms that some CfW programmes enhance the employability of participants even in the long run but not the effective employment rate because there are hardly any vacancies on the Jordanian labour market.
Quite many CfW participants mentioned the importance of skills for employment.The acquisition of new skills was even a major motivation for participants to apply for CfW employment.According to the DEval study, 44 per cent of CfW participants feel better prepared for future jobs because of the skills they acquired during their employment (Roxin et al. 2020, 74-75).However, these statements need to be discussed in more detail as to the participants' most pressing needs and the actual nature of skills acquired.
CfW participants look mainly for income to cover their cost of living as their most immediate need.In Roxin et al.'s (2020: Figure 9) survey, 69 per cent of the respondents said that their primary need was higher income.Just 34 per cent mentioned employment, 9 per cent better accommodation, 8 per cent better food, 3 per cent education, 2 per cent access to health services and less than 1 per cent better security, political participation or other issues.CfW programmes hence meet exactly the needs of their target groups: In all three programmes investigated by the study, at least 65 per cent of the participants perceive the generation of income as the primary benefit of the programmes (ibid.).
In terms of skills provision, CfW projects in Jordan differ substantially.Only few have an explicit training component.Some help participants acquire some basic skills just because of the nature of the works these dofor example if they repair schools, health centres or street lights.A CfW participant said: "I learnt a lot.Now, I can deal with electricity, fix things in the house, paint and much more.Now I can do it alone without external help."Another stated: "I have already learned a lot about modern farming techniques."86 per cent of the respondents of the GIZ Post-employment Survey reported that they were (very) satisfied with the training they had received.Just 8 per cent said that they were (very) dissatisfied, the respective share being significantly higher among university graduates (21 per cent).
Still, many CfW participants consider soft skills more useful than technical skills: In the GIZ Post-employment Survey, 63 per cent considered "ability to work in teams" as one of the most helpful things they had learnt, 48 per cent named "commitment," 23 per cent "patience" and only 22 per cent "new technical skills," another 18 per cent "time management."The results of Roxin et al. (2020) are similar: 96 per cent of the respondent said that participation in a CfW programme had improved their selfesteem.91 per cent valued that CfW programmes had helped them structure their lives.91 per cent highlighted that participation had helped them establish additional social contacts.89 per cent valued the respect that CfW project staff had given to them.The aforementioned 44 per cent that appreciated mainly the skills learnt during CfW employment are rather few in comparison.
However, some CfW projects seem not to convey any tangible skills.Several interviewees said frankly that they had not learnt anything during CfW employment.Many argued that the short durations of employment contracts did not allow for sufficient training, while others highlighted that the kind of work done (e.g.waste collection and street cleaning) was too basic to allow for any kind of learning.
In addition, in a context such as the Jordanian, even the most appropriate training might still not be enough to help CfW participants get a job afterwards.The survey conducted among participants of ILO's Labour Intensive Infrastructure Programmes found that employment rates and average wages directly after CfW participation were worryingly low: before participating in CfW, 57 per cent of men and 43 per cent of women had a job, with an average daily wage of 7-14 JOD, while right after CfW employment only 32 per cent and 13 per cent, respectively, had a job, earning them on average 5-12 JOD per day.Out of these, just 10 per cent felt that their CfW employment had helped them find the new job.And yet even worse, in these few cases, CfW participation had been helpful mostly because of the personal networks that workers had built during their CfW employment rather than the skills acquired (NAMA & ILO 2019).
Likewise, according to Roxin et al. (2020), just 25 per cent of CfW participants have a job immediately after their CfW employment, while 40 per cent had had a job before.This drop in the employment rate might be because many CfW participants aspire a new job with similar conditions: a salary of 220 JD per month, limited working hours, health insurance, a written labour contract and a meal at midday.They accept more informal or less well paid jobs only once they have realised that they have no realistic chances to get the same working conditions on the regular labour market (ibid.).Hence, the employment rate among former CfW participants might go up again only several months after the end of their CfW employment, but we have no tangible information on such a trend.
In addition, the GIZ Post-employment Survey shows that relatively few former CfW participants look for future employment in the informal sector: 77 per cent of respondents said that they were looking for a new CfW job.38 per cent hoped for another job with a formal contract.20 per cent were looking for participation in a longer-term training programme.Just 4 per cent said they considered also an informal sector job (multiple answers were possible).
What is for sure, the situation on the Jordanian labour market continued to deteriorate during the period under research, which explains to some degree why it was so difficult for CfW participants to find a follow-up employment.
CfW may, however, open up possible employment opportunities by raising participants' motivation to work in sectors they had previously discarded.Jordan's so-called "shame culture," labelling certain manual labour as disreputable, is an important factor in job searches and even referred to in Jordan's National Employment Strategy 2011-2020 (MoPIC 2011).Several interviewees highlighted that CfW stretched these boundaries by providing reputable work with good working conditions in generally repudiated sectors.Mostly, the waste sector was mentioned, though CfW programmes can have similar effects for the agricultural or construction sectors.One participant said: "Large numbers of people [were] looking for jobs.They changed their perception about what is shameful."While economic pressure played some role, the change in perceptions was mainly due to CfW programmes' designa formal contract, favourable working conditions, skills trainings and well-targeted awareness campaigns: "In the beginning, there was this stigma about collecting waste.But now, […] the people love working here.[…] They [CfW programmes] made working with waste something that is not shameful" (CfW participant).
Changed perception also played a role for women's labour market access.Jordan's female labour market participation is among the lowest in the world; in 2016, it stood at 17 per cent for Jordanian and merely 4 per cent for Syrian women (Krafft, Assaad, and Keo 2018).
We found evidence that CfW helps female participants to overcome wide-spread societal reservations towards females working in public.A Syrian non-participant noticed that the CfW programmes "removed the shame culture on working women."Many interviewed women see CfW programmes as the only realistic entry point into the labour market: "[Female labour force participation] has become more acceptable.In the past, it was more shameful.The opportunities specifically for women arose" (female CfW participant).They described CfW projects as "safe work environments"much better than other workplaces because all workers are treated respectfully and labour rights are adhered to.Respondents of both nationalities stressed that higher acceptance of female employment was particularly important for Syrian women.A Jordanian woman said: It is very normal among Jordanians that girls work outside the house.Syrians, in contrast, think that women cannot work for pay and that they should stay at home.[…] When we sit together, I try to convince the other girls that it is normal for us to work.[…] I try to convince them that they have to be more educated and self-reliant.
In the GIZ Post-employment survey, the vast majority of female participants (76 per cent) had been unemployed before their CfW participation but their future career plans were remarkably similar to those of men, with a slightly higher share hoping for future participation in another CfW project.Women's reservations to being economically active had thus decreased.
In this way, CfW programmes seem to contribute even to a more general change in gender roles.Some interviewees stated that the existence of these programmes increased the overall recognition of women because they employ a considerable share of females and let them do similar work to men.A Syrian man stated that the local CfW project has "changed the attitudes [so] that the woman is an active part of the society […].The project has transformed the whole community."A Jordanian woman said: "We became equal with men.We are doing the same work.In the past, there were plenty of taboos about women going out and working.Now, this has changed.[…] Now we have an income.We can contribute to improve the household." And yet, economic necessity is sometimes the real reason for women to apply for a CfW employment.Some interviewees stated that female wage employment was culturally inappropriate but a last option for poor families to survive.Several female participants stressed that they would rather stay at home.Once the economic framework conditions improve, these women are likely to end their wage employment.While most women and men in our sample found female labour force participation acceptable or even positive, some (seven men, two women) insisted that this requires an acceptable nature of work and a safe work environment, and a few (three men, one woman) found females should not work in public.These results are in line with a study by Thorne (2021) on a joint skills development programme in Southern Jordan, where local and refugee women acquired important skills but could not overcome barriers connected to entrenched socio-cultural norms.The positive effect of CfW programmes on female employment might thus not be sustainable.

Conclusions and policy implications
We conclude that CfW programmes can foster local economic growth even in contexts of flight and migration.They satisfy the primary needs of participating migrants and vulnerable locals of both sexes by providing access to wage and employment and sometimes also upgrade their skills.In addition, they have positive effects on non-participating households in the host communities through multiplier effects, when CfW wages are spent and re-spent several times locally, and probably also the creation and maintenance of public goods such as streets, dams, municipal parks, school buildings and nature reserves.The effect on investment of additional income into new income-generating activities seems to be limited in Jordan so far but this might change if CfW participants are employed for a longer period.While some of these effects (especially the direct effect on income) are likely to be temporary, others are more sustainable (particularly the creation and maintenance of public goods, the investment of income in income-generating activities and the upgrading of workers' skills).
Other refugee-hosting countries might thus consider setting up CfW programmes as well, which should also cover both refugees and locals of both sexes in order to prevent negative effects on social cohesion.The international donor community should support host countries in such efforts.Many interviewed CfW participants stressed that they liked the fact that CfW programmes provide a wage for a work effort.According to them, unconditional cash transfers are like alms while wages are more decent.Also, the employment distracts the workers from their day-to-day worries, and guards against feelings of boredom, frustration or anger.It brings people together and protects them from feelings of loneliness, isolation and uselessness.
In addition, CfW programmes have an in-built self-targeting mechanism that works often better than the targeting of cash transfer schemes by means-tests (Sakketa and von Braun 2019): Only poor and vulnerable households apply for CfW programmes because better-off households are not willing to do the hard work that these programmes offer.CfW programmes, in contrast, cannot support work-disabled people and they are often more expensive than unconditional cash transfer schemes because of the costs caused by designing and planning the activities, buying building material and monitoring the work.
To reap their additional benefits, it is important, however, that CfW programmes do not just provide wage and employment but also foster economic growth through the creation and maintenance of public goods, multiplier and investment effects and the provision of relevant skills to workers.For this goal, project planners should (i) carefully decide what kind of infrastructure can effectively promote economic growth, (ii) employ cash workers for sufficiently long periods, (iii) try to enable CfW participants to learn something useful during their employment, and (iv) ensure that the work environment is adequate for women as well.

Notes
1.A recent European Journal for Development Research special issue discusses this link, see Burchi et al. (2022).2. The interview guidelines are published in full length in: Loewe, Zintl, et al. (2020).3. Five of the research sites are in the North of Jordan (Al-Mafraq, Irbid, Kafr Asad, Kafr S  awm, Umm al-Jimāl), where the majority of Syrian refugees live; three are in the centre of Jordan (Al-Azraq, Deyr 'Allā, Tal al-Rummān) and one is in the South (Faqū'a).4. For more detailed information, see: Loewe, Zintl, et al. (2020).5.The models and results of the estimates have been published in Loewe, Zintl, et al. (2020).