You Can(’t) Always Get the Job You Want: Employment Preferences in the Peruvian Horticultural Export Chain

Abstract

Employment in high-value agro-export sectors has been recognised to entail the potential to contribute to poverty reduction in rural areas of developing countries. Concerns have yet been raised about the quality of the created employment and worker preferences have often been overlooked in the literature. We use a discrete choice experiment, in which we relate stated and revealed employment preference of agro-industry export workers in Peru. We explain employment (mis)matches as a function of personal and employer characteristics. Results suggest that employment preferences are heterogeneous, but that some groups of workers are systematically less likely to meet their ideal employment expectations. We formulate policy recommendations for both agro-industry employers to increase their workers’ job satisfaction, and for development agencies concerned with employment quality in high-value export sectors.

Acknowledgements

The authors gratefully acknowledge funding from the FWO – Research Foundation Flanders – and from KU Leuven, Department of Earth and Environmental Sciences. We are indebted to all the workers that dedicated time to our survey. We thank Isabel Lambrecht and Goedele Van den Broeck for valuable input during the design of the choice experiment. Data and codes will be provided on request to bona fide researchers.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. We use a choice experiment to analyse workers’ preferences. Older studies on employment choices, have used a ‘direct estimate methods’ by asking workers to rank attributes (see Bartol & Manhardt, 1979; Wiersma, 1990) or a ‘policy-capturing method’ by providing suitability judgements based on attribute description (see Einhorn, 1971; Zedeck, 1977). Both methods have been criticised for excessive subjectivity, participants fatigue, and inconsistency of results across studies (Slaughter, Richard, & Martin, 2006).

2. We have carried out previous research on private standards, employment conditions, worker empowerment, sourcing strategies and export performance in the Peruvian horticultural export sector. In light of this, we have implemented multiple rounds of interviews with companies, workers and stakeholders in the sector since 2010. The first author of this paper has written a PhD dissertation on various aspects of the horticultural export sector in Peru.

3. Attrition is 11 per cent and is due to migration of respondents and incorrect contact information. Mean comparison tests do not reveal differences in observable characteristics between re-surveyed and dropout respondents. Reasons of the 114 respondents to not enter employment are: no serious intention to start employment (27%), pregnancy (14%), under age 18 (16%), illness (19%), found other work (9%) or studying (15%). Reasons are largely exogenous to employment conditions and ex-ante employment preferences of these respondents (results not reported) are similar to those of the other respondents.

4. We follow a similar approach as Reynolds (2003). In a multivariate regression framework, they identify the socio-economic and job determinants for the gap between workers’ actual and ideal working hours. Our choice experiment set-up allows us to take into account more than one employment attribute. Individual, instead of average, WTPs account for personal characteristics in shaping work preferences.

5. Respondents were asked to rate how employers treat them on a scale from 1 to 5: if within company average values range between 1 and 2.49 the company was classified as ‘bad treatment’; between 2.5 and 3.49 as ‘fair treatment’; and above 3.5 as ‘good treatment’.

6. Individual WTP values by region are reported in Table A1 of the Appendix.

7. In Table A2 of the Appendix we report distances between WTP values in the preferred versus the actual employment, separately for each employment attribute. The positive values reveal that on average workers’ too high expectations are not fulfiled (i.e. respondents are willing to give up some salary to receive the attribute level of the preferred employment).