Patterns of employment relationships: the association between compensation policy and contractual arrangements

Abstract

Firms respond differently to labour market regulations and develop an employment relationship accordingly. We use linked employer–employee data to examine the relationship between compensation policies and contractual arrangements in large-sized firms in Portugal. In this country, the wages are regulated through minimum wage and collective agreement, while employment is protected by stringent employment legislation. The empirical analysis starts with a fuzzy clustering to identify typical compensation policies. Three major segments emerge from this analysis: Competitive, Internal Labour Markets and Incentive. The first segment comprises low-wage firms, which are highly responsive to market conditions. The other two reveal properties of internal labour markets, although the incentive-based firms reinforce the use of discretionary power to differentiate the workforce. Subsequently, we estimate a regression model to examine how the compensation policy interacts with contractual arrangement. Empirical evidence confirms the segmentation predictions, i.e. low, flexible wages and flexible contracts prevail in the same firms. Furthermore, vulnerable categories like young workers and female workers are over-represented in Competitive firms, while high-wages are associated with incentive devices benefiting white-collar employees. Apparently, firms foster inequality among segments of workers and often penalise or favour the same category of workers.

Keywords:

• Compensation policy
• contractual arrangements
• labour market segmentation
• fuzzy clustering
• Tobit regression

Acknowledgement

The authors gratefully acknowledge the financial support from FLEX FCT grant PTDC/EGE-ECO/108547/2008, ‘Flexible wages for flexible contracts? The dynamics of the relationship between wage policy and employment contracts at the firm level’. This research was possible thanks to the kindness of the Office for Strategy and Studies (GEE), the Ministry of Economy and Employment for access to the data, Quadros de Pessoal. The authors particularly thank Marta Luís Pereira and Diana Alves for their support.

Notes

1. The data-set is property of Office for Strategy and Studies (Gabinete de Estratégia e Estudos – GEE), Ministry of Economy of Portugal.

2. Two dummy variables are created for manufacturing and KIFS that take the value one if the firm belongs to the respective industry, and zero otherwise (herein, dummy_manufactuting and dummy_kifs). Then, taking female as an example, we use two interaction variables: dummy_manufactuting* female and dummy_kifs* female. Finally, with these two additional variables we re-run the base regression in Table 4. A similar procedure is repeated for the variables blue-collar and young workers. The results are available upon request.