1 Neighbourhood inequality as a health risk: empirical evidence from Swedish registers Preliminary version Authors: Sören Edvinsson, Centre for Population Studies, Umeå University Erling Häggström Lundevaller, Department of Statistics, Umeå University Nawi Ng, Department of Public Health and Clinical Medicine, Umeå University Gunnar Malmberg, Department of Geography and Economic History, Umeå University Introduction We live in a world where life expectancy is rapidly increasing, leading to an ageing population. There are however still large differences between different parts of the world as well as within countries. These differences represent a mix of spatial and social factors that are intertwined which makes it difficult to sort out the exact mechanisms and pathways that result in different health outcomes. Many individual traits, being risk factors for mortality or disease, have been identified. Demographic characteristics such as age, sex and marital status are obviously vital for understanding individual health risks, but research has documented the important role of individual social traits such as social status, wealth and income, social networks and education. During recent decades there is a growing interest in context variables as well. Health is not determined only by the pure individual characteristics but also the individual in relation to his or her context. This becomes apparent when considering the role of physical geography where environmental conditions such as climate, sanitary conditions, pollution, and population density have been identified as potential health risks from ancient times. There are however other context variables related to the social environment where we live that can be of importance. During the last decades the possible negative impact of economic inequality on health has been studied and analysed, first and foremost by Wilkinson and developed further by him and his colleague Pickett (Wilkinson 1996; Wilkinson and Pickett 2007, 2008, 2009). They argue that economic inequality leads to stress that in its turn causes bad health. Their studies have inspired into a lively academic debate where their results have been widely discussed and disputed. A large number of studies have been performed and we now know more about the issue, but there is however still no consensus among researchers in this matter (Deaton 2003; De Mayo 2012; Lynch 2004; Mackenbach 2002; Subramanian and Kawachi 2004; Wagstaff and van Doerslaer 2000; Wilkinson and Pickett 2007). The debate continues. To resolve the issue, good data and a clear idea about possible pathways is required. Some of the central questions concerns time and space. The time dimension is obviously very important in epidemiological studies. In order to identify causality, cross-sectional studies are problematic, in particular when it comes to disease and mortality. It takes days, months, years