New ways of valuing ecosystem services: big data, machine learning, and the value of urban green spaces

There is considerable policy interest to integrate the value of ecosystem services into systems of national accounts. Urban green spaces are traditionally valued by merging spatial household data with administrative data on land use to obtain their amount around households; this is then related to residential wellbeing or real estate prices to arrive at a monetary valuation. This traditional approach, however, neglects not only the large heterogeneity in the quality of urban green spaces but also issues of measuring outcomes and issues of reverse causality. We discuss new data and methods to overcome some of these issues. We focus on the use of high-frequency experience-sampling methods on wellbeing or web-scraped real estate prices to better understand impacts; big data from crowdsourced imagery of urban green spaces or satellite imagery of chlorophyll activity to better understand quality; and machine learning to make better use of data. Finally, we discuss the potential of field experiments and quasi-experiments to deal with reverse causality. Together, these approaches can greatly complement our traditional toolkit for valuing ecosystem services.