Using a Choice Experiment to Assess the Multiple Values of Land in Agricultural Uses in a Peri-urban Area: An Application to Edmonton, Canada

Over the last 30 years, the Alberta Capital Region (the City of Edmonton and 23 surrounding cities, towns, villages and municipal districts) has experienced rapid population growth, economic development, and conversion of agricultural land into alternative land uses. As a result, some of the province’s most productive farmland has been converted into residential and industrial development. Between 2000 and 2012, growth rates for population and developed areas in Alberta Capital Region are approximately 30% and 50%, respectively. For all newly added developed areas, almost 90% were converted from agricultural land (Haarsma, 2014). Concerns about the pace and pattern of development and conversion have thus led to the creation of the Capital Region Board in 2009 and the provincial Land Use Framework in 2008. Despite the historical rates of conversion and the policy attention it has prompted, little research has been conducted to examine what values are being lost as a result of agricultural land conversion. This research has thus been undertaken to assess the multiple values of land in agricultural uses in the Alberta Capital Region, Canada. Some values (e.g., the market value of agricultural commodities) accrue mainly to private individuals and firms, while others (e.g., biodiversity conservation values) accrue to society in general. Values of some agricultural uses, such as those associated with the production of “local food”, regulation of water and air quality, or maintenance of peri-urban green spaces, may be weighed very differently by different interest groups. Based on existing literature, we extend the valuation with an application of ecosystem goods and services that Millennium Ecosystem Assessment (2005) proposes. The objectives of the study are three-fold: (1) Estimate values that residents in the region place on conserving land in different agricultural uses; (2) Explore the links between those values and residents’ affinity with different ecosystem goods and services; and (3) Identify areas and strategies that are of particular interest to the public for conservation in agricultural uses. The study began with a series of focus groups. Three focus group discussions were held with selected experts to define the context, scope and objectives of the empirical study, and one focus group with a random selection of study area residents to pre-test the survey. The second part of the study involved an internet-based survey with a panel of Alberta Capital Region residents recruited by the survey firm, Qualtrics. The survey instrument includes background information on the respondents’ attitudes toward conservation, and an attribute-based choice experiment. This method defines non-market valuation such as the values of environmental goods or services in terms of various attributes including price, and then assesses the respondents’ Willingness To Pay (WTP) for specific bundles of attributes (Grafton et al. 2003). The choice experiments ask respondents to consider an alternative conservation strategy for land in a specific agricultural use, in a specific type of area, with a specific cost, as opposed to the status quo that would result in no policy change. The conceptual model is derived from the standard random utility specification in which utility is divided into observable and unobservable components (Hanemann 1984). In the model, utility contains a deterministic component that consists of the observable attributes (In our case, that is, type of agricultural use, acres conserved, adjacent area, location proximity, and one-time cost), and a random unobservable component. The empirical estimation starts with a simple Multinomial Logit Model. We also use a Multinominal Logit Model with interaction terms to evaluate the effects of individual characteristics such as gender, residence, shopping behaviors, and attitudes towards government policies. More advanced models, such as Latent Class Model and Random Parameter Model, are also estimated to provide further insight into heterogeneity. This research contributes to identifying agricultural regions of outstanding conservation values so that they can be protected against future land conversion. The results indicate that relative to land adjacent to primary highways, land adjacent to conservation buffers is generally preferred for conservation in agricultural uses. Additionally, residents place higher values on land within a 10-kilometer buffer to currently developed areas over land within city limits. Regarding agricultural uses, livestock grazing on native pasture has the highest values, with hay land ranking second. Values for vegetable farms vary from group to group, and residents who do not typically get food from farmers’ markets, community gardens or farms have the lowest WTP for vegetable farms. Women generally place considerably higher value on farmland conservation than men. Residence, whether the respondents are from Edmonton or surrounding counties, does not seem to make a difference on values of land in agricultural uses. Furthermore, concerns for local food production, water purification and air quality are the top reasons for conserving land in agricultural uses. Further research in this study will use the welfare measure from the WTP estimation in cost-benefit analyses to inform decisions on land use changes, including the creation, restoration and compensation of agricultural or natural areas. The non-market values of ecosystem goods and services associated with different agricultural uses can also be compared to the financial costs of such projects. Local governments have already expressed interest in those analyses.