Day, B.H.; (2000) A recreational demand model of wildlife-viewing visits to the game reserves of the Kwazulu-Natal province of South Africa. (CSERGE Working Papers GEC 20). Centre for Social and Economic Research on the Global Environment (CSERGE): London, UK.

Preview

PDF 17590.pdf Download (773kB)

Abstract

In recent years, random utility models (RUMs) have become an increasingly popular approach to estimating the welfare benefits derived by visitors to recreational sites. Researchers using such models have tended to concentrate on the choice between sites; explaining a visitor’s decision by means of the different qualities of the available sites and the different costs of travelling to those sites. This is all well and good for ‘day trips’ but for recreational trips characterised by visits lasting a number of days, concentrating solely on the choice between sites may be a gross oversimplification. For such ‘away-breaks’, a visitor’s choice of accommodation and length of stay may be just as important as the qualities of the site and the length of the journey in determining the costs and benefits that result from the trip. This paper describes the application of a RUM known as a nested multinomial logit model (NMNL), which distinguishes the three dimensions of choice that characterise away-breaks; duration of stay, choice of recreational site and choice of accommodation type. Four costs are important in determining choice for such trips; the cost of travel to the recreational site, the cost of accommodation at the site, the cost of time whilst travelling and the cost of time whilst on-site. Previous applications have frequently assumed that travel time can be valued at some exogenously determined proportion of the wage rate, whilst at the same time ignoring the value of time spent on site. The specification of the indirect utility function in the model presented here, allows for the value of time to be inferred from the data by estimating the proportion of the wage rate that most appropriately values a unit of time spent in different activities. The model is applied to a unique dataset that records details of trips made by domestic households to the game reserves of the KwaZulu-Natal province of South Africa. These trips are typical of away-breaks, since visitors tend to travel fairly large distances to visit the reserves and typically stay one or more nights on site. Each of the game reserves affords visitors different wildlife-viewing opportunities and provides a variety of accommodation facilities that vary greatly in their quality and price. Geographical information system (GIS) techniques have been used to establish exact door to gate distances and provide accurate estimates of travel costs and travel times that take account of assumed road speeds. GIS techniques have also been employed to garner socio-economic data on the households in the dataset. One further novel feature of the data is the use of complex computer algorithms to accurately establish the choice sets faced by individual households. The three-level NMNL model is estimated using a full information maximum likelihood (FIML). The results of the analysis support the work of De Serpa (1971) in that they suggest that recreationists have a positive willingness to pay to save time in travelling to a reserve, but as would be expected, are not willing to pay anything to save time spent on site. The model is used to calculate welfare estimates for continued access to the different game reserves. Average per-trip estimates of the consumer surplus enjoyed by domestic visitors range from around $15 for one reserve, up to almost $50 for another. Boot-strapping techniques have been employed to calculate standard errors for these benefit estimates.

Download activity - last month

Export as CSVJSONXML

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item