This paper examines the effects of different specifications, functional forms and estimation methods on the results in a hedonic land price analysis for Vienna. We investigate purchase price transactions with undeveloped residential buildings plots and demolition objects in the period 1987 to 2003 by applying a series of parametric estimation methods, such as OLS, Box-Cox, cluster-fixed effects or methods of spatial econometrics. The results are assessed on the basis of appropriate statistical tests. As the data structure is of the pooled-cross-section type, we experiment with suitable spatio-temporal weight matrices applied in the spatial models. We compare the results with the standard forms nearest neighbor and Delaunay. Overall, we find some very high fluctuation of individual parameters. Especially for spatial location attributes and neighborhood characteristics which are asymmetrically distributed in space an adequate modeling of spatial interdependencies or lack of spatial information seems to be crucial.
Austrian house prices have risen rapidly in recent years and there has been considerable debate on the underlying factors. Much of the debate focuses on increasing trends in real house price indexes and the ratio of house prices to disposable income as a measure of affordability. While it is generally accepted that the low interest rate environment is a key driver of house prices, there is uncertainty as of the sustainability of the level of house prices in some European countries. The OeNB fundamental indicator for residential property prices, launched in January 2014, points to an overvaluation in property prices by 21.7% in Vienna in the first quarter of 2018. For Austria as a whole, the indicator reached 11.2%. Several international studies also point to overvaluation in Austrian housing markets. House prices can fluctuate more than fundamentally justified because agents overreact to current fundamentals as well as past returns and are influenced by their sentiment. Reliable valuation metrics are therefore very important for monitoring residential property markets. While different approaches to identify overvaluation have strengths and weaknesses, theoretical consistency should be a prerequisite of any model of house price behaviour. Furthermore, examining national price indices is an ineffective means of early detection of housing bubbles. Speculative overvaluations arise in individual local and regional markets before spreading to the national market. This paper analyses regional house price developments in Austria. We follow Bourassa et al. (2016) and use an asset pricing approach to compare actual price levels with implied fundamental or equilibrium levels. We model prices for existing condominiums as a function of the present value of expected market rents, allowing for a time-varying risk premium and state dependent rental growth expectations. Model results are then compared with results produced by alternative methods to assess house price developments. We also discuss the theoretical weaknesses of the OeNB fundamental indicator and the critical assumptions underlying our model.
Abstract. A variety of methods have been used to investigate the empirical relationship between research and development (R&D) spending and the productivity of firms. The most widely employed frameworks are the production function and the associated productivity framework. In these settings, productivity growth is related to expenditures on R&D, and an attempt is made to estimate statistically the part of productivity growth that can be attributed to R&D activities. This article surveys the expansive body of empirical literature on this subject and finds a large and significant impact of R&D on firm performance on average. However, the estimated returns vary considerably between the different studies due to differences across data samples and econometric models, as well as methodological and conceptual issues. A meta‐analysis on the studies surveyed reveals that the estimated rates of return do not significantly differ between countries, whereas the estimated elasticities do. Furthermore, the estimated elasticities are significantly higher in the 1980s and consistently higher in the 1990s compared with the 1970s. Hence, contrary to a widely held belief, we find no convincing evidence of an exhaustion of R&D opportunities in the last two decades.
This paper examines the effects of different specifications, functional forms and estimation methods on the results in a hedonic land price analysis for Vienna. We investigate purchase price transactions with undeveloped residential buildings plots and demolition objects in the period 1987 to 2003 by applying a series of parametric estimation methods, such as OLS, Box-Cox, cluster-fixed effects or methods of spatial econometrics. The results are assessed on the basis of appropriate statistical tests. As the data structure is of the pooled-cross-section type, we experiment with suitable spatio-temporal weight matrices applied in the spatial models. We compare the results with the standard forms nearest neighbor and Delaunay. Overall, we find some very high fluctuation of individual parameters. Especially for spatial location attributes and neighborhood characteristics which are asymmetrically distributed in space an adequate modeling of spatial interdependencies or lack of spatial information seems to be crucial.
