ABSTRACT
Friedman’s (1966) first formulation for optimal bidding has been criticised as demanding unrealistic amounts of data to estimate the model parameters. Hanssmann and Rivett (1959) partially solve this by reducing the number of parameters in the model and thus the data demands, but with loss of predictive power. Multivariate methods offer a means of better utilisation of all available data, depending on the adequacy of certain assumptions concerning the statistical properties of bids. Recent empirical studies (Skitmore, 1991) indicate that, with suitably transformed data, these assumptions may not be unduly violated in construction contract auctions. This paper considers the use of one such multivariate approach for deriving optimal mark up values against both single and multiple competitors.
