Australia's cash economy: Are the estimates credible?

The method of 'excess sensitivity' of Bajada (1999, 2000, 2001) indicates a large underground economy in Australia, with estimates of unrecorded income approximately 15 per cent of official gross domestic product. These estimates concern policy-makers, especially those agencies responsible for national accounts, tax collection, economic stabilisation and law enforcement. We show that the method exhibits a severe form of non-robustness, in which the results change markedly with a simple change in the units of measurement of the variables. There is a separate problem in which a key parameter is set to an unrealistic value that makes the estimates many times too high.