For an analysis of the effect of utilizing an alternative parametric framework on the estimates of multifactor productivity, see Baldwin, Gaudreault and Harchaoui (2001).

While it is possible that public capital provision affects the return on private capital investment, and may influence income shares, sufficient data are not currently available to test this supposition. It is not currently clear how business sector incomes should be distributed if the elasticities of public capital, private capital and labour should be divided if they exhibit constant returns to scale. While previous studies have used capital and labour share of income to split the return on public capital between business sector inputs, this approach raises capital and labour income, but it does not change their shares. Consequently, there does not appear to be a large penalty from maintaining currently held assumptions about business sector income shares.

For more information see Gu et al. 2003.

For more information see Baldwin and Gu 2007; Gellatly, Tanguay and Yan 2003; and Harchaoui and Tarkhani 2003.

Our public capital stock includes all investment in fixed assets by governments. All asset classes are employed for the sake of completeness. However, public fixed investment is dominated by investment in roads and bridges. For more information see Baldwin and Dixon 2007.

The mid-point estimates of 17% are remarkably consistent with estimates of the internal rate of return to business sector capital investment in Canada. Using the Input/output estimates of industry surplus and the capital stock estimates in KLEMS, Baldwin and Gu (2007) calculate the internal rate of return for capital in the business sector as 15%. In light of the uncertainty surrounding the estimate of public capital’s rate of return reported in Macdonald (2008), this suggests that the rates of return do not differ greatly between the business and public sectors.