Is that factor just lucky? Australian evidence

Assessing the usefulness of daily and monthly asset‐pricing factors for Australian equities

“…It generates economically and statistically significant reductions in pricing errors across test assets, even after allowing for the possibility of data mining. Our findings for monthly returns, therefore, are highly consistent with Harvey and Liu (2020) and Hoang et al (2019). However, our daily analysis reveals that essentially the same assetpricing specification is also appropriate in research contexts involving shorter return horizons.…”

“…In addition, differences in methodological approaches to factor construction are inevitable and further hinder comparisons from one paper to the next. Nevertheless, our mean monthly returns are not dissimilar to those reported in Hoang et al (2019) for PS, GP, ROE HXZ , NOA, MAX, IVOL and ROA. Of the 27 factors studied in this paper, 16 factors have mean monthly returns that are statistically different from zero at the 5 percent level of significance.…”

“…In a recent paper, Hoang et al (2019) provide the first out-of-sample testing of Harvey and Liu's (2020) finding. Moreover, they utilise this novel approach to assess the usefulness of asset-pricing factors in the Australian equity market.…”

“…Moreover, they utilise this novel approach to assess the usefulness of asset-pricing factors in the Australian equity market. In what is by far the most comprehensive study of potential factors to date, Hoang et al (2019) construct 91 factors that have been identified as anomalous in prior international work. Applying Harvey and Liu's (2020) methodology, 3 Other issues with the GRS test include: (i) the distributional assumptions on which it is based, (ii) the lack of precision with which the variance-covariance matrix of regression errors can be estimated, (iii) the difficulty in implementing the GRS test when the number of test assets is large, and (iv) implausibly large short positions implied for some assets in the mean-variance efficient portfolio (see Harvey and Liu, 2020, p. 4).…”

“…Our paper is closely related to Hoang et al (2019) and builds on their work in a number of ways. First, we consider which factors are useful when data are sampled at daily frequency.…”