Here's a question that sounds academic until you realise real money is riding on the answer: does momentum on the ASX actually work, and if so, which flavour of it? Cross-sectional versus time-series momentum is one of those distinctions that separates quant traders who've read the literature from those who've stress-tested it on thinly traded Australian equities with gaps, halts, and survivorship landmines everywhere.
The honest answer is that both strategies have empirical support — but the statistical significance of each depends heavily on how you handle the data constraints unique to the ASX. Australia's relatively small liquid universe, high concentration in financials and resources, and shorter historical depth compared to US datasets make replication genuinely hard. What looks like alpha in a backtest often dissolves under proper testing once you account for those realities.
Cross-sectional momentum — the Jegadeesh and Titman 1993 framework — ranks assets by prior 12-month returns, goes long winners, shorts losers, and rebalances monthly. On the US market with thousands of stocks, the law of large numbers does a lot of the statistical heavy lifting. On the ASX, you might have 150 genuinely liquid names. That's not a portfolio; that's a dinner party. Your t-statistics on mean returns need to be treated with real scepticism when cross-sectional correlation is high and your effective sample size is tiny.
Time-series momentum — Moskowitz, Ooi, and Pedersen's 2012 contribution — asks a different question: is this asset's return positive over the past 12 months? If yes, go long; if negative, go short or flat. This approach sidesteps the ranking problem entirely, which is genuinely useful on a small exchange. Each stock becomes its own signal, and you can test significance stock-by-stock using standard t-tests on the time series of returns. The catch? You need enough history per stock, and ASX data before the mid-1990s is patchy. That constrains your effective degrees of freedom badly. Proper significance testing here means using Newey-West standard errors to correct for autocorrelation — skipping that step is how researchers fool themselves. For a rigorous grounding in the statistical mechanics, the full treatment in Investopedia's momentum explainer and the original Wikipedia article on time-series momentum are solid starting points, while the theoretical underpinnings of cross-sectional ranking live in cross-sectional momentum theory — read them alongside the original papers, not instead of them.
The practical takeaway: before arguing about which strategy is better on the ASX, fix your data first. Source point-in-time constituent lists, include delisted stocks, and run a Newey-West corrected t-test on your strategy's return series. If your t-stat is under 2.0 out-of-sample, you don't have a strategy — you have a story.
Statistical significance isn't the finish line; it's just proof you're not fooling yourself with noise dressed up as edge.
This content is for educational purposes only and does not constitute financial product advice. Past performance is not indicative of future results. Profit Logic Ltd (ACN 688 669 936) accepts no responsibility for errors or omissions in this content or anywhere on this website. Always seek advice from a licensed financial adviser before making investment decisions.