Search Results for « Time-series momentum (TSMOM) »

There are all kinds of quantitative portfolios used in the stock investing such as momentum, mean reversion, liquidity and so on. Cross-sectional momentum and Time-series momentum are the two main method of momentum strategies. They are also the basic methods that were used to construct the portfolio of investing. The essay analyses the difference of the theoretical definition and compares the two methods’ return, sharp ratio, maximum drawdown and some other indicators in a relative stable and prosperous environment which is simulated by well performed stocks in the 10 years’ period without considering the transaction cost. The result finds that the time-series momentum and cross-sectional momentum both generate positive return in the environment that is simulated. Moreover, the time-series momentum even performs better than cross-sectional momentum strategies in the stable and prosperous environment. The essay also points out the limitation of the experiment and presents some useful advice for the investors who want to invest in the stock market.

ABSTRACT Shiller's CAPE, dividend yield, and the government bond yield curve slope are measures of value in equity and bond markets. We find that equity market time‐series momentum performs well in mid‐valuation regimes, but breaks down near historical valuation extremes, where the direction of the equity market commonly turns. Controlling for equity market return reversals—when valuation measures are near their 10‐ or 20‐year historical extremes—increases the R 2 of a predictive regression of equity market returns by up to 110% and the R 2 of a predictive regression of time‐series momentum returns by up to 550%.

We develop and empirically test theoretical conditions for time-series momentum (TSM) and reversal across the capital structure. The conditions organize a unified pattern of predictability: equity TSM is concentrated in high-risk firms, low-risk firms reverse, corporate bonds reverse, and past equity returns predict future bond returns. The mechanism arises in a model where equity is a call option on firm assets and variance risk is priced. Variance risk generates momentum, leverage generates reversal, and the former dominates in sufficiently levered or high-variance firms. Aggregating these firm-level forces at the portfolio level explains and predicts momentum crashes.

Purpose – The purpose of this paper is to contain an empirical application of the concept of “time series momentum” – as developed by Moskowitz et al. (2012) – to commodity markets with daily data during 1995-2012. Design/methodology/approach – The paper applies the new concept of “time series momentum” to the sphere of commodity markets. Findings – The paper extends the results previously obtained by Moskowitz et al. (2012) to a second category labeled “breakout strategy.” Research limitations/implications – Further management strategies can be elaborated for investment management purposes, based on the suggested inclusion of the “time series momentum” in commodities. Practical implications – The empirical evidence gathered in this paper bears practical significance for portfolio managers and commodity tradings advisors relying on trend following strategies. Originality/value – Commodity markets are quickly developing to an alternative asset class for investors. Discovering their properties and characteristics has a broad appeal in finance.

This paper studies how time-series momentum (TSMOM) should be implemented when the traded assets belong to the same equity market. We compare conventional portfoliospecific Momentum with a common market-level Momentum signal, Long-Short with Long-Only positioning, and strategies with and without inverse-volatility scaling across a broad set of U.S. characteristic-sorted equity portfolios. The results show that a common market Momentum signal is systematically more effective than portfolio-specific Momentum, suggesting that these portfolios are primarily exposed to a shared market state rather than distinct portfolio-specific regimes. The preferred trading rule depends on the performance objective. Market-Momentum Long-Only delivers the strongest performance when evaluated by the Sharpe ratio, whereas Market-Momentum Long-Short performs best when evaluated by six-factor alpha. Conditional-state analysis explains this difference: nonpositive Market Momentum identifies a state with lower but still positive expected returns and substantially higher risk, while its factor-adjusted expected return is negative. Finally, volatility timing mainly improves weaker Long-Short strategies and adds little to the bestperforming specifications. Overall, effective TSMOM implementation within a common equity market requires reconsidering both the source of the Momentum signal and the appropriate response to unfavorable market states.

We study a fixed-lookback (L = 30 day) long/cash time-series-momentum (TSMOM) sleeve on liquid crypto-majors (BTC, ETH, BNB, XRP, ADA, SOL, DOGE), overlaid with a market-drawdown-state de-risking rule. The construction follows Moskowitz, Ooi and Pedersen (2012): hold each asset only while its trailing 30-day cumulative log-return is positive, otherwise rotate that allocation to cash; then scale the whole book's gross exposure down (×0.5) when the aggregate crypto market is more than 15% off its trailing peak. The headline finding is a drawdown-control result, not an alpha result. Across an offline 96-configuration grid, 100% of configurations produced a smaller maximum drawdown than equal-weight buy-and-hold; in an out-of-sample (OOS) walk-forward test with the lookback fixed a priori at L = 30, the sleeve delivered net-of-25bps Sharpe 1.41 (Harvey-Liu-Zhu t = 3.21) versus buy-and-hold Sharpe 0.70 (t = 1.60), with maximum drawdown cut roughly in half (-44.6% vs-84.4%). Adding the drawdown-state de-risk overlay (×0.5 beyond 15% off-peak) further reduced maximum drawdown to about-28% and lifted Calmar from roughly 1.56 to 1.98 while holding Sharpe near 1.53 (t = 3.69). The effect survives a frozen long-history (survivorshiphardened) universe and a 50bps round-trip cost assumption. We are explicit that this is risk-managed β-a documented trend premium with crash protection-not market-neutral alpha; that vol-targeting / Kelly sizing on top of this sleeve was tested and SETTLED NULL (it de-levers the payoff engine); that the L = 30 optimum is a recorded, accepted sharp-parameter fragility; that absolute returns are survivorship-inflated; and that the forward Harvey t > 3 promotion bar has NOT yet been cleared (the sleeve is paper-only and forward-accruing).