MomSignal: Momentum Trading Signal

Description Usage Arguments Details Value Author(s) References See Also

View source: R/MomSignal.R


Function to generate several momentum trading signals. Signals currently implemented are:

All the signals are as defined in Baltas-Kosowski (2012).

Also, to each signal can be associated a so called momentum speed, which is an activity to turnover-ratio used to assess signals trading intensity. Letting X the signal, its speed is defined as

SPEED_{X} = √{\frac{E[X^2]}{E[(Δ X)^2]}}

The higher the speed, the larger the signal activity and thus the portfolio turnover.


MomSignal(X, lookback, signal, cutoffs, speed = FALSE, ...)



A list of xts objects, storing assets data. See 'Details'.


A numeric, indicating the lookback period in the same frequency of X series.


A character, specifying the momentum signal. One of SIGN, MA, EEMD, TREND, or SMT.


A numeric vector, with positional cutoffs for Newey-West t-statitics and R^2, see 'Details'.


A boolean, whether or not to compute the chosen momentum signal speed.


Any other pass through parameter.


Data strictly needed in X depends on the signal chosen. SIGN is based on assets returns. MA, EEMD, TREND, and SMT are price-based momentum signals.

For the TREND, Newey-West t-statistics lower and upper cutoffs can be provided. With SMT, cutoffs can additionally provide the lower R^2 cut-off. Defaults are set at -2, 2 for Newey-West t-statistics and a minimum R^2 = 0.65.

SMT over sub-periods is not currently supported.


A list of xts objects, consisting of the chosen momentum signal for the corresponding assets data X provided. Signals are {-1, 0, 1} for short, inactive, and long positions, respectively. TREND and SMT are the only signals that can result in inactive positions. With speed, additionally the chosen momentum speed for the given assets.


Vito Lestingi


Baltas, A. N. and Kosowski, R. (2012). Improving time-series momentum strategies: The role of trading signals and volatility estimators. EDHEC-Risk Institute.

Bryhn, A. C and Dimberg, P. H. (2011). An operational definition of a statistically meaningful trend. PLoS One.

Luukko, P. JJ. and Helske, J. and Rasanen, E. (2016). Introducing libeemd: A program package for performing the ensemble empirical mode decomposition. Computational Statistics.

Moskowitz, T. J. and Ooi, Y. H. and Pedersen, L. H. (2012). Time series momentum. Journal of Financial Economics.

Wu, Z. and Huang, N. E. (2009). Ensemble empirical mode decomposition: a noise-assisted data analysis method. Advances in Adaptive Data Analysis.

See Also

sandwich::NeweyWest(), Rlibeemd::eemd()

JustinMShea/ExpectedReturns documentation built on Sept. 27, 2020, 5:41 p.m.