No momentum factor after accounting for cross-sectional liquidity

What causes momentum, or what is associated with momentum? A new paper suggests that there is a strong link between liquidity and momentum, and that changes in liquidity precede momentum gains.

The paper, titled “Momentum Returns and the Role of Liquidity Improvements” by Jeppe Bro, demonstrates that the traditional stock market momentum anomaly is actually driven by cross-sectional liquidity dynamics rather than representing an independent risk premium. 

The novel idea is that past winners systematically see their trading liquidity improve before a portfolio is formed, while losers see liquidity deterioration. The momentum price drift is the market adjusting to these new liquidity states. The author calls this the Liquidity Improvement Factor, and when including this factor, there is no momentum alpha. The data shows mixed results pre-2000 data relative to more recent data. Any momentum effect is drift toward high-liquidity stocks and is a byproduct of liquidity dynamics. This is the most recent paper that attempts to explain the momentum factor. 

This is a very interesting thesis. Investors should track or follow liquidity changes to enhance any measure of momentum. I have some issues with the Amihud measure of liquidity, which looks at absolute return divided by vol and is manipulated to form liquidity differences. Still, I do not have a better alternative at this time. 

Expectation Bias and short-term Momentum

 

Machine learning can be used to predict analyst forecast errors. These forecast errors can predict cross-sectional returns and abnormal returns. There is an underreaction to fundamental news, which is amplified by overconfidence, sticky beliefs, and information uncertainty. This expectation bias can explain short-term price momentum in high volatility stocks. From expectation bias comes a rationale for trading patterns in price. See "Expectation Bias and Short-term Momentum". Past forecast errors have a critical impact relative to other features.

The problem of bimodality and deep mometum

 

Momentum is considered one of the financial factors that shows consistency throughout time. Other factors come and go, but not so with momentum. This does not mean that momentum will work at all times. Additionally, momentum is subject to significant risks. It depends on the regime, with momentum profits greater during an expansion than during a recession. There is also a greater likelihood of crash risk with momentum trades. This is the premium investors are paid to hold assets showing momentum. Finally, it is found that momentum exhibits a bimodal distribution of relative returns. Past winners may be likely to persist, but there is also a greater likelihood of becoming losers. This U-shaped distribution creates specific risks when holding momentum stocks.

The paper, "Bimodality Everywhere: International Evidence of Deep Momentum,” explores this specific momentum feature by using a deep momentum technique (RET) to mitigate or exploit it. Deep momentum is a two-step process: first, neural networks generate probabilities of future return declines; second, stock performance is based on returns or the Sharpe ratio to form long-short portfolios. It is shown that using specific machine learning techniques yields significant improvements over traditional methods for forming momentum portfolios. 

There is a lot going on with this paper, so the devil is in the details, but it shows that machine learning can be used to improve over a simple momentum strategy.

Nonlinear momentum - Increase positions with signal strength

 

Many firms are using trend-following. Their distinctions are often based on the length of trend signals, the set of markets used, and the risk management techniques employed. A recent paper, Nonlinear Time Series Momentum, measures the nonlinear relationship between trend and risk-adjusted returns using machine learning techniques. Using techniques to exploit nonlinear relationships within momentum outperforms simple linear methods. This nonlinear value-added is observed across all asset classes, frequencies, and horizons and lookback periods. This is especially true during market downturns. 

For modelers, this research concludes that simple nonlinear transformation of momentum signals will improve strategy performance, and signal strength interacts with predictability. However, as signals move to extremes, the extra risk from adding to position sizing diminishes, and at some point is not worth taking the extra risk. Hence, there is a complex nonlinear function between signal strength, sizing, and optimal risk-adjusted returns. Increase position exposure when signals are stronger, but reduce the signal exposure as you move to extremes.  

Momentum is persistent - A history

The paper, "Momentum factor investing: Evidence and Evolution", provides a history of research on the momentum factor. It has expanded into many research areas, and all have shown strong momentum effects. Momentum is across all asset classes, all time periods, and all regions of the world, and is shown to come in many forms, both price-based and fundamentals-based. When considering these different approaches to momentum, it is clear that many are unique and not associated with the same fundamental drivers. Nevertheless, we can agree that momentum is closely tied to behavior through the slow reaction to news.

 

Momentum across different factors

 

Momentum can be displayed across different forms, and it seems as though each has different forms of persistence: 

• beta
• country 
• factor/style 
• industry 
• stock 

The factor and industry momentum are strongly persistent over the short and intermediate run. Stock-specific momentum is persistent over the intermediate but not the short run. Beta and country momentum do not show persistence. This has important implications for how we use the momentum factor. These issues are described in the paper, "Optimizing the persistence of price momentum: which trends are your friends." The critical takeaway from this work is that while momentum is pervasive, there are areas or pockets of momentum persistence that can be exploited. These points of persistence are extreme with factor and industry tilts. Industry behavior moves together as well as factor behavior. This is not found in the overall market or with individual stocks.

Is it always about momentum but need style/factor diversification

 

There is a bandwagon effect with the momentum factor. The momentum factor performance is closely associated with cumulative flows into an asset or asset class. Still, these strategies do not last forever, which is why an investor should diversify their sources of alpha.

The simple case is to match the fundamental and systematic components to achieve a smoother blended return. 

Money flows and market behavior will change, and it is hard to find these turning points, so the first pass for effective factor management is to use more than one factor and ensure they show natural low correlation.

Return stacking - an easy approach for return enhancements

 

Return stacking is not hedging. Return stacking is not a solution to higher returns. Return stacking is a means of efficiently using capital through mixing core assets with the benefits from futures margin structures. It is a variation on portable alpha strategies, or portable beta by another name. An investor gets capital efficiency through gaining exposure in futures for a core asset. The unlocked capital can then be used to buy another asset. See "Return Stacking &Portable Alpha: An Investor's Guide".

The core approach can allow capital to be used to create a better return-to-risk trade-off efficiently. Of course, the choice of the stack will determine the return-to-risk. The foundation of this work is based on choosing assets in the stack that are uncorrelated.

The power of combining value and momentum factors

 

One of the more interesting combinations of factor exposures is value and momentum. Value and momentum have both delivered strong excess returns, and blending them together also has strong benefits, even though the value and momentum strategies are negatively correlated. 

An interesting new paper looks at the combination by trying to explain the excess returns through a pricing kernel that uses nine latent variables from the combined value and momentum cross-section, see "Value and Momentum Leftovers". The combination has proven to be much better than investing in each factor by itself. However, there is an issue of what the pricing kernel should be to help combine these two factors because there is a significant alpha after trying to price the combination. Additionally, the idea that you can just blend the value and momentum as a 50/50 combination may not be effective. 

We will not go into the details of this "leftover" argument other than to say that the use of latent variables can help price these two key factors and can lead to better factor combination outcomes. The blend of value and momentum can be improved through well-defined pricing techniques. 

Trend market performance not all the same

PremiaLab Research has issued a new paper titled "Navigating Market Shifts with QIS Trend and Momentum Strategies," which analyzes trend and momentum behavior in 2025 and the last five years using their pure factor models. First, trend-following has not performed well across all major asset classes; however, an equity cross-sectional momentum approach did very well. The directional changes associated with the high uncertainty in 20205 have hurt those following trends. Trend-following is a pure directional strategy, while a momentum strategy focuses on relative movement to create a long/short portfolio.

Over the longer term, equity momentum continues to perform well, but there were also substantial gains focused on short-term interest rates. Clearly, there are no exploitable trends in all assets, and the secret to good trend-following is to balance exposures across all asset classes. 

 

Momentum and reversal related to turnover

 

Momentum is a key factor found across all asset classes. It is persistent and perhaps the most used of the major risk factors. A paper I had not read before provides more light on the topic by showing that you can have both reversal and momentum with 1-month returns if you sort on turnover. The low turnover decile shows short-term reversals, while the high turnover decile shows momentum. See "Short-term Momentum".

Reversal and momentum can coexist once you account for turnover. Large liquid stocks are less sensitive to price pressure effects. The price pressure effect is the main driver for short-term price moves and reversals, but turnover does matter. Don't fight short-term reversal unless you account for the turnover issue.

Fama-French factors and economic drivers - Watch macro

 

The Fama-French factors have been used extensively to describe any set of returns; however, further work is needed to explain the economic drivers behind these factors. The paper "Understanding Asset Pricing Factors" takes a novel approach to analyzing the connection between economic events and factor moves. The authors analyze days with significant factor returns and then classify them by linking them to new articles the following day. Macroeconomic news, monetary policy, and corporate earnings reports are the main drivers of returns for factors, as should be expected. 

The Fama-French factors include four major factors beyond market risk: SMB (small minus big size effect), HML (high minus low value effect), RMW (robust minus weak profitability effect), and CMA (conservative minus aggressive investment effect). It is found that HML value premium is related to the macro factor. CMA is related to the commodity factor. SMB is correlated with the exchange rate factor and the unknown factor. RMW is related to shocks in individual companies. The authors use AI and human coders to classify news events that were tied to the FF risk factors.  

The categorization shows that humans and AI do a similar job and there is a clear distinction between the events that drive factor returns. Unsurprisingly, macro is the dominant driver of large moves. This resurrects the issue that even equity investors who may focus on factors RV should follow what is happening in the macroeconomy.

Factor investing - The 3 P's framework

 

How many factors should you have in an equity?  The factor zoo is constantly expanding, so you must be continually on the lookout for the next new factor, yet there should be a limit on the number. So, how do you determine the correct number of factors? Felix Goltz suggested that the number should be based on the three Ps - Pervasive, Plausible, and Practical. This is simple, easy to remember, and can be implemented by any investor. Of course, there is much room for interpretation concerning these three P's, but that is a deeper discussion.

Pervasive - A factor should stand the test of time, apply across different asset classes, and be verified through independent research.

Plausible - A factor should have a strong theoretical foundation and a straightforward narrative that can explain its presence. 

Practical - A factor should still be present despite transaction costs and should not deteriorate as more investors use it. 

If you have only 3 factors, you will likely be too stringent with your three P's. If you are getting to a number greater than a dozen, it is likely you should review your criteria. Perhaps these numbers are not wide enough, but this is a good start for forming a process for testing and inclusion. 

Portable alpha - A useful tool, but questions

 

We have always been in favor of portable alpha as the best way to gain both beta and alpha exposure. Stop thinking about picking managers and pick the amount of beta and potential alpha. Don't pay a premium for beta. Pay for your alpha only. The following graphs are from Blackrock and Pimco and provide the basic framework.

The idea that you can engineer extra value always has appeal for investors, yet there is a need to look at the risk portion of the equation and the fact that alpha generation must meet a minimum to offset the costs of the engineering. 

One, the risk is based on the amount of leverage used. Do you want to have 100% equity exposure plus alpha or do you want to undertake this strategy without a higher notional value? The correlation for the alpha strategy with the market matters. Second, given there are costs with the strategy, there must be an expectation that alpha will be both positive and above the financing costs. There is also a need for the alpha to be uncorrelated with the market return so that any beta shortfall will be offset by alpha and lower risk.  

More talk about portable alpha - Is this right time?

 

AQR came out with a new research piece called Portable Alpha: Why Now? The Problems: Elevated Equity Valuations and Macro Volatility. Their argument is that given the high valuations, investors should consider portable alpha to gain the extra exposure of hedge funds. Keep your 60/40 stock bond exposure and add in the hedge fun exposure on top of the core. This is a sound argument, but it begs the question of why hold the classic 60/40 mix. 

If you think equities are overvalued, you can change the asset allocation mix and add the hedge fund exposure. For example, you could increase the bond exposure through futures and then add the alternative exposure as a return kicker for added value. 

This can be the right time for portable alpha, but it does not have to be based on holding equity exposure constant.

The portable alpha approach can be used numerous ways as a method for both return enhancement and risk reduction. The process of adding alpha to core exposure can come in many forms. 

Trend-following and equity markets - control the costs

 

There is money to be made trading long-only trends in equities but like all trading it is not easy and driven significantly by cost assumptions. In fact, it is the cost control that may be the most important alpha producer. Take what may seem like a good model without transaction costs and then add realistic slippage and trading costs and you will see significant alpha deterioration. Redo the analysis but then add a set of rules that account for turnover and costs to reduce the number of trades, and you can add back alpha. Of course, you will never get back to the original theoretical performance without any trading costs, but the number may still look attractive. There is no free lunch and there is no hidden lunch. Perhaps the easiest place to find alpha is through optimizing for costs. 

While cost analysis was not the main goal of the paper, "Does Trend Following Still Work on Stocks" it may be the key takeaway. The model is simply based on looking for new highs for stocks with an ATR stop exit strategy. There are no special features, but it does work until we add reasonable cost assumptions. Nevertheless, simple rule changes can get the strategy back on the right track through minimizing turn-over. Along with entry and exit signals, place turnover constraints into the model.

Factor and price momentum across the globe

 

Many have tried to to explain away price momentum as driven by factor effects, but a recent study that looks at a large set of anomalies across a large set of countries fins that price momentum is still a key factor, see "Factor momentum versus price momentum: insights from international markets" The question answered in this paper was simple, does factor momentum drive stock price momentum? 

Now, factor momentum is strong across most international markets, but factor momentum cannot entirely explain stock and industry momentum; however, factor momentum based on principal components does a better job than traditional factor modeling. Nevertheless, price momentum often does a better job of explaining factor momentum. So, we can conclude that price momentum is distinct and cannot just be a combination of factor momentums. The paper provides an interesting display of comparisons between price and factor momentum that makes the story more compelling.

30 years of momentum research and it is only getting stronger

 

It has been over 30 years since the first major work on momentum was presented in a leading finance journal (Jegadeesh and Titman). Prior to that study, the conventional wisdom of efficient markets believed there was no trend or momentum effect. The world has changed and over the last three decades we have extensive research that has improve the initial research and strong explanations for why momentum exists. There are both behavioral and risk-based explanation for momentum and they have stood the test of time. Markets will under and overreact to new and will have different time dimensions. There is also a systematic risk component that can provide a reasonable explanation for momentum.

Momentum does not just exist in equity markets but is present around the global, in commodities, currencies, and fixed income. There is also commonality in stock momentum through industries and factors. Momentum is one of the strongest, most pervasive, and well researched factors in the marketplace. 

Of course, the strong long-term research results do not mean that momentum will always work as has been tested, momentum works strongest when there is a strong overall market sentiment. When more investors expect momentum is a given, it is more likely to have a poor performance year. There is a range of performance. 

All if this work is nicely presented in a review, see "Momentum: what do we know 30 years after Jegadeesh and Titman's seminal paper?" 

The out of control momemtum factor - can it continue?

 

The exceptional factor for 2024 is momentum. There are many ways to calculate this factor, but a simple equity benchmark is from S&P which has a broad suite of equity factors. The momentum factor is up 48% through the first eleven months of the year and is more than 50% higher than any other year and any other factor. Strong momentum factor performance does not mean a reversal in the next year, but the size of this move is extraordinary. Of course, the driver a just a few stocks which means that is not likely that this can continue if these key stocks have any slowdown or reversal. The high momentum winners may see a reversal after the one-year mark, but the follow-through on these names which have lasted for more than a year suggests that it is not a given that there will be a January effect. 

Realize that factor extremes can happen but does not suggest that there will be continuity. The rebalancing of names can support factor returns but looking at distribution properties leads to caution in holding the momentum factor.

Factor investing across different regimes

Factor risk premiums are time varying and a simple approach of breaking the economy into a four quadrant macro regime world will have significant benefit. The mayor regime is based on rising or falling inflation and the composite leading indicators for the US. Both are easily obtained, and the four quadrants can be easily generated using monthly information.  Based on the factor premium indices from S&P Global Intelligence, we can identify changing factor return profiles. For the full story, see "A Historical Perspective of Factor Index Performance Across Macroeconomic Cycles". 

Specific regimes may last for long time periods only to see significant uncertainty as regime shifts come frequently. These factor returns will vary significantly. Clearly falling growth and rising inflation is worst environment followed by falling growth and falling inflation. The best environment is when growth is rising and falling inflation. 

The quality factor index shows the best returns overall returns followed by the low volatility strategy. While outperformance and hit ratios are clear during different regimes, the gains may seem small relative to transaction costs, yet a simple strategy of regular rebalancing will lead to significant gains over the longer run.