Advertisement IC Markets

Cross-Sectional Momentum: Buying Last Month's Winners

Discuss 1-minute to 15-minute price action setups, fading intraday momentum, key support/resistance zones, and proven short-term trading methodologies.
Post Reply
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Cross-Sectional Momentum: Buying Last Month's Winners

Post by Fairman »

Cross-Sectional Momentum: Buying Last Month's Winners

Cross-sectional momentum compares currencies against each other. It buys those that have performed best recently, and sells those that have performed worst, expecting the ranking to persist for a while.
diagram.png
diagram.png (38.61 KiB) Viewed 255 times
The procedure is simple. Choose a group of major currencies, measure each one's performance against the dollar or against a basket over a lookback period, such as the last one to twelve months, rank them, and take long positions in the top group and short positions in the bottom group. Hold for a period, then rebalance.

Academic studies have documented momentum in currencies over horizons of several months, though the effect varies across periods and has weakened at times. Explanations include slow adjustment to information and the behavior of central banks, which move policy gradually.

The strategy produces periodic losses, particularly at turning points, when trends reverse. It also requires patience, with holding periods of weeks or months.

Practical considerations include transaction costs, which are small relative to the holding period, and swap, which can offset returns if you are long low-yielding currencies and short high-yielding ones.

Risk management is vital. Equal-risk weighting by volatility avoids dominance by the most volatile currencies. Position limits prevent excessive concentration.

You can test a basic version in a spreadsheet using monthly closing prices, which makes it accessible without special software.

Practical step: rank eight major currencies by their three-month performance against the dollar, and see which pairs the ranking would suggest.
It’s Fairman :geek:
Recommended broker for automated trading & scalping IC Markets
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Re: Cross-Sectional Momentum: Buying Last Month's Winners

Post by Fairman »

Monthly closes in a spreadsheet really is enough to test this. No special software, just rank and record the next month's result. It's a good weekend project. Take eight or nine major currencies against the dollar, get the month end closes for a few years, calculate the three month change for each, rank them, and then note what the top two and bottom two did the following month. After a couple of hours you'll have a rough idea of whether the effect shows up in recent years or not. Don't expect smooth results. Even when it works, there are long stretches where it doesn't, especially around big turning points.
It’s Fairman :geek:
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Re: Cross-Sectional Momentum: Buying Last Month's Winners

Post by Fairman »

Watch the swap side. Momentum can put you long a low yielder and short a high yielder, and paying carry every night on a trade held for weeks adds up. That's a real conflict, because carry and momentum sometimes point in opposite directions. If the strongest currency recently is also a low yielder, the momentum long costs you swap. Some people combine the two signals, only taking momentum trades where the carry is neutral or positive. That reduces the number of trades but removes the steady drain. Either way, include the swap in your test results. Ignoring it can make a momentum strategy look a lot better than it would have been.
It’s Fairman :geek:
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Cross-Sectional Momentum: Buying Last Month's Winners

Post by Shadow Trader »

Measuring eight currencies against the dollar leaves the dollar itself unranked, which matters in years when it is the strongest or weakest of the lot. Measuring all nine against an equal weighted basket of the others puts it back in.

Sample size is the other limit of the weekend version. Three years of month ends gives only 36 results, and if the monthly returns swing by about 2 percent, the standard error of the average is about 0.33 percent, so a mean below roughly 0.7 percent a month cannot be told apart from zero.

Going back further helps, though the effect has weakened in some periods, so a long sample mixes regimes. Splitting it into halves and checking the sign in each is a cheap compromise.

On the swap conflict above, the cost can be put in the same units. A pair costing 3 percent a year in carry drains 0.25 percent a month, which is a large share of a momentum edge that might only be a few tenths of a percent.

Does anyone skip the most recent month when ranking? In equities that gap is standard because of short term reversal, and it would be interesting to know whether currencies show the same thing.
FXS Prop Desk
Posts: 1693
Joined: Wed Oct 07, 2026 9:52 am

Re: Cross-Sectional Momentum: Buying Last Month's Winners

Post by FXS Prop Desk »

Momentum strategies in FX can suffer sudden reversals, especially when central banks change direction. Combining momentum with carry or value signals, as some academic work suggests, can reduce those reversal losses compared with momentum alone.

Testing the strategy with realistic retail costs, including swaps, is important, since academic results often assume institutional pricing.

For retail traders, rebalancing monthly rather than weekly keeps costs reasonable, since spreads on several pairs add up quickly.
FXS Official Team | The Idea Lab & Strategy
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Cross-Sectional Momentum: Buying Last Month's Winners

Post by Shadow Trader »

Rank churn near the cut off is a cost the spreadsheet version hides. A currency that sits second one month and third the next gets closed and reopened, paying the spread twice for a signal that barely changed.

A buffer rule handles most of it. Buy a currency when it enters the top two, but only sell it once it falls below fourth, and do the mirror image at the bottom.

Say the plain version changes two of its four legs in a typical month and the buffered one changes one. That halves the trading cost, and the test will show whether the slower exits give back more than they save.

Blending lookbacks is the other cheap improvement. Averaging each currency's rank over one, three and six months stops a single sharp month from flipping the whole basket.

It also softens the turning point problem a little, since the six month rank resists a sudden reversal. The trade off is that the basket is later to join a new trend.

Has anyone compared a single three month rank with a blended one over the same years? The difference in turnover alone would be interesting.
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Cross-Sectional Momentum: Buying Last Month's Winners

Post by Shadow Trader »

Textbooks say the London open and the US open are the volatile times. Real data for a specific pair on a specific platform is more detailed, and it often reveals surprises.

The method is simple. Export a few months of one-minute bars, calculate the average range for each minute or each five-minute block across the day, and plot the result. The chart typically shows clear spikes at known events, including data releases and session opens, plus lower activity at lunch periods and late in the day.

Beyond the obvious, it can show patterns that are unique to the pair. A spike at 8:30 New York that is bigger than at the London open. A secondary burst at the Tokyo fix. A quiet period at an hour that most traders assume is active.

With the profile in hand, setups can be tested within the hours that actually offer enough range to cover costs, which sharpens results without changing the strategy itself.

Overlapping several months of data on the same chart reveals which spikes are stable and which came from one-off events. A spike that appears every week at the same time is structural. One that appeared once may simply be a single large news day skewing the average.

A related refinement is to build separate profiles for each weekday, or at least to separate Friday from the rest. The profile for a Monday often looks quite unlike that of a Thursday, particularly around the first hours of the week and around data releases. With a profile for each, session filters can be much sharper.

Median values are more robust than averages, since one wild day can distort the mean. Plotting both lines shows how often the market surprises relative to its typical behaviour.

What does the intraday volatility profile look like for the pair you trade, and did anything surprise you?
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Cross-Sectional Momentum: Buying Last Month's Winners

Post by Shadow Trader »

Sometimes price jumps several pips in seconds with no obvious catalyst, often on a spike in tick volume. Scalpers who trade momentum try to jump on and ride it for a few pips.

The challenge is that the entry arrives after the burst has already started. Slippage and spread cut into the gain. If the burst is the whole move, the late entry has nothing left to earn.

A distinction that helps: bursts that break a clear level and hold often continue for a bit longer, while bursts into the middle of nowhere often fade. Level context tells something about which is which.

Execution speed matters more here than in almost any other style. A delay of a second can change the outcome, so hotkeys, a stable connection and a low-latency setup become part of the strategy.

Exit rules need to be fast too. Momentum trades should have a time stop, since a burst that doesn't continue within a few seconds is probably over.

Is momentum scalping part of the approach, and what is the typical holding time?
Post Reply