USD/JPY
Since 1996, USD/JPY has gained most in March (average +0.79%, higher in 60% of 30 years) and done worst in August (-0.94%). November has averaged +0.78%.
Grades describe the data. They are not buy or sell ratings.
Seasonality
Average return by calendar month, 1996 to 2026
Sell in May: November to April averaged +2.65%, May to October -1.06%; winter did better in 70% of 30 years. Compare markets
Month-by-month table
| Month | Average | Median | Up years | Last 10 years | Best | Worst | Years | Pattern |
|---|---|---|---|---|---|---|---|---|
| January | 0.00% | -0.70% | 47% | -0.74% | +5.89% | -4.84% | 30 | no clear pattern |
| February | +0.62% | +0.44% | 53% | +0.85% | +8.59% | -4.43% | 30 | no clear pattern |
| March | +0.79% | +0.80% | 60% | +0.58% | +7.53% | -6.56% | 30 | weak |
| April | +0.30% | -0.18% | 43% | +0.91% | +7.44% | -4.61% | 30 | no clear pattern |
| May | +0.08% | -0.02% | 50% | +0.01% | +4.68% | -8.14% | 30 | no clear pattern |
| June | +0.14% | +0.13% | 57% | +1.77% | +5.48% | -6.86% | 30 | no clear pattern |
| July | -0.42% | -0.59% | 43% | -1.15% | +4.93% | -6.94% | 30 | no clear pattern |
| August | -0.94% | -0.69% | 37% | -0.06% | +4.41% | -5.28% | 30 | strong |
| September | +0.33% | +0.08% | 57% | +1.17% | +5.27% | -4.52% | 30 | no clear pattern |
| October | -0.25% | -0.01% | 50% | +1.45% | +6.61% | -14.60% | 30 | no clear pattern |
| November | +0.78% | +0.24% | 57% | +0.01% | +7.90% | -6.13% | 30 | no clear pattern |
| December | +0.17% | -0.01% | 50% | -0.30% | +7.72% | -7.71% | 30 | no clear pattern |
Seasonal windows
Picked on older years, then checked on the last 10 years
Held in 6 of the last 10 years (2016 to 2025). Those years were not used to pick the window. Any window of this length went the same way in 56% of them.
We test every start date and every length from 10 to 90 days, so some windows look good by chance. To guard against that, we pick windows using the older years only and publish a window only if it also held in at least 6 of the last 10 years, which played no part in picking it. Each square is one year; green means the window worked. All open and upcoming windows
Technicals
50-day average below the 200-day
No active signal today. We watch 52-week highs and lows, RSI extremes and 50/200-day crossovers, and show what followed each in the past.
Reactions to events
Average move from the close before each event
| Event | Cases | Day 1 | Day 5 | Day 20 | Up after 20 days |
|---|---|---|---|---|---|
| Fed hikes | 45 | -0.08% | +0.02% | -0.02% | 51% |
| Fed cuts | 36 | -0.01% | +0.07% | +0.17% | 67% |
| First Fed hike of a cycle | 6 | +0.16% | +0.49% | +1.03% | 67% |
| First Fed cut of a cycle | 5 | +0.38% | -0.60% | -1.50% | 60% |
| Fed cuts of 50 bp or more | 18 | -0.41% | -0.33% | +0.80% | 78% |
| ECB hikes | 30 | -0.02% | +0.24% | +0.18% | 60% |
| ECB cuts | 30 | -0.41% | -0.33% | -0.29% | 43% |
| Tariff escalation headlines | 8 | -0.62% | -1.46% | -1.07% | 38% |
Year by year
Monthly moves for the last 16 years, most recent first
What drives USD/JPY
- The gap between US and Japanese yields; Bank of Japan policy shifts and intervention risk.
- Japanese fiscal year end in March brings repatriation flows.
Latest insights
Questions
What is the best month for USD/JPY?
Since 1996, March has had the highest average move for USD/JPY at +0.79%, rising in 60% of 30 years. March has the highest share of positive years (60%).
What is the worst month for USD/JPY?
August has had the weakest average move at -0.94%, rising in 37% of 30 years.
How does USD/JPY usually do in November?
USD/JPY has averaged +0.78% in November (median +0.24%), rising in 57% of 30 years. We rate the pattern as no clear pattern.
Questions, ideas or corrections?
Write to us. We read every email.
Not investment advice. Market Seasonality publishes historical statistics, news summaries and scenario analysis for information and education. Past patterns do not guarantee future results. Nothing here is a recommendation to buy, sell or hold any asset, and scenarios describe possible paths without predicting them. See the disclaimer and methodology.
