Forex spreads change with the trading day because the supply of executable quotes changes. When more dealers compete and orders can be matched easily, spreads tend to narrow. When participation thins or prices become harder to manage, they can widen.
That makes trading time part of execution cost. A setup that looks affordable during the London–New York overlap may carry a different price near the daily rollover. The currency pair, upcoming news and your broker’s pricing still matter; a session label is not a promise of cheap trading.
This guide focuses on those timing effects. For the underlying pricing mechanics, see our guide to forex spreads.
Why Forex Spreads Change Throughout the Day
A dealer quoting a currency pair must manage the risk of accepting a trade and then offsetting it. That is generally easier when other participants are actively quoting both sides of the market. More competing prices and more opportunities to trade can support a smaller difference between bid and ask.
During quieter periods, a dealer may have fewer attractive prices available for hedging. Quotes can become wider, available trade sizes can shrink, or both. The same order can therefore cost more without any dramatic movement in the exchange rate.
Liquidity and volatility are not opposites. A busy market can move quickly while maintaining tight spreads because buyers and sellers remain active. A sleepy market can show little price movement but still offer expensive execution.
Research on AUD/USD has documented narrower spreads during normally active periods and wider spreads around the thin transition after New York trading winds down. These are historical market observations, not promised retail prices: see the RBA analysis of intraday forex liquidity.
Forex Trading Session Hours: Use Reference Windows, Not Opening Bells
Spot forex has no single exchange opening bell. Sydney, Tokyo, London and New York are useful labels for regional activity, but trading does not switch between them like a relay baton. Singapore, Hong Kong and other centers also contribute.
The table uses illustrative observation windows for comparing spreads. It does not define official trading hours. The RBA study of intraday currency trading also treats session boundaries as indicative rather than formal market openings and closings.
| Market center | Local observation window | UTC during local standard time | UTC during local daylight time |
|---|---|---|---|
| Sydney | 08:00–17:00 | 22:00–07:00 | 21:00–06:00 |
| Tokyo | 09:00–18:00 | 00:00–09:00 | No seasonal clock change |
| London | 08:00–17:00 | 08:00–17:00 | 07:00–16:00 |
| New York | 08:00–17:00 | 13:00–22:00 | 12:00–21:00 |
Sydney’s UTC window starts on the preceding UTC calendar day and crosses midnight. Its daylight saving season also falls in a different part of the year from London’s and New York’s. Avoid treating “summer hours” as one shared global setting.
How the Main Sessions Affect Spreads
Asia-Pacific: Judge the Pair, Not Just the Session
The Asian session is not uniformly quiet. Yen, Australian dollar and New Zealand dollar pairs can receive locally relevant orders and economic news during these hours. Historical intraday research shows that their activity profiles differ from those of European currencies.
For a practical comparison, examine USD/JPY and EUR/GBP separately rather than assuming both should have their narrowest spreads at the same time. The relevant question is whether enough competing liquidity is available for the pair you actually trade.
Nor should the first quotes after the weekend or daily handover stand in for the whole Asian session. A short transition period and established regional trading are different conditions.
London and the New York Overlap
London hours and the London–New York overlap are sensible periods to investigate when comparing execution costs in major pairs. Institutional research identifies the London opening period and overlapping London–New York hours as periods of particularly strong currency-market liquidity. See the BIS analysis of forex trading locations and liquidity.
More participation can produce closer competing quotes, but the benefit is not automatic. A scheduled announcement, an abrupt repricing or one-sided order flow can interrupt otherwise favorable conditions.
Once European participation recedes, reassess rather than assuming the earlier spread will persist. For trading decisions, the quote available for your order matters more than the reputation of the session.
The London–New York Overlap Moves With Daylight Saving Time
Using the reference windows above, London and New York overlap from approximately 13:00 to 17:00 UTC when both cities use standard time. When both use daylight time, the corresponding window is approximately 12:00 to 16:00 UTC.
The awkward weeks occur because their clocks do not change together. US daylight saving time begins on the second Sunday in March and ends on the first Sunday in November. In 2026, those dates are March 8 and November 1, confirmed in the NIST daylight saving time schedule.
During the mismatch, New York is on daylight time while London is on standard time. Under the same reference windows, the overlap becomes approximately 12:00 to 17:00 UTC. A trading routine pinned to one UTC hour can therefore sample a different part of the session after a clock change.
The UK changes clocks on the last Sunday in March and October. Its 2026 dates are March 29 and October 25, listed in the official UK clock-change calendar.
Record UTC time and the relevant local time in a trading journal. Also check the platform’s server offset. A chart marked 15:00 does not tell you which financial centers are active unless you know what that timestamp means.
Why Spreads Can Widen Around the New York Rollover
The daily rollover deserves separate attention because it is a processing boundary, not simply another session opening. A common reference point is 17:00 New York time. For example, the CME Group EBS trading schedule lists a 17:00 Eastern Time trade-date roll for most spot FX instruments, with a different convention for NZD pairs.
That reference point corresponds to 22:00 UTC during New York standard time and 21:00 UTC during daylight time. Retail platforms can have their own maintenance periods and instrument schedules, so check the terms for the account you use.
During the thin handover, a wider quote can increase the cost of opening or closing a position even if the midpoint barely moves. Do not assume the spread will normalize after a fixed number of minutes; observe the actual quotes.
Keep spread widening separate from overnight financing. The spread is embedded in bid and ask. A financing debit or credit is a separate account adjustment. They may occur around the same time, but they are not the same charge.
When News and Holidays Override the Session Pattern
Session averages hide short periods of poor execution. Around a major announcement, a dealer faces the risk that a quote becomes outdated before it can be adjusted. That can encourage wider pricing or smaller quoted sizes, even during an otherwise active overlap.
Do not interpret “high activity” as “low cost” without checking the spread. A market can process many trades while prices jump and immediately available liquidity deteriorates. The distinction is covered further in our guide to spread widening during market volatility.
Holidays create another exception. A normal weekday timestamp does not guarantee normal participation. Research using interdealer EUR/USD and USD/JPY data found that UK holidays reduced trading volume even among dealers outside the UK. The effect extended beyond the location taking the day off: see the Federal Reserve study of forex trading activity and holidays.
A holiday is therefore a reason to check conditions, not a rule that every pair must trade badly. Treat weekend reopening, year-end trading and regional holidays as separate categories when reviewing your own spread records.
What a Session Difference Can Cost
Consider a hypothetical EUR/USD account quoting a spread of 0.8 pips during an active period and 2.4 pips during a thin period. These figures illustrate the calculation; they are not market averages or broker offers.
For a position of €100,000, one pip is $10. If the midpoint stays unchanged and a buy is immediately closed at the bid, the spread-related loss would be $8 at 0.8 pips or $24 at 2.4 pips. That is a $16 difference for the same position size.
This example excludes commissions, financing and slippage. With the same spread at entry and exit, a completed buy-and-sell trade carries one full spread’s equivalent cost relative to an unchanged midpoint, not two full spreads. Different entry and exit conditions require a more careful comparison.
The difference matters most when expected price moves are small or trades are frequent. Our guide to calculating forex spread costs in pips and money covers the position-size calculations.
How to Compare Spreads Across Sessions
A useful comparison starts with consistent observations from the same account and instrument. Comparing one broker’s advertised minimum with another account’s rollover quote tells you very little.
Collect observations across several ordinary trading weeks, separating announcement periods and holidays from routine conditions. Record the bid and ask together, not whichever price happens to appear on the chart.
- Use consistent timestamps. Record UTC, the session and any change in server offset.
- Keep the account conditions stable. Compare the same pair, pricing plan and intended order size.
- Measure typical and expensive periods. Review the median spread, an upper percentile and how long wider quotes persist.
- Check execution separately. Record commissions, requested prices and filled prices rather than treating the displayed spread as the complete cost.
Where possible, use regularly sampled quotes rather than a few screenshots. One observation can miss a brief spike; collecting every quote update without adjustment can give more weight to periods when updates arrive faster. Consistency matters more than a large but uneven pile of data.
Measure the hours in which your strategy actually produces orders. A low average spread at lunchtime is not much help if every entry and exit occurs elsewhere. Include exits, too: a trade opened during a liquid overlap may still need to close during a thinner period.
Fixed spread accounts also need care in this comparison. A steady displayed spread does not establish that execution conditions stayed unchanged. Read the pricing conditions and compare actual fills rather than judging the account from the spread column alone.
The practical rule is to use session timing as a starting filter, then confirm the live cost. Regional activity, clock changes, rollover and news all affect the context. None replaces checking the executable quote before placing an order.