International Fund NAV: Fair Value Pricing

An international fund's NAV looks stale for a simple clock reason: the stocks inside it stopped trading hours before the fund put a price on them. Tokyo's regular session ends before 3:00 a.m. New York time and London's before noon, yet the fund strikes one net asset value at 4:00 p.m. ET. Fair value pricing is the board-approved process that carries those stale foreign closes forward to the 4:00 p.m. strike, using information that arrived after each home market shut.

Why a 4:00 p.m. NAV rests on prices set overnight

A mutual fund prices once a day. It totals the market value of what it holds, subtracts what it owes, and divides by shares outstanding; that arithmetic is walked through in how a mutual fund NAV is calculated. Every order that beats the cutoff receives that single price, whatever time it was entered, a rule covered in when mutual funds trade.

For a fund holding US stocks, the closing prices feeding that calculation are minutes old. For a fund holding Japanese stocks, the last regular-session print in Tokyo is more than thirteen hours old by the time it is used. European holdings are roughly four and a half hours stale, since London and Frankfurt both finish around 11:30 a.m. New York time. The world stock market hours guide maps those overlaps.

How much price information arrives after the foreign close

A company that trades in Tokyo and also carries a US listing gives a clean measure of what a stale close misses. Its US line keeps trading through the New York day, long after Tokyo has gone dark, so the move from the 9:30 a.m. open to the 4:00 p.m. close is information no Tokyo closing price can contain. The panel splits three years of daily moves for 6 such names into two pieces: the overnight gap, prior close to next open, which wraps the home-market session, and the New York open to close move, which sits entirely after it.

WITH daily AS
(
    SELECT
        ticker,
        date,
        toFloat64(open)  AS open_px,
        toFloat64(close) AS close_px,
        lagInFrame(toFloat64(close)) OVER (PARTITION BY ticker ORDER BY date
            ROWS BETWEEN 1 PRECEDING AND CURRENT ROW) AS prior_close_px
    FROM global_markets.stocks_daily_aggs
    WHERE ticker IN ('SONY', 'TM', 'ASML', 'SAP', 'HSBC', 'KO')
      AND date >= '2023-08-01'
      AND date <  '2026-08-01'
)
SELECT
    ticker,
    round(quantileDeterministic(0.5)(abs(open_px / prior_close_px - 1) * 100, toUInt32(date)), 3) AS overnight_gap_pct,
    round(quantileDeterministic(0.5)(abs(close_px / open_px - 1) * 100, toUInt32(date)), 3)       AS open_to_close_pct
FROM daily
WHERE prior_close_px > 0
  AND open_px > 0
GROUP BY ticker
ORDER BY indexOf(['SONY', 'TM', 'ASML', 'SAP', 'HSBC', 'KO'], ticker)

Read the bars in pairs. SONY, a Tokyo name, carries a typical New York open to close move of 0.558%, against 0.888% for the overnight gap that holds the whole Tokyo day. The European listings in the middle of the chart cover a shorter gap, four and a half hours rather than thirteen. The New York control at the end, KO, inverts the shape: 0.232% overnight against 0.542% in the session it actually trades. These are medians of absolute moves, so one dividend or corporate action cannot tilt them.

Prices keep moving long after the home market shuts

Zoom into the New York clock and the missing information shows up half hour by half hour. The panel below takes the US listings of a Japanese and a German company over the three months to August 2026, groups every minute bar by its New York half hour, and measures the average size of the move inside each bucket in basis points. One basis point is one hundredth of a percentage point.

WITH buckets AS
(
    SELECT
        ticker,
        toDate(toTimeZone(window_start, 'America/New_York')) AS et_day,
        formatDateTime(toStartOfInterval(toTimeZone(window_start, 'America/New_York'), INTERVAL 30 MINUTE), '%H:%i') AS et_time,
        argMin(toFloat64(open), window_start)  AS bucket_open,
        argMax(toFloat64(close), window_start) AS bucket_close
    FROM global_markets.delayed_stocks_minute_aggs
    WHERE ticker IN ('SONY', 'SAP')
      AND window_start >= '2026-05-01 00:00:00'
      AND window_start <  '2026-08-01 00:00:00'
    GROUP BY ticker, et_day, et_time
)
SELECT
    et_time,
    round(avgIf(abs(bucket_close / bucket_open - 1) * 10000, ticker = 'SONY'), 1) AS japan_adr_bps,
    round(avgIf(abs(bucket_close / bucket_open - 1) * 10000, ticker = 'SAP'), 1)  AS europe_adr_bps
FROM buckets
WHERE bucket_open > 0
GROUP BY et_time
HAVING countIf(ticker = 'SONY') >= 40
   AND countIf(ticker = 'SAP')  >= 40
ORDER BY et_time

The earliest bucket that prints on most days is 04:00 New York time, where the Japanese listing averages 45.8 bps and the German one 56.7 bps. Tokyo had shut hours before that bucket opened. The last bucket in the panel, 16:30, still averages 21.1 bps on the Japanese line. A fund valuing that same company at its Tokyo close carries a price struck before any of it happened.

What fair value pricing does to an international fund's NAV

US funds value a holding at market when a current market quotation is readily available. When one is not, the fund's board, or the valuation designee the board appoints under the SEC's fair value framework in the Investment Company Act, sets a fair value in good faith and documents the method. A foreign stock whose exchange closed thirteen hours ago is the standard case, and most international funds handle it with a systematic model rather than a hand adjustment.

The inputs are things that did trade after the foreign close: futures on the home-market index, US-listed shares of the same or comparable companies, the currency, and broad US sector moves. From those the model estimates, holding by holding, where the price would plausibly have stood at 4:00 p.m. ET. Many funds gate it with a trigger. If the reference basket moves past a set threshold during US hours, the adjustment applies for that day; under the threshold, the raw foreign closes stand.

The purpose is defensive. Without an adjustment, a foreign fund's 4:00 p.m. NAV is a price anyone can see is out of date, and an order entered at 3:59 p.m. buys the previous Tokyo close with a full New York day of information in hand. That is time-zone arbitrage, and the money it makes comes out of the shareholders who sit still.

The co-movement that makes the model's inputs usable is measurable. The panel sorts five years of sessions by how much a basket of five US large caps moved between the open and the close, then averages what the two foreign listings did over the identical hours.

WITH day_moves AS
(
    SELECT
        ticker,
        date,
        (toFloat64(close) / toFloat64(open) - 1) * 100 AS open_to_close_pct
    FROM global_markets.stocks_daily_aggs
    WHERE ticker IN ('AAPL', 'MSFT', 'JPM', 'XOM', 'KO', 'SONY', 'SAP')
      AND date >= '2021-08-01'
      AND date <  '2026-08-01'
      AND open > 0
),
paired AS
(
    SELECT
        date,
        avgIf(open_to_close_pct, ticker IN ('AAPL', 'MSFT', 'JPM', 'XOM', 'KO')) AS us_large_cap_pct,
        avgIf(open_to_close_pct, ticker = 'SONY')                                AS japan_pct,
        avgIf(open_to_close_pct, ticker = 'SAP')                                 AS europe_pct
    FROM day_moves
    GROUP BY date
    HAVING countIf(ticker IN ('AAPL', 'MSFT', 'JPM', 'XOM', 'KO')) = 5
       AND countIf(ticker = 'SONY') = 1
       AND countIf(ticker = 'SAP')  = 1
)
SELECT
    multiIf(us_large_cap_pct <= -1.0, 'US large caps: -1% or lower',
            us_large_cap_pct <= -0.3, 'US large caps: -1% to -0.3%',
            us_large_cap_pct <   0.3, 'US large caps: -0.3% to +0.3%',
            us_large_cap_pct <   1.0, 'US large caps: +0.3% to +1%',
                                      'US large caps: +1% or higher') AS us_market_move_bucket,
    round(avg(japan_pct), 3)  AS japan_adr_pct,
    round(avg(europe_pct), 3) AS europe_adr_pct,
    count()                   AS observation_count
FROM paired
GROUP BY us_market_move_bucket
ORDER BY min(us_large_cap_pct)

On the sessions where the US basket lost more than 1% between the open and the close, 105 of them, the Japanese listing averaged -1.272% and the German one -1.264% over the same window. On the sessions it gained more than 1%, the pair averaged 1.038% and 1.337%. Tokyo and Frankfurt were closed for every minute of it. That co-movement is what a fair value model is built to capture.

Why the fund's daily move will not match the index

The headline number for a foreign market is built from that market's own closing prices. A fair valued NAV is built from those closes plus an estimate of the hours that followed. The two cover different windows on the clock, so their daily percentages differ, sometimes by a wide margin.

Take a hypothetical two-day sequence. A Tokyo index closes Tuesday down 1.0%, and during Tuesday's New York hours the reference basket falls a further 0.8%. A fund that fair values its Japanese sleeve books something near 1.8% of decline in Tuesday's NAV, while the index headline for Tuesday still reads 1.0%. Wednesday, Tokyo opens lower and finishes down 0.8%: the index headline shows that full 0.8%, and the fund's NAV, having already taken the hit on Tuesday, moves much less. Across the pair of days the two measures land in the same neighbourhood. Read either day alone and the fund looks broken.

That gap is a timing difference in the same information, not a pricing error and not the slow drift that tracking error describes. It appears on the days the model triggers and it unwinds over the session that follows.

Does this happen to international ETFs as well?

International ETFs strike a NAV the same way and fair value the same holdings. Their market price also trades continuously in New York, so the 4:00 p.m. price already carries the post-close information the NAV has to estimate. A published premium or discount on an international ETF is partly a comparison of two different answers to one question, which is one reason those readings can run wider than on a domestic fund. ETF premium and discount to NAV takes that apart, the wrapper differences sit in mutual funds versus ETFs, and the cash timing after a sale is in mutual fund settlement time.

WITH daily AS
(
    SELECT
        date,
        toFloat64(open)  AS open_px,
        toFloat64(close) AS close_px,
        lagInFrame(toFloat64(close)) OVER (ORDER BY date
            ROWS BETWEEN 1 PRECEDING AND CURRENT ROW) AS prior_close_px
    FROM global_markets.stocks_daily_aggs
    WHERE ticker = 'SONY'
      AND date >= '2016-01-01'
      AND date <  '2026-08-01'
)
SELECT
    toString(toYear(date)) AS year,
    round(quantileDeterministic(0.5)(abs(open_px / prior_close_px - 1) * 100, toUInt32(date)), 3) AS overnight_gap_pct,
    round(quantileDeterministic(0.5)(abs(close_px / open_px - 1) * 100, toUInt32(date)), 3)       AS open_to_close_pct
FROM daily
WHERE prior_close_px > 0
  AND open_px > 0
GROUP BY year
ORDER BY year

FAQ

What is fair value pricing in a mutual fund?

It is the process a fund follows when a current market quotation for a holding is not readily available. The board, or the valuation designee it appoints under the SEC's fair value framework, sets a good faith value with a documented, tested method. For international funds the everyday case is a foreign exchange that closed hours before the 4:00 p.m. ET NAV strike.

Why doesn't my international fund's daily move match the foreign index?

The index is calculated from the home market's closing prices. The fund's NAV can carry a fair value adjustment for the hours that passed after that close. The two numbers cover different windows on the clock, so they disagree on the day, and the difference narrows once the home market has traded again.

Does fair value pricing apply to international ETFs?

Yes. An ETF's NAV is struck at 4:00 p.m. ET from the same kind of inputs. The difference is that the ETF's own market price trades all day in New York, so post-close information is already in the price a buyer pays, which is what a published premium or discount measures against.

Is a fair value adjustment the same as tracking error?

No. Tracking error describes a fund's drift from its benchmark over time, from costs and portfolio sampling. A fair value adjustment is a timing difference: the same information reaches the NAV a session before it reaches the index print.

How can I tell whether a fund uses fair value pricing?

The prospectus and the statement of additional information describe the valuation policy, including the trigger and the vendor model where one is used. Day to day, the visible sign is a NAV change that parts company with the home market index on a session when New York hours were busy.

Every panel here ships with the SQL that produced it, so the open to close split can be rerun on any company with a US listing. Ask for it in plain English on the Strasmore terminal.