According to the Family Income and Expenditure Survey (Income and Expenditure section) released by the Statistics Bureau of Japan in early October, households of two or more people spent ¥6,204 per household on lodging in August 2026, up 10.0% from ¥5,641 a year earlier. The January–August total reached ¥29,605, 23.1% above the ¥24,056 recorded in the same period of 2025, and not a single month in 2026 has come in below the prior year. Meanwhile, Japanese guest nights in the Japan Tourism Agency’s Overnight Travel Statistics Survey fell 3.6% over the same eight months — a seemingly contradictory picture in which “fewer people are staying, yet household lodging spend is rising.” This article first explains how to read the statistic, then splits spending growth into a price factor and a frequency/mix factor using the lodging CPI, and lays out the months in which the two official statistics move in opposite directions, tracing the gap to differences in population. We also show the distribution of annual averages across 52 prefectural capitals and designated cities, last winter’s (Nov–Jan) levels, and the next release date.

How to read “lodging” in the household survey — a monthly figure per household, including households that did not stay

Lodging in the Family Income and Expenditure Survey is not an average of households that stayed at hotels or ryokan. It is “expenditure per household per month,” obtained by summing spending across every surveyed household of two or more (including those that did not stay anywhere that month) and dividing by the number of households. The ¥6,204 figure therefore does not mean “households that stayed paid ¥6,204 on average”; it means “spreading the lodging spend of households that stayed thinly across all households of two or more in Japan gives ¥6,204 a month.” The number rises in the same way whether the share of households that stay (frequency) goes up or the payment per stay (price) goes up.

Four points are worth keeping in mind. First, the sample is small. According to the Statistics Bureau’s survey outline, 8,076 households of two or more are surveyed nationwide, and each household is replaced after six months. For an item like lodging, where only a limited number of households spend and each payment is large, a handful of households taking an expensive trip can move the monthly average. That is why single-month YoY can swing to +48% (February 2026) or −16% (July 2025), and this article therefore also shows a three-month moving average. Second, package tours are a separate item. The price of travel products bundling accommodation and transport is recorded under “9.4.2 Package tours,” not lodging. If buying habits shift from packages to independent bookings, lodging spend rises even with no change in nights. Third, from the January 2025 data the expenditure classification switched to the 2025 revision. The item code for lodging continued, so the time series remains comparable, but the layout of the published tables and the grouping of neighboring items differ before and after. Fourth, the population is limited to households of two or more and excludes single-person households. This matters later when we consider the gap with the JTA statistics.

Monthly trend, Jan 2024–Aug 2026 — August is 15.7% of the year, followed by Nov, Mar, Jul and Oct

First, the seasonal shape. On a three-year average for 2023–2025, August accounts for 15.7% of annual spending, standing far above every other month (5.5–9.4%). Next comes November (9.4%), followed almost evenly by March (8.8%), July (8.8%), October (8.8%) and May (8.7%). The troughs were June (5.5%) and February (5.7%). Summer-vacation family trips push up per-household spending, while March and May — with the fiscal-year changeover and holidays — and the autumn-foliage month of November form a second group of peaks. This shape broadly matches lodging-property seasonality, but note that the household survey records spending in the month the household paid, so prepaid lodging is booked in the month of payment rather than the month of stay.

Source: Compiled by HotelBank Editorial Team from the Ministry of Internal Affairs and Communications, Family Income and Expenditure Survey (Income and Expenditure section, households of two or more, 9.4.1 Lodging)

On an annual basis, the yearly average (12-month mean) rose from ¥2,530 in 2023 to ¥2,863 in 2024 and ¥3,000 in 2025 — up 18.6% in two years. In 2026, the Jan–Aug total of ¥29,605 is 23.1% above the same period of 2025 (¥24,056) and 29.5% above 2024 (¥22,858). Given that the 2025 Jan–Aug total was only 5.2% above 2024, the size of the increase since the start of 2026 stands out. By month, March 2026 reached ¥4,097 — the first time a month other than August has reached the ¥4,000 range — and May was ¥3,870, well above ¥3,227 a year earlier.

Reading 2026 growth by single month and 3-month moving average — Jan–Aug +23.1%, August alone +10.0%

Single-month YoY growth ran +29.8% in January, +48.1% in February, +26.3% in March, +22.7% in April, +19.9% in May, +20.3% in June, +25.6% in July and +10.0% in August. Smoothing the volatility with the YoY change in the three-month moving average gives +26.4% in February, a peak of +33.6% in March, then +31.4% in April, +23.0% in May, +20.9% in June, +21.9% in July and +16.1% in August — growth peaked in spring and has narrowed since. August’s +10.0% is the slowest single-month growth of the year, but August is also the month with the largest spending, and its absolute increase (+¥563) falls short of February (+¥1,089) and March (+¥853) while exceeding April (+¥560) and June (+¥417). It is too early to read the slowdown as a stall; it is more reasonable to see it as a reaction to the high level of August a year earlier (¥5,641, +4.9% YoY).

Source: Compiled by HotelBank Editorial Team from the Ministry of Internal Affairs and Communications, Family Income and Expenditure Survey (households of two or more, lodging). The 3-month moving average is the YoY change in the average of the latest three months, including the current month

Source: Compiled by HotelBank Editorial Team from the Family Income and Expenditure Survey and Consumer Price Index (Ministry of Internal Affairs and Communications) and the Overnight Travel Statistics Survey (Japan Tourism Agency). “Growth ex-price” = (1 + spend YoY) ÷ (1 + CPI YoY) − 1, an estimate

Splitting growth into “price” and “frequency/mix” — of the Jan–Aug +23.1%, the price factor is +3.8%

For lodging properties, an increase in lodging spend means something different depending on whether households paid more per night for the same number of nights (price factor) or stayed more often or with more people (frequency/mix factor). Here we use the lodging index of the Consumer Price Index (2020 = 100) and treat the residual from dividing nominal YoY growth by CPI YoY growth as an approximation of the frequency/mix factor. The lodging CPI averaged 168.0 in Jan–Aug 2026, up 3.8% from 161.8 a year earlier. Deflating the nominal +23.1% by this leaves +18.6%. In other words, most of the Jan–Aug growth cannot be explained by price increases and is estimated to lie on the “volume and mix” side — how often people stay, how many travel together and the grade of property they choose.

By month, lodging CPI YoY growth gradually narrowed from +5.9% in January and +6.0% in February to +2.0% in July and −0.2% in August, essentially flat. The August CPI a year earlier was 181.4, the highest since 2020, and it has not risen from there. As a result, August’s nominal +10.0% translates almost one-for-one into a real +10.2%, and price contributed effectively nothing. In January–March, by contrast, 5–6 points of nominal growth came from price and the remaining 20–40 points from frequency/mix. This residual also includes upgrades the CPI does not capture — “choosing a pricier room or a meal plan at the same property” — so reading it as a real 19% increase in nights would overstate it. It should be treated as an estimate with a margin. Note also that the lodging CPI moved to a 2025 base from the July 2026 data, and YoY changes differ by up to 1.6pt between the old and new bases. To keep the time series continuous, this article uses the 2020-base index throughout; the impact of the rebase is covered in “Japan Lodging CPI Rebased to 2025: YoY Cut by up to 1.6pt vs Your ADR.”

Source: Compiled by HotelBank Editorial Team from the Family Income and Expenditure Survey and Consumer Price Index (lodging) (Ministry of Internal Affairs and Communications) and the Overnight Travel Statistics Survey (Japan Tourism Agency)

The gap between the lodging CPI and our estimated settled ADR is broken down in “Lodging CPI +25.6% in 3 Years vs Settled ADR +10.9% — 14.7pt Gap.” The CPI is a price index that holds quality constant, and it moves differently from the market’s average settled rate. Using the CPI to deflate the household survey removes changes in the price households pay for lodging of the same quality — not changes in market ADR.

Opposite in direction to JTA’s Japanese guest nights — the sign differs in all eight months of 2026

According to the Japan Tourism Agency’s Overnight Travel Statistics Survey, Japanese guest nights changed by −2.3% in January 2026, −1.6% in February, −1.5% in March, −5.5% in April, −0.9% in May, −6.3% in June, −4.1% in July and −5.8% in August (first preliminary, 50.04 million guest nights), for a Jan–Aug total of −3.6%. Household lodging spend was positive in all of the same eight months, so the sign did not match in a single month. Setting the household survey, with a frequency/mix factor estimated at +19%, beside the JTA statistics, with nights down 3.6%, looks contradictory.

The gap arises not because one of them is wrong but because they measure different things. First, the populations differ. The household survey covers only households of two or more and excludes single-person households. In the 2020 Census, single-person households made up about 38% of private households, so stays by single people who travel on business often, or by younger and older solo travelers, do not appear in household lodging spend. Second, the payer differs. The travel statistics count the total number of guests who stayed at properties and so include company-paid business stays, whereas the household survey records only what households paid themselves. If corporate business travel declines while personal and family travel increases, guest nights can fall while household spending rises. Third, the household survey measures money, not people; larger parties and higher grades raise spending without adding nights. Fourth, from the January 2026 survey the Overnight Travel Statistics Survey changed its stratification basis from number of employees to number of rooms, so YoY comparisons may include the effect of this methodological change. Fifth, as noted above, the household survey’s sample is small and single months swing widely.

The two statistics are therefore best understood as measuring “Japanese lodging demand” from different angles. Guest nights are total volume seen from the property side; the household survey is the wallet of households, centered on families and couples. The fact that the former has fallen while the latter has risen in 2026 suggests a change in the quality of domestic lodging demand: “fewer people are staying, but spending per household on family trips is getting thicker.” Even within domestic demand, Japanese overnight travel consumption reported in the JTA’s Travel and Tourism Consumption Trend Survey (April–June quarter) is a separate statistic from both the household survey and the travel statistics, and differs from the figures in this article in scope and method. The pattern that survey shows for FY2025 — “fewer actual travelers but record total spending” — is covered in “Japan FY2025 Travel Spend ¥8.39tn: Growth Went Outside the Room Bill.” The annual outlook for domestic travelers is covered in “Japan Guest Nights Fall in 2026: Jan–Aug −5.0%, Full-Year Outlook,” and the combination of volume and price by prefecture in “Japan July 2026: Guest-Nights −3.5%, ADR −4.4% Across 47 Prefectures.”

Annual averages across 52 prefectural capitals and designated cities — median ¥2,637, half between ¥2,039 and ¥3,214

The household survey publishes results for prefectural capitals and government-designated cities (52 cities in total). Lining up the 12-month averages for 2025, the median is ¥2,637, the first quartile ¥2,039 and the third quartile ¥3,214, so half of the cities fall between ¥2,039 and ¥3,214. The minimum was ¥1,385 and the maximum ¥6,308 — 4.6 times the minimum. The national figure (¥3,000) is above the median because it is weighted by household numbers including households outside these cities, and because a small number of high-level cities pull the average up.

Source: Compiled by HotelBank Editorial Team from the Ministry of Internal Affairs and Communications, Family Income and Expenditure Survey (households of two or more, lodging, 2025 monthly average, 52 prefectural capitals and government-designated cities)

Source: Compiled by HotelBank Editorial Team from the Family Income and Expenditure Survey (Ministry of Internal Affairs and Communications) (N=52 cities)

We do not rank individual cities here because each city’s sample is only around a hundred households, and it is not unusual for a city’s annual average to move by tens of percent from the previous year. Indeed, several cities saw their annual average rise by 50% or more between 2024 and 2025, and several fell by 30% or more, so single-year rankings are not reproducible. What can be read is the shape of the distribution — the sense of level that “most cities sit in the ¥2,000–3,000 range per month, and only five exceed ¥4,000.” Households in the Tokyo area and other large cities tend to spend more, but this reflects not only income levels but also the longer distances of trips and homecomings that involve overnight stays; it does not indicate where the lodging properties are located.

Looking toward winter — last Nov–Jan totaled ¥9,456 over three months; next release November 10

As material for thinking about winter demand, here are last year’s levels. Lodging spend was ¥3,459 in November 2025 (+17.7% YoY), ¥2,795 in December (+4.9%) and ¥3,202 in January 2026 (+29.8%), for a three-month total of ¥9,456, 17.2% above ¥8,071 a year earlier. November, with autumn foliage and a three-day weekend, and January, when New Year homecomings and trips fall within the month, are the heavier spending months of winter. December, by contrast, tends to be relatively light for lodging because year-end spending goes to other categories.

Source: Compiled by HotelBank Editorial Team from the Ministry of Internal Affairs and Communications, Family Income and Expenditure Survey (households of two or more, lodging)

Whether winter 2026 beats last year depends on where the narrowing of growth since the autumn comes to a stop. YoY growth in the three-month moving average has shrunk from +33.6% in March to +16.1% in August; if that slope continues, it would reach single digits by year-end — but because last November and January were high, the same yen amount will show lower YoY growth. In practice it makes more sense to benchmark against yen levels (¥3,459 in November, ¥3,202 in January) rather than growth rates. The next release covers September 2026 and, according to the Statistics Bureau’s release schedule, will be published at 8:30 a.m. on Tuesday, November 10, 2026, together with the July–September quarterly average. September will be compared with ¥2,817 a year earlier (+3.9% YoY).

Implications for operators and investors — how to serve demand with “fewer people, thicker wallets”

To sum up, lodging spend by households of two or more rose 23.1% in Jan–Aug 2026, of which the price factor was +3.8%; the remaining roughly +19% is estimated to reflect changes in frequency, party size and grade mix. At the same time, Japanese guest nights fell 3.6%. From the property’s perspective, a strategy chasing total guest volume and a strategy capturing per-trip spending by family and couple households are now pointing in different directions.

First, the fact that the second peak in household spending is spread across November, March and May can inform product design for busy periods other than August. March in particular exceeded ¥4,000 for the first time outside August, at ¥4,097 in 2026, suggesting that spring-break, graduation and fiscal-year-end family trips are becoming established as lodging spend. Second, growth in the frequency/mix factor means households increased spending even in August, when the lodging CPI was flat — there is room to raise spend per guest through combinations of room type, meals and length of stay without raising prices. Third, trends among single-person households and corporate travel, which the household survey does not capture, are likely showing up on the declining side of guest nights; for weekday demand at business hotels, the travel statistics and a property’s own booking data are better leading indicators than the household survey. Reading each statistic by “whose spending, on what, counted how” is the precondition for not misjudging changes in the quality of demand. How to read the Overnight Travel Statistics Survey is explained in “Overnight Travel Statistics Survey Guide: 61.6% OCC, Prelim vs Final.”

Frequently Asked Questions