ASSETCASTING

After a 10% decline in real property prices, how long until recovery to the observed peak?

Japan, Korea, United States and China · −10% from peak → recovery to prior peak

Total 8 · Completed 4 · Ongoing 4 · Excluded 0

Last observed 2023 Q4 · Updated 2026-10-09

Completed cases4 casesOngoing and excluded cases are not included
Duration Median41 quartersMean 38.8 quarters
Completed outcome range16 quarters ~ 57 quartersMinimum ~ maximum

Same conditions, different outcomes

Completed and elapsed durations are distinguished. Cases with unconfirmed endings are excluded from completed statistics. Historical averages do not predict the remaining duration. Open a case to explore its context.

China · 2023 Q4 · −10% from observed real-price peak · End unconfirmedEnd unconfirmed · 2023 Q4—

Observed 2023 Q4 → End unconfirmed · As of 2026-03-31

Country and data-definition segment
China · Existing-home prices in 70 cities
Signal confirmation quarter
2023 Q4
Previously observed peak within the same data segment
2021 Q3 · 112.9908
Real-price index at entry
101.4751
Quarter confirming the exit criterion
Not confirmed
Last comparable quarter
2026 Q1
Treatment of subsequent data
Prices and exit status after this quarter are not estimated

The exit condition was not confirmed at the observed price of 85.1336 through 2026 Q1. The actual subsequent end is unknown, and the duration is not extended to the verification date, 2026-10-09. This concerns the real index in the relevant China data segment, not recovery of an individual home’s nominal price or invested principal.

United States · 2007 Q3 · −10% from observed real-price peakCompleted · 2007 Q357 quarters

Observed 2007 Q3 → 2021 Q4 · As of 2026-03-31

Country and data-definition segment
United States · Existing-home prices; domestic input series seasonally adjusted
Signal confirmation quarter
2007 Q3
Previously observed peak within the same data segment
2006 Q1 · 153.9192
Real-price index at entry
135.3987
Quarter confirming the exit criterion
2021 Q4
Last comparable quarter
2026 Q1
Treatment of subsequent data
Prices and exit status after this quarter are not estimated

The difference from 2007 Q3 to 2021 Q4 is 57 quarters. These are boundaries identified from quarterly average indices, not actual events on the first day of a quarter. This concerns the real index in the relevant United States data segment, not recovery of an individual home’s nominal price or invested principal.

Japan · 1993 Q1 · −10% from observed real-price peak · Tracking stopped at a data-definition transitionEnd unconfirmed · 1993 Q1—

Observed 1993 Q1 → End unconfirmed · As of 2008-03-31

Country and data-definition segment
Japan · Urban residential land prices
Signal confirmation quarter
1993 Q1
Previously observed peak within the same data segment
1991 Q1 · 189.2763
Real-price index at entry
169.9309
Quarter confirming the exit criterion
Not confirmed
Last comparable quarter
2008 Q1
Treatment of subsequent data
Peak comparisons discontinued after changes in data definitions, such as land versus housing

Data definitions change after 2008 Q1, so peaks and recovery status cannot be compared for the same underlying asset coverage. The case is retained as one without a confirmed end in the prior data segment; it is not asserted to have remained unrecovered afterward or through the present. Elapsed duration is not provided. This concerns the real index in the relevant Japan data segment, not recovery of an individual home’s nominal price or invested principal.

  • BIS Real Residential Property Prices · Japan ↗BIS; Ministry of Land, Infrastructure, Transport and Tourism; Japan Real Estate Institute · Checked 2026-10-09 · BIS Q.JP.R.628; real index2010=100, deflated by CPI; comparison segment 1955 Q1–2008 Q1; national sources: Ministry of Land, Infrastructure, Transport and Tourism; Japan Real Estate Institute
  • BIS Selected Residential Property Prices · compilation and national sources ↗Bank for International Settlements (BIS) · Checked 2026-10-09 · 2026-07-31 version; CPI deflation and quarterly observations; historical interpolation warning; source breaks JP2008Q2/KR1986Q1/US1975Q4/CN2016Q1
United States · 1992 Q3 · −10% from observed real-price peakCompleted · 1992 Q330 quarters

Observed 1992 Q3 → 2000 Q1 · As of 2026-03-31

Country and data-definition segment
United States · Existing-home prices; domestic input series seasonally adjusted
Signal confirmation quarter
1992 Q3
Previously observed peak within the same data segment
1989 Q4 · 94.0594
Real-price index at entry
84.0731
Quarter confirming the exit criterion
2000 Q1
Last comparable quarter
2026 Q1
Treatment of subsequent data
Prices and exit status after this quarter are not estimated

The difference from 1992 Q3 to 2000 Q1 is 30 quarters. These are boundaries identified from quarterly average indices, not actual events on the first day of a quarter. This concerns the real index in the relevant United States data segment, not recovery of an individual home’s nominal price or invested principal.

Korea · 1992 Q1 · −10% from observed real-price peak · End unconfirmedEnd unconfirmed · 1992 Q1—

Observed 1992 Q1 → End unconfirmed · As of 2026-03-31

Country and data-definition segment
Korea · Existing-home prices
Signal confirmation quarter
1992 Q1
Previously observed peak within the same data segment
1991 Q2 · 151.2834
Real-price index at entry
134.9463
Quarter confirming the exit criterion
Not confirmed
Last comparable quarter
2026 Q1
Treatment of subsequent data
Prices and exit status after this quarter are not estimated

The exit condition was not confirmed at the observed price of 105.6515 through 2026 Q1. The actual subsequent end is unknown, and the duration is not extended to the verification date, 2026-10-09. This concerns the real index in the relevant Korea data segment, not recovery of an individual home’s nominal price or invested principal. Nationwide existing-home and historical land indicators are used, so this cannot be interpreted as claiming that the nominal price of an individual Seoul apartment has not recovered.

United States · 1982 Q3 · −10% from observed real-price peakCompleted · 1982 Q316 quarters

Observed 1982 Q3 → 1986 Q3 · As of 2026-03-31

Country and data-definition segment
United States · Existing-home prices; domestic input series seasonally adjusted
Signal confirmation quarter
1982 Q3
Previously observed peak within the same data segment
1979 Q4 · 82.2821
Real-price index at entry
74.0087
Quarter confirming the exit criterion
1986 Q3
Last comparable quarter
2026 Q1
Treatment of subsequent data
Prices and exit status after this quarter are not estimated

The difference from 1982 Q3 to 1986 Q3 is 16 quarters. These are boundaries identified from quarterly average indices, not actual events on the first day of a quarter. This concerns the real index in the relevant United States data segment, not recovery of an individual home’s nominal price or invested principal.

Korea · 1980 Q1 · −10% from observed real-price peak · Tracking stopped at a data-definition transitionEnd unconfirmed · 1980 Q1—

Observed 1980 Q1 → End unconfirmed · As of 1985-12-31

Country and data-definition segment
Korea · Residential and non-residential land prices
Signal confirmation quarter
1980 Q1
Previously observed peak within the same data segment
1978 Q3 · 119.1278
Real-price index at entry
104.3959
Quarter confirming the exit criterion
Not confirmed
Last comparable quarter
1985 Q4
Treatment of subsequent data
Peak comparisons discontinued after changes in data definitions, such as land versus housing

Data definitions change after 1985 Q4, so peaks and recovery status cannot be compared for the same underlying asset coverage. The case is retained as one without a confirmed end in the prior data segment; it is not asserted to have remained unrecovered afterward or through the present. Elapsed duration is not provided. This concerns the real index in the relevant Korea data segment, not recovery of an individual home’s nominal price or invested principal. Nationwide existing-home and historical land indicators are used, so this cannot be interpreted as claiming that the nominal price of an individual Seoul apartment has not recovered.

Japan · 1974 Q3 · −10% from observed real-price peakCompleted · 1974 Q352 quarters

Observed 1974 Q3 → 1987 Q3 · As of 2008-03-31

Country and data-definition segment
Japan · Urban residential land prices
Signal confirmation quarter
1974 Q3
Previously observed peak within the same data segment
1973 Q3 · 153.1654
Real-price index at entry
136.6606
Quarter confirming the exit criterion
1987 Q3
Last comparable quarter
2008 Q1
Treatment of subsequent data
Prices and exit status after this quarter are not estimated

The difference from 1974 Q3 to 1987 Q3 is 52 quarters. These are boundaries identified from quarterly average indices, not actual events on the first day of a quarter. This concerns the real index in the relevant Japan data segment, not recovery of an individual home’s nominal price or invested principal.

  • BIS Real Residential Property Prices · Japan ↗BIS; Ministry of Land, Infrastructure, Transport and Tourism; Japan Real Estate Institute · Checked 2026-10-09 · BIS Q.JP.R.628; real index2010=100, deflated by CPI; comparison segment 1955 Q1–2008 Q1; national sources: Ministry of Land, Infrastructure, Transport and Tourism; Japan Real Estate Institute
  • BIS Selected Residential Property Prices · compilation and national sources ↗Bank for International Settlements (BIS) · Checked 2026-10-09 · 2026-07-31 version; CPI deflation and quarterly observations; historical interpolation warning; source breaks JP2008Q2/KR1986Q1/US1975Q4/CN2016Q1

How to interpret this

This is recovery of the prior peak of the CPI-adjusted index, not recovery of nominal housing prices or invested principal. Tracking of the case concerning Japan’s 1991 land-price peak stops at the 2008 data-definition transition; it is not spliced to the subsequent housing index to claim non-recovery through 2026.

These are fixed price rules for examining bubble and overheating risks, not officially diagnosed bubbles. This fixed sample of 4 countries—Japan, Korea, the United States and China—does not represent all of Asia or the world. The real index divides nominal housing prices by CPI; it differs from investment returns incorporating inflation, housing costs and rents, leverage, taxes and transaction costs. Differences in asset coverage, seasonal adjustment and data starting dates make direct cross-country level comparisons difficult. Historical BIS data may contain quarterly interpolations of annual source data and revisions in the current vintage. This is not a trading strategy recreating signals actually published at the time. Averages limited to completed cases omit long-running unended cases, and statistical independence across countries and periods is not guaranteed. This Korean reprocessed analysis is neither an official BIS translation nor an approved result of BIS or national statistical agencies. Users are not charged additional fees for the included BIS statistics.

Conditions, methodology and sources

Condition: entry occurs in the first quarter at least 10% below the highest index already observed within the same data-definition segment. The reference peak immediately preceding entry is fixed; exit occurs in the first quarter recovering to or above that level. Repeated declines before recovery are not counted as new cases. If only the segment’s first observation is available as the reference peak, earlier price history is unknown and the case is excluded from statistics. The full original quarterly history is used, but tracking stops at official data-definition transitions in Japan 2008Q2, Korea 1986Q1, United States 1975Q4 and China 2016Q1; peaks and annual growth rates are recalculated. Only one case is open at a time per country and data segment; overlapping entries during that case are recorded in audit data but not counted as separate cases. Entry dates are not selected using future prices or future peaks. The criteria—annual 10% growth and a 10% decline from the peak—were fixed before reviewing outcomes and are not official bubble thresholds. Annual growth is calculated only against the value 4 quarters earlier within the same definition segment; if the first calculable value already exceeds the threshold, the case is excluded for an uncertain starting boundary. Recovery is not asserted across data gaps or definition transitions, and unconfirmed cases are retained. Quarter labels refer to source observation quarters; day 1 in months 1, 4, 7 and 10 is a storage normalization. Completed duration is the difference in observation-quarter numbers, not a transition time measured by month or day. Each case’s reference date is the end of the last comparable data quarter; the source verification date is 2026-10-09. Reported peaks are updated sequentially using only values observed by that point.