Key takeaways
- Across 18 advanced economies from 1950 to 2020, real house prices rose an average 0.5% in the 110 country-years when inflation ran above 10%, and they rose in exactly half of those years.
- Equities did worse in the same 1,176 country-years: an average real total return of -2.0% when inflation was above 10%, positive in only 34.5% of those years.
- The UK's 1970s show both halves of the story. Real house prices rose 40.7% over 1970 to 1979 despite cumulative inflation of 209%, but fell 28.9% across 1974 to 1977, when annual inflation peaked at 22.7%.
- The 2021 to 2023 spike was a failed test in the UK: Nationwide's RPI-adjusted series shows real house prices down 14.1% between 2020 Q4 and 2026 Q2, even though nominal prices rose 21.3%.
- In the US, the Case-Shiller national index rose 22.0% in real terms from June 2020 to June 2022, then lost 3.0% in real terms in the four years to June 2026 while CPI rose 12.7%.
Does housing keep up with inflation? Over decades, yes. In the inflation years themselves, mostly not
Do house prices keep up with inflation? Is a house the hedge people say it is? The answer depends entirely on the horizon you ask about, and that's the whole finding.
Over long stretches, housing has beaten inflation comfortably. Nationwide's own RPI-adjusted series has UK real house prices up 99.5% between 1975 Q1 and 2026 Q2, which is 1.35% a year above retail prices. In the US, Robert Shiller's index of real home prices sits at 215.7 in June 2026 against 100 in 1890, a real gain of 0.56% a year over 136 years.
In the years when inflation is actually running hot, the record is different. Across 18 advanced economies from 1950 to 2020, the 110 country-years with inflation above 10% saw real house prices rise by an average of 0.5%. The median year was flat. Real prices rose in exactly half of those years and fell in the other half. In the UK and the US specifically, the high-inflation years were worse than that average. So the housing inflation hedge is real as a decade-long store of value, and unreliable as protection in the year you'd want it. The rest of this piece is the evidence for both halves, and then the strongest case against reading it that way.
The housing inflation hedge across five inflation regimes, 18 countries, 1950 to 2020
The test uses the Jordà-Schularick-Taylor Macrohistory Database, Release 6, which carries annual nominal house price indices, consumer price indices and total returns on housing and equities for 18 advanced economies since 1870. Each country-year was sorted into a regime by that year's consumer price inflation, and the real change in house prices was computed by deflating the nominal index by the CPI. Post-1950 is the cleaner sample: the earlier decades contain two world wars and 26 country-years of inflation at 50% or more, which are excluded from every full-sample figure below. That leaves 1,176 country-years from 1950 to 2020.
| Inflation regime | Country-years | Average inflation | Real house price change, mean | Real house price change, median | Years real prices rose | Real housing total return | Real equity total return | Years equities rose in real terms |
|---|---|---|---|---|---|---|---|---|
| Below 0% (deflation) | 64 | -0.6% | 3.6% | 2.7% | 75.0% | 7.4% | 12.8% | 67.2% |
| 0% to 2% | 342 | 1.1% | 3.2% | 3.3% | 71.3% | 7.7% | 14.1% | 68.7% |
| 2% to 5% | 443 | 3.2% | 2.6% | 2.5% | 68.2% | 7.5% | 7.1% | 56.9% |
| 5% to 10% | 217 | 7.1% | 3.0% | 1.3% | 57.1% | 8.3% | 5.8% | 50.7% |
| Above 10% | 110 | 13.9% | 0.5% | 0.0% | 50.0% | 4.3% | -2.0% | 34.5% |
The pattern is monotone in the column that matters. The share of years in which real house prices rose falls from 75.0% in deflation to 71.3% in low inflation, 68.2% in moderate inflation, 57.1% between 5% and 10%, and 50.0% above 10%. The chart above plots the mean real change by regime, and the drop from around 3% to 0.5% at the top end is the finding. The 5% to 10% band has a mean of 3.0% but a median of 1.3%, so a few strong years are carrying a typical year that barely kept pace.
Above 10%, the distribution is wide as well as centred on zero. Of those 110 country-years, real house prices fell by more than 10% in 17 of them and rose by more than 10% in 14. A 15.5% chance of a double-digit real loss is not what "hedge" usually means.
Now the comparison column. Real equity total returns in the same high-inflation years averaged -2.0%, and equities delivered a positive real year only 34.5% of the time. Real returns on housing, once net rent is counted as well as price, averaged 4.3% above 10% inflation. Relative to shares, housing was the better place to be in an inflationary year. Relative to inflation itself, price alone was a coin toss.
The full sample from 1871 is harsher still. Across 2,012 country-years with inflation below 50%, the 213 years above 10% inflation saw real house prices fall by an average of 3.3%, with real gains in only 35.2% of them. It includes the wartime and interwar episodes, which is why the post-1950 table leads.
In the UK and the US, the high-inflation years were worse than the 18-country average
The UK series in the database runs from 1900 to 2020 and contains 14 years with inflation above 10%. Real house prices fell in those years by an average of 4.5%, and rose in only 28.6% of them, which is 4 years out of 14. The US series runs from 1891 to 2020 and has 8 such years. The average real change was -4.3%, and real prices rose in just 1 of the 8.
Those are small samples, and the UK count is dominated by the decade that gave housing its reputation.
The 1970s built the reputation, and the inflation years inside the decade contradict it
Take the UK from 1970 to 1979. Cumulative inflation was 209%, so the price level roughly trebled. Real house prices ended the decade 40.7% higher. Real equity total returns were -12.1% over the same ten years. Counting net rent, the real total return on housing was 96.9%. On a decade view, housing beat inflation and beat shares by a wide margin.
Now look inside the decade. From 1974 to 1977, the four years around the inflation peak, cumulative inflation was 89% and annual inflation reached 22.7% in 1975. Real house prices fell 28.9% over those four years. Nationwide's RPI-adjusted quarterly series, which begins in 1975, shows the same thing at a finer grain: a typical property worth £139,749 in today's money at 1975 Q1 was worth £114,424 by 1977 Q2, a real fall of 18.1% in nine quarters.
The decade's real gain came from the boom of 1972 and 1973, before inflation peaked, and from the recovery after 1977. The hedge held over ten years because the sequence happened to run boom, inflation, recovery. Anyone who bought in 1974 got the inflation without the boom.
The US tells a quieter version. From 1973 to 1982, cumulative inflation was 120%. Real house prices ended the ten years up 0.3%, so they kept pace and no more. Real equity total returns were -14.6%. The real total return on housing including rent was 64.0%. Again housing beat shares and matched inflation, by the narrowest of margins on price alone.
2021 to 2026: the most recent inflation, and UK real house prices lost 14.1%
The database ends in 2020, so the latest episode has to be measured from the index publishers directly. Nationwide deflates its own index by the ONS Retail Prices Index. On that basis a typical UK property cost £229,819 at 2020 Q4 and £278,784 at 2026 Q2, a nominal rise of 21.3%. In real terms it fell 14.1%. Real prices reached their post-2020 high at 2022 Q1 and have lost 17.6% since. UK CPI inflation peaked at 11.1% in October 2022, with RPI at 14.2% in the same month, and between December 2020 and June 2026 CPI rose 30.5% while RPI rose 41.0%.
The deflator matters here. Nationwide's choice of RPI, which has risen faster than CPI over the period, makes the real fall larger. Deflating the same 21.3% nominal rise by the 30.5% CPI increase instead gives a real fall of 7.0%. On either measure the answer is a real loss. The 17.6% real drawdown from the 2022 Q1 high sits in the same family as the two previous UK real bear markets in Nationwide's series: 37.4% from 1989 Q2 to 1995 Q4, and 25.9% from 2007 Q3 to 2013 Q1. Those two came with nominal falls that everyone noticed. This one arrived with nominal prices still rising, which is why it's felt so little.
The US split the episode in two. The S&P Cotality Case-Shiller national index rose from 219.769 in June 2020 to 308.065 in June 2022, a nominal gain of 40.2%, while the CPI rose 14.9%. That's a real gain of 22.0%. Then inflation peaked at 9.1% in June 2022 on the not seasonally adjusted CPI-U index, and from that month to June 2026 the index rose only 9.3% to 336.663 while the CPI rose 12.7% to 333.952. Real house prices fell 3.0% over those four years. Meanwhile the S&P Composite index in Shiller's data returned 79.3% in real terms over the same four years, against 12.5% in the two years before.
The mechanism is visible in the timing. US house prices surged in 2020 and 2021, before consumer inflation arrived. Once inflation showed up, the policy response was rates. The effective federal funds rate went from 0.08% in January 2022 to 5.33% in August 2023, and Bank Rate went from 0.25% in December 2021 to 5.25% in August 2023. The price of a mortgage payment rose with them. Inflation reached house prices through the central bank, and the transmission was negative. That's the same shape as the UK 1970s: real gains arrive before the inflation, and the inflation years themselves are the drawdown.
Price indices miss the rent, and the rent is what kept housing positive
Every house price index, from Nationwide to Case-Shiller to the database's own series, measures price and nothing else. Housing's total return also includes net rent, and in the high-inflation regime that leg is most of the return. Above 10% inflation, the real change in house prices averaged 0.5% while the real total return on housing averaged 4.3%. The gap is rent, which the database counts net of maintenance costs, ground rent and other irrecoverable expenditure. For a leaseholder those costs arrive as a charge somebody else budgets, and no statute caps how far a service charge increase can run ahead of the index.
Whether that rent reaches you depends on what you own. A landlord banks it, and a landlord's rents can reprice with inflation. An owner-occupier receives it as shelter, which is genuinely valuable but doesn't compound inside the asset and can't be rebalanced into anything else. The full decomposition of that return, and what it means for treating your house as an investment, is a separate piece. For the hedge question, the point is narrower: the part of housing's return that has held up in inflation is the part a homeowner never sees as money.
There's a second leg most people feel more directly. A mortgage is a fixed nominal debt. When the price level rises, the real value of that debt falls, whatever the house does. That's a benefit of inflation to the borrower, not a property of housing as an asset, and it accrues to anyone with fixed-rate debt of any kind. It's often what people mean when they say property protected them, and it shouldn't be confused with the price of the house keeping up, which in the UK since 2020 it hasn't.
The strongest case for the hedge comes from the dataset's own authors
The economists who built the database, in the working-paper version of The Rate of Return on Everything, 1870-2015, read their own data as supportive of housing. Their words: "equity co-moved negatively with inflation in the 1970s, while housing provided a more robust hedge against rising consumer prices." And more broadly: "apart from the interwar period when the world was gripped by a general deflationary bias, equity returns have co-moved negatively with inflation in almost all eras."
That reading is correct, and the table above agrees with it. Housing's total return stayed positive in the high-inflation regime and equities' did not. Real returns on bills, the cash alternative, averaged -3.1% in the same years. If the question is "which of the big asset classes lost least when inflation was high", housing wins, and it wins on the price-only measure too.
The disagreement is about the word "hedge". A hedge is something that pays off when the thing you fear happens. What the record shows is an asset that loses less than the alternatives, and that recovers its purchasing power over the following decade, provided the sequence cooperates. Those are different properties, and the second one is a claim about ten-year patience rather than about inflation. The same distinction runs through the record for gold, where the case for how much gold in a portfolio rests on a decades-long real drawdown as much as on the 1970s, and through the much shorter record for a bitcoin inflation hedge.
There's also a longer history that cuts against treating the post-1950 gains as a law of nature. Knoll, Schularick and Steger's No Price Like Home, published in the American Economic Review in 2017, assembled house prices for 14 advanced economies since 1870. Their abstract: "real house prices stayed constant from the nineteenth to the mid-twentieth century, but rose strongly and with substantial cross-country variation in the second half of the twentieth century." They attribute about 80 percent of the post-war boom to rising land prices. In the working-paper version they note that between 1899 and 1938, UK house prices fell on average by 1 percent per year. Shiller's US series says the same: real home prices went from 100 in 1890 to 112.3 in January 1997, a real gain of 0.11% a year over 107 years. Almost all of the 0.56% a year since 1890 was earned after 1997. Housing's long-run real return is a story about land scarcity in one 70-year window, not about inflation.
What this evidence can't tell you
These are national indices, deflated annually, across countries with different mortgage markets, tax treatments and planning regimes. Your house is one property in one street, bought with leverage and sold with costs, and none of that is in the table. The database's house price series is a nominal index deflated by the CPI, so it doesn't capture quality change, transaction costs, or the rent an owner consumes; the total return column is the authors' attempt at the last of those and is the only one that includes it. The fees behind a vehicle crossover are a small example of the costs that sit outside an index like this entirely.
The regime table sorts each year by that year's inflation. It doesn't ask what happened in the year after, and the 1970s sequence shows that the recovery often came once inflation had passed. A three-year window would give a kinder answer for housing and a harsher one for the "in the year you need it" framing, and both would be true.
The samples are small where it matters. The 110 high-inflation country-years since 1950 are concentrated in the 1970s and early 1980s, and 14 of the UK's high-inflation years and 8 of the US's are a thin base for country-level claims. The database ends in 2020, so the 2021 to 2023 episode isn't in the regime table at all; it's measured separately above from Nationwide and Case-Shiller, using RPI in one case and CPI in the other. A backtest of past inflation is not a forecast of the next one, and the next one might arrive with a different central bank reaction, which is the point the closing section makes.
What would change the conclusion
If the policy response to inflation changes. The channel that broke the hedge in 2022 was interest rates. Inflation led to higher rates, higher rates led to smaller mortgages, and smaller mortgages led to lower real prices. An inflation that central banks chose not to fight would run through house prices differently, and the 1970s UK, where real house prices rose 40.7% across the decade, is partly a record of that.
If you measure total return rather than price. On the database's total-return series, housing's real return above 10% inflation was 4.3%, positive and well ahead of equities. For a landlord with repricing rents, the hedge claim survives in a weaker form. For an owner-occupier, the price-only number is the one that governs what the house is worth. A second property kept for holidays is in the same position, and the second home total return leans almost entirely on price as a result.
If the deflator is CPI rather than RPI. Nationwide's real series uses RPI, which rose 41.0% between December 2020 and June 2026 against 30.5% for CPI. On CPI, the UK real fall since 2020 Q4 is 7.0% rather than 14.1%. The sign doesn't change. The size does, and a reader who prefers CPI is reading the same data correctly.
If the next inflation comes with a different sequence. The decade-long gains in the 1970s came from a boom that preceded the inflation. If the next episode starts from a market that has already fallen in real terms, as the UK has since 2022 Q1, the arithmetic of a recovery starts from a lower base. Nationwide reported annual house price growth of 1.6% in August 2026, against CPI inflation of 3.1% in the same month, which means the real series is still falling as of the latest print.
The number to watch is the real one. Nationwide publishes its RPI-adjusted series every quarter, and it's the only headline house price figure in the UK that answers the hedge question directly. LedgerTouch holds property at nominal value alongside the rest of the balance sheet; whether that value has kept up is a question the real series answers and the nominal one hides.