When housing prices climb, a familiar story tends to follow: homeowners feel richer, open their wallets, and the broader economy gets a boost. Central bankers and policymakers have long treated this “wealth effect” as one of the transmission belts that connects real estate markets to consumer spending. But a closer look at Korean household data suggests the story may be far narrower than the headline version implies.
A new study published in Applied Economics Letters finds that the boost to consumption from rising home prices shows up almost entirely among one group: households that own more than one property. Renters and people who own a single home, by contrast, show either no measurable response or a slightly negative one.
The question behind the research
Earlier work has established that housing wealth effects vary across households, but researchers have often studied average responses across whole populations, or looked at heterogeneity without tying it to a specific economic mechanism. Dong-Jin Pyo of Changwon National University set out to connect two threads that have tended to run in parallel: a structural model of household decision-making, and empirical evidence from household-level panel data.
The central question is straightforward. When home prices rise, which households actually translate that paper gain into higher spending, and why? Pyo argues that the answer hinges on liquidity, or how easily a household can turn a housing gain into cash it can spend today. A homeowner’s equity may swell on paper, but tapping it usually requires refinancing, selling, or taking on more debt, each of which involves costs, credit checks, and loan-to-value limits.
Building a model of three household types
Pyo built a computational model that simulates households making consumption and saving choices over time. Each simulated household falls into one of three categories: renters, single-homeowners (people who own the home they live in), and multi-homeowners (people who own at least one additional property beyond their primary residence). The model incorporates realistic frictions such as mortgage amortization, property taxes, down-payment requirements, transaction fees, and a cap on how much households can borrow against their homes.
Income, interest rates, and housing prices in the simulation fluctuate randomly, with drift parameters calibrated to Korean data. Renters pay rent tied to housing prices. Single-homeowners hold equity but cannot easily access it without transacting. Multi-homeowners collect rental income from their additional properties and can shuffle collateral across holdings.
When Pyo simulated the model across many possible price paths, a clear pattern emerged. Multi-homeowners showed consumption growth rising in step with housing price growth, with an estimated slope of about 0.11. Single-homeowners showed a slope near zero and slightly negative. Renters showed essentially no response at all.
The intuition offered by the model is that for renters, higher home prices mostly signal higher future rents and future purchase costs, which prompts them to hold back on spending. For single-homeowners, the equity gain is real but locked up, and higher future housing-related expenses push in the opposite direction of any wealth boost. Only multi-homeowners have both the rental income and the collateral flexibility to convert price gains into spendable resources.
Testing the prediction on Korean households
To see whether real households behaved the way the model predicted, Pyo turned to the National Survey of Tax and Benefit, a Korean household panel dataset running from 2013 to 2021. The dataset tracks income, consumption, assets, debts, and housing values at the household level. The final sample used for estimation included 3,489 observations.
Pyo estimated fixed-effects regressions that related changes in real consumption to changes in real housing prices, allowing the relationship to differ by homeownership type. He also included controls for income growth, age, household size, and employment, and used lagged values of financial assets and debts as instruments to address concerns about reverse causation between spending and balance sheet changes.
The baseline estimates lined up with the model. A one-percentage-point increase in housing price growth was associated with a 0.103-percentage-point increase in consumption growth among multi-homeowners, a statistically significant result. For single-homeowners the estimate was slightly negative (−0.016) and not statistically distinguishable from zero. For renters the estimate was −0.039, also statistically insignificant.
Slicing the data by debt, income, and assets
Pyo then examined whether these patterns held once households were further sorted by their financial condition. He ran three additional specifications that interacted housing price growth with debt-to-asset ratios, income terciles, and asset terciles.
Among households with high debt-to-asset ratios (above 70 percent), multi-homeowners still showed a positive response of 0.123, while high-debt single-homeowners actually showed a sharply negative reaction of −0.534. One reading, consistent with the model, is that highly indebted single-homeowners facing rising home prices anticipate greater future housing costs and pull back on current spending.
Income slicing produced a similar picture. High-income multi-homeowners showed the strongest response at 0.141. Low-income single-homeowners showed a significantly negative response of −0.261. Middle-income households of all types showed muted reactions.
The asset breakdown revealed one striking figure: multi-homeowners with relatively low measured assets showed a large positive response of 0.653, while asset-poor single-homeowners showed a large but statistically insignificant negative estimate. High-asset multi-homeowners showed a smaller but statistically significant 0.093. Renters, even wealthy ones, never showed a positive response.
What it means for housing-based stimulus
The practical takeaway concerns policy. Governments sometimes lean on housing markets to stimulate demand, either by supporting home prices, easing credit for homebuyers, or reducing transaction frictions. The implicit assumption is that rising home values will feed through into consumer spending across the broad population of homeowners.
Pyo’s results suggest that channel is narrow. If the households most responsive to housing wealth gains are multi-homeowners, and if that group is relatively small in the overall population, then the aggregate boost from a housing rally may be limited. As Pyo puts it, “When gains accrue primarily to unconstrained multi-homeowners — a relatively small group — housing-based stimulus is unlikely to generate strong aggregate demand.”
A few caveats are worth noting. The findings come from Korean household data covering 2013 to 2021, a period with its own particular credit regulations, tax rules, and housing market dynamics. Loan-to-value caps, property tax rates, and the prevalence of multi-property ownership differ across countries, so the magnitudes reported here may not translate directly to other settings. The analysis also relies on observational panel data with instrumental variables, so the estimated relationships are best read as associations disciplined by the model rather than the results of a controlled experiment.
