How COVID-19 Reshaped the Hidden Network Linking China’s Housing Markets

When a pandemic sweeps across a country, its economic shockwaves rarely stay confined to the place where the first case appears. Housing markets, in particular, are stitched together by invisible threads: commuters, investors, migration flows, and the collective mood of millions of buyers and sellers. A new study published in Humanities and Social Sciences Communications has now mapped those threads with unusual precision, tracking how the COVID-19 outbreak altered the way housing price volatility travels between 38 major Chinese cities. The findings reveal a market that briefly fell silent, then reorganized itself in ways that carry important lessons for financial regulators everywhere.

The research, conducted by Tingguo Zheng of Xiamen University’s Center for Macroeconomic Research, Mo Zhou of Jiangxi University of Finance and Economics, and Shiqi Ye of the Chinese Academy of Sciences, tackles a question that has haunted economists since early 2020: does a public health crisis make housing markets more interconnected and therefore more fragile, or does it sever the links that normally transmit price swings across regions? The answer, it turns out, is neither simple nor permanent. The pandemic’s effect on inter-city connectedness started out negative but gradually dissipated and turned positive after roughly 40 weeks, while the level of cross-regional spillover of housing price volatility dropped significantly during the shock.

To reach these conclusions, the team built one of the more technically ambitious datasets in the housing economics literature. They used second-hand housing price indices for 38 major Chinese cities, capturing the resale market where prices respond quickly to changing expectations rather than being smoothed by long development cycles. Second-hand markets are often considered the pulse of urban real estate because they reflect the real-time bargaining of individual households rather than the slow, regulated pricing of new construction. By feeding these weekly indices into a large time-varying parameter vector autoregression model, the researchers could quantify how much of the volatility in one city’s housing market is explained by shocks originating elsewhere.

The vector autoregression framework, or VAR, is a workhorse of macroeconomics: it treats every variable in a system as both a cause and a consequence of every other variable, estimating how shocks propagate over time. The twist in this study is the time-varying parameter component. Instead of assuming that the relationships between cities remain constant across the sample, the model allows the strength of each pairwise connection to evolve week by week. This flexibility matters enormously when the system is hit by something as disruptive as a pandemic, because the entire architecture of the network may shift. A model frozen at pre-pandemic parameters would miss precisely the dynamics that matter most.

From the model’s output, the researchers constructed a time-varying complex network of housing price volatility spillovers. In this network, each city is a node, and a directed edge runs from one city to another when volatility in the first city transmits measurable pressure to the second. The total connectedness of the network, a standard measure borrowed from financial contagion research, summarizes how tightly coupled the whole system is at any point in time. When connectedness is high, a disturbance in one market can cascade widely; when it is low, markets absorb shocks locally. This framing treats China’s housing market not as a collection of isolated cities but as a single, evolving ecosystem, much like the networks used to monitor systemic risk in banking.

The headline result is striking. In the immediate aftermath of the pandemic shock, overall connectedness among the 38 cities fell, suggesting that lockdowns, travel restrictions, and widespread uncertainty temporarily decoupled regional housing markets from one another. Buyers stopped moving, transactions froze, and the channels through which price signals normally traveled went quiet. But this isolation did not last. As the pandemic wore on, the negative effect gradually dissipated, and after approximately 40 weeks the net effect on overall connectedness turned positive. In other words, once activity resumed, the housing markets of China’s major cities ended up more tightly linked than before the crisis, a pattern consistent with the idea that shared macroeconomic conditions and coordinated policy responses can bind markets together in the recovery phase.

Meanwhile, a second measure told a different story. The cross-regional spillover level of housing price volatility, which captures how much volatility leaks across regional boundaries, dropped significantly under the pandemic shock. This apparent paradox, lower spillover alongside eventually higher connectedness, makes sense when one considers the mechanics of the disruption. During the acute phase, reduced urban population mobility meant fewer cross-city property purchases, fewer investors reallocating capital between regions, and fewer channels for local sentiment to spread. Volatility became more contained within each city even as the long-run architecture of the network was being rewired.

The mechanism analysis is where the study becomes particularly relevant beyond China. The authors find that the impact of COVID-19 on city housing price volatility operated mainly through two channels: urban population mobility and the psychological and behavioral factors of market participants. Population mobility is the physical substrate of housing market integration. When people can move freely, they carry demand from one city to another, search for property remotely, and respond to price differences across space. Mobility restrictions cut those channels abruptly. Behavioral factors add the emotional layer: fear, uncertainty, and herding behavior can amplify or suppress price responses in ways that pure fundamentals cannot explain. Search-attention indices, which track how intensely people in different cities are looking for housing information online, provide a window into this collective psychology.

These findings arrive at a moment when policymakers are still digesting the lessons of the pandemic for financial stability. Real estate is the largest asset class held by Chinese households, and housing-related lending is a cornerstone of the banking system. A tightly connected housing network is not inherently dangerous, but it creates the conditions for synchronized downturns: if many cities’ markets move together, a national correction becomes more likely, and the tools available to local governments become less effective. The authors frame their work as a contribution to guarding against systemic risks, arguing that monitoring the connectedness structure of housing markets should become a routine part of risk surveillance during major public health events and comparable shocks.

The study also demonstrates a methodological template that other countries can adopt. Weekly housing indices, mobility data from telecommunications or transportation sources, and online search metrics are now available in most economies, and time-varying network models can be estimated with them at relatively low cost. The Chinese experience suggests that the first weeks of a crisis are when markets decouple and when conventional contagion models are most likely to fail, while the recovery phase is when hidden reconnections form. Regulators who watch only aggregate price levels may miss both phases; those who watch the network itself gain an early-warning system.

There are, of course, limits to what any single-country study can establish. China’s housing market is shaped by distinctive institutions, including local government land finance, capital controls, and strong policy intervention, so the exact timing and magnitude of the effects may differ elsewhere. The authors relied exclusively on pre-existing, aggregate city-level secondary data, which protects privacy but cannot capture household-level heterogeneity in how people responded to the crisis. Still, the core insight travels well: major public health events do not merely depress housing prices, they temporarily restructure the relationships between markets, and those structural changes follow a predictable arc from disruption to reconnection.

As the world prepares for the possibility of future pandemics, the study offers a concrete warning wrapped in a technical package. The housing market’s response to COVID-19 was not a simple dip and rebound but a dynamic reorganization, with connectedness dipping, then overshooting to a higher level within about 40 weeks of the shock. Understanding that arc, and the mobility and behavioral forces that drive it, may prove essential for the next time a public health emergency tests the stability of the places where most families keep most of their wealth.

Subject of Research: The impact of the COVID-19 pandemic on inter-city housing price volatility spillovers and connectedness among 38 major Chinese cities

Article Title: Impact of major public health events on inter-city connectedness and cross-regional spillover of housing prices

Article References: Zheng, T., Zhou, M., & Ye, S. (2026). Impact of major public health events on inter-city connectedness and cross-regional spillover of housing prices. Humanities and Social Sciences Communications. https://doi.org/10.1057/s41599-026-08901-6

Image Credits: AI Generated

DOI: 10.1057/s41599-026-08901-6

Keywords: COVID-19, housing prices, spillover effects, complex networks, vector autoregression, China real estate, systemic risk, population mobility, behavioral economics, market connectedness, public health events, second-hand housing