The COVID-19 pandemic significantly impacted global cities, leading to speculation about their potential demise. Initial reports showed temporary population decreases and increased real estate transactions outside major urban areas in several countries. Contributing factors included higher infection rates in dense areas, anxieties about crowded spaces, business closures, and the shift to remote work. While some studies suggest a temporary urban exodus that reversed after restrictions eased, others indicate increased outmigration in certain contexts, such as from Tokyo's city center to suburbs and from inner cities in Stockholm and other cities. However, research largely focuses on before-and-after comparisons of residential preferences, neglecting the mid- to long-term impacts of the pandemic on perceptions of city living. This study addresses this gap by investigating whether the pandemic and increased remote work have altered residential preferences in Tokyo, a city that did not implement strict lockdowns, using a randomized conjoint experiment to analyze residents' preferences for hypothetical residential profiles with varied attributes, including reminders of potential pandemic re-occurrence and remote work options.

Existing research on the pandemic's impact on urban areas offers conflicting perspectives. Some scholars argue that cities, historically resilient, will emerge stronger, with studies showing a reversal of outmigration trends after restrictions lifted. Conversely, other research shows increased outmigration from city centers to suburbs and rural areas in various cities globally, including Tokyo, where migration from the center to the suburbs increased post-pandemic. A study in Japan found that net inter- prefectural migration was negative in major prefectures like Tokyo and Aichi in 2020. Studies in Australia, Germany, and Sweden also report similar trends. Most existing studies primarily focus on comparing changes in residential preferences during and before the pandemic, for example, noting increased preference for detached houses with gardens among households with children in Tokyo and Osaka due to increased home-working during the pandemic. Another study in Poland showed a shift in preferences from owning land outside the city to prioritizing proximity to green spaces. A Korean survey highlighted concerns about urban living among Seoul residents, though commuters did not show a strong desire to relocate. This study builds upon existing literature by going beyond simple before-and-after comparisons using a conjoint experiment to understand the long-term impacts.

This study employed a conjoint experiment to examine residential preferences in Tokyo Metropolis (excluding island regions), differentiating it from the broader Tokyo Metropolitan Area. The experiment involved a randomized controlled trial (RCT) with a 2x2 matrix design (Reminder and Remote treatments). 1333 valid responses were obtained from a survey conducted using Rakuten Insight's online platform. Participants were randomly assigned to one of four groups: Control-Control, Reminder-Control, Control-Remote, and Reminder-Remote. Each group completed seven sets of conjoint experiments, choosing between two hypothetical residential profiles with randomized attributes and the option of staying in their current residence. The 'Reminder' treatment involved a paragraph highlighting the potential for future pandemics, while the 'Remote' treatment offered a scenario with extensive remote work options. Attributes included neighborhood type (inner city, within 23 wards, outside 23 wards but within Tokyo, outside Tokyo), housing type (single/semi-detached, terrace house, apartment/mansion), public transport access, amenity access, green space, commute mode, community support, and noise levels. The Average Marginal Component Effects (AMCEs) were analyzed to determine the influence of each attribute on residential choice, considering free-choice and force-choice scenarios and the impact of the RCT groups. Heterogeneity analysis examined the effects of income, age, and self-reported mental health status on preferences.

The study revealed a 40% preference to relocate from current residences. Despite reminders of potential future pandemics and the remote work option, participants generally preferred living in Tokyo, particularly in the inner city wards and areas within the 23 wards. The reminder treatment generally did not influence preferences significantly compared to the control group. However, the remote work option led to a statistically significant increase (3.8%) in the intention to relocate. Regarding specific attributes, terrace housing was consistently unpopular, while apartment/mansion housing was slightly preferred. Inner city wards were the most desirable, with a decreased probability of selecting residences outside the 23 wards, especially among those with freedom to choose their residence. Those who received a pandemic reminder had a stronger aversion to living outside the 23 wards and preferred quieter residences. The combination of remote work and pandemic reminder produced similar results, with slightly reduced aversion to locations outside the 23 wards. Analysis of socio-demographic factors revealed: * Higher-income groups showed more reluctance to relocate due to remote work options; with pandemic reminders, they preferred quieter residences. * Younger individuals (20-39) were more reluctant to live outside the inner wards, preferring better amenities and community support. * Individuals with worsened mental health post-pandemic displayed neutral preferences towards relocating outside Tokyo and greater concern over noise levels. * Low-income individuals in the Control-Control group strongly preferred public transportation access, this preference diminishes when a reminder is given or a remote working option is available.

The findings suggest that, unlike some other global cities experiencing significant central area decline, Tokyo's inner city may be unaffected by the pandemic and remote work trends. This might be attributed to differences in policies and perceptions of remote work, lower remote work prevalence in Japan, and cultural preference for face-to-face communication. The lack of preference for public transport and terrace housing aligns with some findings but contradicts others, highlighting context-specific factors. The reminder of potential pandemic re-occurrence heightened aversion to areas outside the 23 wards and preferences for quieter environments. Socio-demographic differences in housing preferences show varied responses to public health concerns and employment changes, reflecting the complexities of urban dynamics. The preference for inner-city living among younger people counteracts concerns about widespread urban decline.

This study demonstrates that the allure of Tokyo, particularly its inner city and 23 wards, remains strong even after the pandemic, indicating a low likelihood of large-scale outmigration. The findings counter concerns of significant urban decline, suggesting the need for strategic compact urban development to promote sustainability and resilience. However, continued monitoring using updated population data is crucial to confirm these findings and assess the long-term impacts of the pandemic and remote work on urban dynamics. Future research should explore the specific factors driving preferences across different demographic groups and examine the long-term implications of shifting work patterns.

The study relies on stated preferences from a survey, which may not fully reflect actual behavior. The hypothetical scenarios presented in the conjoint experiment might not perfectly represent real-world choices. The study is limited to Tokyo, and findings may not be generalizable to other cities. Although the sample size is relatively large and attempts to control biases were made, the reliance on online panel data might introduce sampling bias.