Paper: arXiv 2405.20715
Authors: Diabul Haque
Abstract
The Japanese real estate market, valued over 35 trillion USD, offers significant investment opportunities. Accurate rent and price forecasting could provide a substantial competitive edge. This paper explores using alternative data variables to predict real estate performance in 1100 Japanese municipalities. A comprehensive house price index was created, covering all municipalities from 2005 to the present, using a dataset of over 5 million transactions. This core dataset was enriched with economic factors spanning decades, allowing for price trajectory predictions. The findings show that alternative data variables can indeed forecast real estate performance effectively. Investment signals based on these variables yielded notable returns with low volatility. For example, the net migration ratio delivered an annualized return of 4.6% with a Sharpe ratio of 1.5. Taxable income growth and new dwellings ratio also performed well, with annualized returns of 4.1% (Sharpe ratio of 1.3) and 3.3% (Sharpe ratio of 0.9), respectively. When combined with transformer models to predict risk-adjusted returns 4 years in advance, the model achieved an R-squared score of 0.28, explaining nearly 30% of the variation in future municipality prices. These results highlight the potential of alternative data variables in real estate investment. They underscore the need for further research to identify more predictive factors. Nonetheless, the evidence suggests that such data can provide valuable insights into real estate price drivers, enabling more informed investment decisions in the Japanese market.
Complexity vs Empirical Score
- Math Complexity: 7.0/10
- Empirical Rigor: 6.0/10 (capped from 8.5: transaction costs not modelled; see penalty flags)
- Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 8.5)
Why this score: The paper uses advanced statistical methods (PCA, OLS regression, transformer models) and a large dataset with specific performance metrics (Sharpe ratios, R-squared), indicating high mathematical complexity and strong empirical backtesting.
Research Flowchart
flowchart TD A["Research Goal<br>Forecast Japanese Real Estate Prices"] --> B["Data Collection & Processing"] B --> C["Alternative Data Modeling"] B --> D["Transformer Model Prediction"] B --> E["Inputs:<br>5M+ Real Estate Transactions<br>Economic & Demographic Factors"] C --> F["Investment Signals<br>(Sharpe Ratio / Annualized Return)"] D --> G["4-Year Risk-Adjusted<br>Return Prediction"] F --> H["Key Findings<br>R²: 0.28 | Low Volatility Returns"] G --> H
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes using alternative data variables combined with transformer models to predict real estate performance and risk‑adjusted returns in Japanese municipalities.
Source sentence
transformer models to predict risk‑adjusted returns 4 years in advance, the model achieves an R‑squared score of 0.28, explaining nearly 30 percent of the variation in future municipality prices.
Assumes 2
- The relationship between property characteristics and prices remains constant over time. model
Source sentence
it assumes that the relationship between property characteristics and prices remains constant over time and requires a large, representative sample of transactions to produce reliable estimates.
- A large, representative sample of transactions is needed to produce reliable estimates. data
Source sentence
it assumes that the relationship between property characteristics and prices remains constant over time and requires a large, representative sample of transactions to produce reliable estimates.
Data
Japanese municipal real estate · other
- House price index (municipality level) market-prices unknown
Source sentence
A comprehensive house price index is constructed for all Japanese municipalities from 2005 to the present, utilizing a dataset containing over 5 million transactions.
- Population and migration dataset alternative public
Source sentence
The population and migration datasets are obtained from the SSDS. They start at 1996 and end at 2022, with yearly values.
- Taxable income dataset fundamentals public
Source sentence
The Taxable income dataset starts at 1985 and ends at 2021. It includes both the number of taxpayers and the total taxable income. It is collected by the Local Tax Bureau and provided by the Statistics Bureau through the SSDS.
Costs
Not modeled
Finds 5
Sharpe 1.50 · annual return 4.6%
- Net migration ratio delivered an annualized return of 4.6% with a Sharpe ratio of 1.5. performance
Source sentence
For instance, the net migration ratio delivered an annualized return of 4.6% with a Sharpe ratio of 1.5
- Taxable income growth yielded an annualized return of 4.1% and Sharpe ratio of 1.3. performance
Source sentence
... recorded annualized returns of 4.1% (Sharpe ratio of 1.3)
- New dwellings ratio produced an annualized return of 3.3% with a Sharpe ratio of 0.9. performance
Source sentence
new dwellings ratio ... recorded annualized returns of
- Transformer model predicting 4‑year‑ahead risk‑adjusted returns achieved R‑squared of 0.28 (≈30% variance explained). predictive-accuracy
Source sentence
the model achieves an R‑squared score of 0.28, explaining nearly 30 percent of the variation in future municipality prices
- Linear models for individual factors showed low R² scores, indicating limited explanatory power. statistical
Source sentence
low R2 scores which indicate significant noise in the dependent variable
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 24 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (20)
- Biases in Appraisal Land Price Information: The Case of Japan (2006) — link
- Aging and Property Prices: A Theory of Very-Long-Run Portfolio Choice and Its Predictions on Japanese Municipalities in the 2040s* (2017) — link
- Pricing Structure in Tokyo Metropolitan Land Markets and its Structural Changes: Pre-bubble, Bubble, and Post-bubble Periods (2007) — link
- Housing Prices in Tokyo: A Comparison of Hedonic and Repeat Sales Measures (2010) — link
- Testing for Granger non-causality in heterogeneous panels (2012) — link
- A Brief Review of House Price Forecasting Methods (2024) — link
- The Moody’s Analytics Case-Shiller Home Price Index Forecast Methodology (2013)
- Seven Sins of Quantitative Investing (2014)
and 12 more
- Immigration and Housing Rents in American Cities (2006) — link
- Immigration and House Prices in the UK (2011) — link
- The Long-Run Relationship between House Prices and Income : Evidence from Local Housing Markets (2003) — link
- Housing Supply and Housing Bubbles (2008) — link
- The Efficiency of Market for Single-Family Homes (1989)
- An Anatomy of Price Dynamics in Illiquid Markets: Analysis and Evidence from Local Housing Markets (2004) — link
- A Spatio-Temporal Model of House Prices in the US (2010) — link
- Attention Is All You Need (2017) — link
- Getting Ahead of the Market: How Big Data is Transforming RE (2018)
- Real Estate Transaction Price Information System (n.d.)
- Methodology of JRPPI: Japan Residential Property Price Index (2020)
- Report on Internal Migration in Japan (n.d.)
0 of 20 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.