Underquoting and the Dual Role of Price Guides: Evidence from Housing Sales

Technical Summary: Underquoting and the Dual Role of Price Guides: Evidence from Housing Sales

Problem Statement
The paper addresses the persistent issue of underquoting in residential property markets, defined as advertising a property at a price below its plausible market value or expected transaction range. While regulatory reforms in Australia (specifically Victoria and New South Wales) have introduced disclosure requirements, underquoting remains a significant source of consumer complaints and enforcement challenges. A central difficulty in measuring underquoting is distinguishing between deliberate strategic distortion and legitimate valuation uncertainty or subsequent price discovery driven by competition. Existing measures, such as listing-to-sale price gaps, often conflate the initial listing decision with the effects of auction competition, bargaining, and market conditions. Furthermore, the economic function of a price guide may differ fundamentally between sales mechanisms: in auctions, guides may serve primarily to attract bidders, whereas in private-treaty sales, they may function more as signals of the seller's asking position and anchors for negotiation.

Methodology
The study employs a unified theoretical and empirical framework to examine the dual role of price guides across auction and private-treaty sales.

• Theoretical Framework: The authors develop a model where a seller chooses a price guide to balance two competing channels: a participation channel (lower guides attract more buyers/inspectors) and an information/bargaining channel (misleading guides attract mismatched buyers or create negotiation frictions). The model posits that the participation benefit is higher in auctions (where a reserve price protects the seller), while the information/bargaining cost is higher in private-treaty sales. This leads to the prediction that auction guides will be more downwardly distorted and less informative about final prices than private-treaty guides.
• Data: The empirical analysis utilizes data from the Melbourne housing market, a large, spatially heterogeneous market with active auction and private-treaty segments.
  • Training Sample (2023–2024): 65,768 detached-house transactions (20,250 auctions, 45,518 private-treaty) used to estimate valuation models.
  • Evaluation Sample (2025): 3,991 matched listings and transactions (1,336 auctions, 2,655 private-treaty) used for out-of-sample testing.
• Valuation Benchmarks: To avoid reliance on advertised guides or agent-selected comparables, the authors construct guide-independent valuation benchmarks using two approaches:
  1. Fixed-Effects Hedonic Model: A global log-linear OLS model relating transaction prices to structural (bedrooms, bathrooms, parking, land size), locational (distance to CBD, transport), and neighborhood (suburb median rent) characteristics.
  2. Geographically Weighted Regression (GWR): A spatially adaptive model allowing marginal effects of housing attributes to vary across local submarkets. This addresses the limitation of global models in capturing spatial heterogeneity.
  • Both models are estimated separately for auction and private-treaty segments.
  • Adjustment: To improve out-of-sample accuracy, the authors apply a real-time local adjustment based on the median prediction error of recent, comparable transactions (same suburb, similar bedroom count) prior to the sale date.
• Underquoting Measure: A conservative classification rule is applied. A transaction is classified as underquoted only if the realized sale price exceeds both the advertised upper guide and the lower bound of the model-implied valuation interval (derived from the GWR benchmark). This ensures the classification relies on independent valuation evidence rather than the listing-to-sale gap alone.

Key Results
The empirical findings strongly support the theoretical predictions regarding the dual role of price guides:

1. Asymmetry in Price Guide Accuracy:

   • Auctions: Only 26.27% of realized auction prices fall within the advertised guide range. Conversely, 65.72% sell above the upper guide, and only 8.01% sell below the lower guide. This indicates a strong upward asymmetry, suggesting guides are set significantly below final transaction prices.
   • Private-Treaty: 56.05% of prices fall within the advertised range, with a more balanced distribution of sales above (19.62%) and below (24.33%) the guide.
   • Model Comparison: The GWR-implied valuation ranges produce a much more balanced distribution of realized prices for auctions (42.44% within range) compared to the advertised guides, suggesting the advertised guides are systematically low rather than the valuation models being systematically high.

2. Guide Distortion Relative to Market Value:

   • Model-implied valuation bounds (both upper and lower) exceed advertised bounds significantly more often in auctions than in private-treaty sales.
   • Under the GWR benchmark, the model-implied upper bound exceeds the advertised upper guide in 77.69% of auction cases, compared to 56.95% for private-treaty sales.

3. Incidence of Underquoting:

   • Using the conservative GWR-based rule, 44.09% of auction transactions are classified as underquoted, compared to only 14.05% of private-treaty transactions.
   • This difference remains statistically significant (approx. 28.7 percentage points) after controlling for property characteristics, location, season, and land use in linear probability models.

4. Spatial Heterogeneity:

   • The GWR model significantly improves out-of-sample prediction accuracy compared to the global OLS benchmark. The median absolute error for auctions drops from ~$138,600 (OLS) to ~$72,300 (GWR), and for private-treaty sales from ~$90,200 to ~$60,300. This confirms the necessity of spatially adaptive benchmarks in a market like Melbourne.

Significance and Contributions
The paper makes three primary contributions to the literature on housing markets and underquoting:

1. Unified Theoretical Framework: It provides a unified explanation for why price guides differ in informativeness across sales mechanisms. It formalizes the "dual role" of guides as both participation devices (dominant in auctions) and information signals (dominant in private-treaty sales), explaining why underquoting is more prevalent and severe in auctions.
2. Comparative Empirical Evidence: It offers new evidence on the informativeness of price guides, demonstrating that auction guides are less closely related to expected transaction values and exhibit strong asymmetric errors compared to private-treaty guides.
3. Scalable Measurement Framework: It develops a scalable, model-based approach to measuring underquoting that reduces reliance on ex-post price gaps or small sets of agent-selected comparables. By combining advertised guides, realized prices, and out-of-sample valuation intervals estimated without using the guides themselves, the method provides a more robust benchmark for regulatory monitoring.

Limitations and Scope
The authors explicitly note that the study does not estimate the causal effect of specific disclosure regimes, nor does it observe internal auction data such as bidder numbers, reserve prices, or inspection counts. Consequently, the theoretical mechanisms are interpreted as economic explanations for observed empirical patterns rather than structurally identified causal effects. Similarly, the underquoting measure identifies transactions satisfying a conservative valuation-based criterion but does not establish the legal intent required for regulatory findings in individual cases. The findings are specific to the Melbourne market under current institutional conditions.