Evaluating information asymmetry effects on hotel pricing: a comparative analysis before and during the COVID-19 pandemic in the Taiwan's market

The study examines how information asymmetry between hoteliers and consumers affects hotel lodging prices and how this changed with COVID-19. Price discrimination and dispersion are common in hospitality due to non-storability, inelastic supply, and multi-channel distribution, and information asymmetry can impose an “information tax” on both buyers and sellers. COVID-19 disrupted tourism worldwide, but Taiwan maintained relatively normal domestic travel in 2019–2020 due to strict border controls, making it a suitable context to compare pre- and post-outbreak dynamics. The authors propose using a two-tier stochastic frontier approach to quantify the relative information mastery of buyers and sellers embedded in transaction prices, contributing a quantitative lens on information asymmetry in hotel pricing. The study situates its inquiry within broader digitalization trends in tourism and Taiwan’s ongoing transition toward integrated ICT systems.

Two streams are reviewed. (1) Economic modeling of hotel pricing: Hedonic pricing studies link room rates to hotel attributes (e.g., star rating, reputation, location, seasonality) and show strong effects of category and online ratings on WTP and prices. Revenue management pricing highlights fairness and transparency effects on WTP, and game-theoretic models analyze interactions among hotels, OTAs, and regulators. Additional determinants include seasonality, location premiums, length of stay, and star ratings. (2) Information asymmetry in hotel pricing: Asymmetric information induces price dispersion and can reduce performance. Tourists’ bargaining often seeks value for money, and electronic-era factors (ratings, eWOM, review volume, reply rates, quality certifications) mitigate uncertainty and asymmetry. The paper proposes a conceptual framework for bargaining dynamics: information availability shapes bargaining intention, power, and outcomes, which in turn influence hotel prices, with determinants grouped into hotelier-related, consumer-related, and market-related factors.

The study adopts a two-tier stochastic frontier model (SFA2tier) to decompose opposing informational effects from hoteliers (who can raise prices by extracting consumer surplus) and consumers (who can lower prices by extracting hotelier surplus). Let the fair price conditional on characteristics x be μ(x). The observed price P is modeled as P = μ(x) + ε, with ε = w − u + v, where w ≥ 0 is the surplus extracted by hoteliers (raising P), u ≥ 0 is the surplus extracted by consumers (lowering P), and v is a symmetric noise term (v ~ N(0, σv²)). Hoteliers’ and consumers’ informational effects are linked to expected surpluses relative to the fair benchmark, summarized by E(1 − e^{-ε}) terms for each side, and a net surplus (NS) capturing overall asymmetry on price. Distributional assumptions: w ~ Exp(σw), u ~ Exp(σu), independent of v. Parameters are estimated by Maximum Likelihood. The deterministic component μ(x) uses a Translog specification: μ(x) = Σ βi ln xi + 0.5 ΣΣ βij ln xi ln xj, where xi = {log(Room), log(Labor_M), log(Labor_R)}, allowing for interactions among capacity and staffing variables. Estimation employs STATA 15 and the SFA2tier routine. Data comprise monthly observations for tourist hotels in Taiwan from July 2019 to November 2020, drawn from the Taiwan Tourism Statistics Database, yielding 2032 observations after cleaning. Descriptive statistics include mean lodging price 3527.4 NTD (SD 2520.0), mean rooms 623.6 (SD 1899.6), managerial staff 34.0 (SD 36.8), and room department staff 59.0 (SD 44.6). Regional average prices generally declined during COVID-19, except in Tainan and the East region.

- Model fit and coefficients: Compared with OLS, the two-tier SFA captures asymmetric informational effects. Significant coefficients in the frontier include log(Room) positive (0.066, p<0.05), log(Labor_R) negative (−0.324, p<0.01), interaction log(Room)×log(Labor_M) negative (−0.023, p<0.05), and log(Labor_M)×log(Labor_R) positive (0.075, p<0.1). Variance parameters indicate identifiable one-sided effects for both agents under the exponential–normal composite error structure. - Magnitudes of informational surpluses (Table 3, total sample): Hoteliers’ surplus averages +30.31% relative to the fair benchmark, while consumers’ information reduces prices by 14.51%, yielding a net surplus of +15.81%. Interpreted in levels, a room with a fair price of 100 NTD would transact at approximately 115.81 NTD due to information asymmetry. - Pre- vs post-COVID comparison: Before COVID-19, net surplus averaged 17.52%; during COVID-19 it decreased to 14.84%, indicating reduced information asymmetry and weaker hotelier advantage, with a corresponding strengthening of consumer informational position. Hoteliers’ surplus declined from 31.43% (pre) to 29.69% (during), while consumer surplus increased from 13.90% to 14.85%. - Distributional perspective: At Q1 the net surplus is approximately 0.59% (near parity), widening at median and Q3, with the largest disparities at Q3. Post-epidemic, the distribution of hotelier surplus shifts left (lower informational extraction), while consumer surplus shifts right (greater informational extraction). - Regional heterogeneity (Table 4): Net surplus decreased in major metropolitan areas post-outbreak, e.g., Taipei from 16.68% to 10.09%, Kaohsiung from 4.32% to 1.86%, Taichung from 4.27% to 3.16%, Tainan from 16.13% to 15.61%, and Taoyuan from 11.95% to 11.22%. In contrast, the East region increased from 8.49% to 12.38% and Scenic areas from 39.46% to 41.65%. - Aggregate effect measure: The overall impact of information asymmetry on final prices is positive (0.6954), indicating a systematic upward pressure on prices relative to the fair benchmark. - Market patterns: Descriptive evidence shows declines in average prices in most regions during COVID-19 (notably Taipei and Kaohsiung), with relatively stable or increasing prices in Tainan and the East region.

The study’s central question concerns how information asymmetry between hoteliers and consumers influences hotel prices and how this changed during COVID-19. The two-tier SFA decomposition shows that hotels typically possess greater informational advantage, raising prices above a fair benchmark. However, during COVID-19, this advantage narrowed, especially in metropolitan areas, implying a movement toward more symmetric information and fairer pricing. Two mechanisms likely drive this shift: (1) a change in the composition of demand—fewer overseas visitors (who possess relatively less local price information) and a higher share of domestic travelers with better information and stronger bargaining power; and (2) supply-side adjustments—hoteliers increasing transparency, promotions, and information disclosure to stimulate demand and comply with heightened public health oversight, thereby reducing the scope for price discrimination. These mechanisms align with observed reductions in net surplus in urban centers and the distributional shift toward higher consumer surplus and lower hotelier surplus post-outbreak. The findings underscore the role of market structure, consumer mix, and information policies in moderating price-setting power and price dispersion in hospitality.

This paper quantifies information asymmetry in Taiwan’s hotel market using a two-tier stochastic frontier model and micro data from 2019–2020, comparing pre- and post-COVID periods. Contributions include: (1) providing an empirical measure of bilateral informational surpluses embedded in transaction prices; (2) demonstrating that hotels generally hold greater bargaining power, lifting prices above fair benchmarks; (3) documenting a post-COVID reduction in information asymmetry, driven by shifts in consumer composition and increased transparency; and (4) revealing regional heterogeneity with greater convergence toward fairness in metropolitan areas. Practical implications: Hoteliers should enhance transparency, communicate pricing logic, and leverage eWOM and quality signals to build reputation and long-term competitiveness. Policymakers can strengthen market supervision, reduce transaction frictions, incentivize transparency (e.g., tax benefits), offer targeted subsidies to weaker-bargaining consumers, and expand public information channels to mitigate asymmetry-related welfare losses. Future research: Extend to panel data frameworks to capture intertemporal dynamics in information mastery, explore additional digital trace variables (ratings, review volumes, response behaviors), and test alternative distributional assumptions or identification strategies within SFA2tier to validate robustness across markets and shocks.

Methodological and data constraints include: potential MLE convergence sensitivity to initial values and sample size in two-tier frontier estimation; reliance on repeated cross-sections rather than true panel data due to quasi-fixed inputs and aggregation in tourism statistics, limiting the ability to estimate fixed or random effects; and possible measurement errors in aggregated monthly indicators. Future work should employ panel datasets and dynamic specifications (e.g., Polachek and Yoon, 1996; Lu et al., 2011) to examine intertemporal changes in buyer–seller information mastery and heterogeneity across segments and channels.