Correcting Information Asymmetry in Non-Standardized Construction Work: A Primary-Source-Based Extension of the Japan Construction Cost Database (JCCDB)

Algorithmic and AI-based cost-diagnostic systems can flag overcharging in residential and commercial construction when a reliable market-price reference exists for the work in question. However, for non-standardized or specialized work—items such as foot-bath spa installations or large studio mirrors—such systems have no native reference and tend to substitute the price of a superficially similar but fundamentally different residential small-item, producing systematic false-これpositive ("overcharge-suspected") judgments. This paper formalizes this failure mode, which we term phantom overcharge, and proposes a correction method that incorporates manufacturers' and installers' primary-source documents—wholesale quotations and dealer price lists—directly into the reference-price layer of a construction cost database. Building on the Japan Construction Cost Database (JCCDB; 65,729 items across 398 categories), we add six specialized reference categories, bringing the deployed price-reference layer from 170 to 176 categories. We present two validation cases—a foot-bath spa installation quoted at ¥4,400,000 and a dance-studio mirror installation—in which an initial automated judgment of "overcharge-suspected" reverses to "appropriate" once primary-source wholesale evidence is incorporated. We further show that the genuine cost concern in both cases lay not in the headline item price but in overhead ratios (12.5% and 17.4%). Our central finding is that the value of correcting information asymmetry is highest precisely where standardized price references are absent—the very region where conventional tools fail.