Environmental vetoes and incomparable retrofit packages: An ELECTRE-I-style investment analysis of an anonymised nonresidential building

A retrofit package with the lowest modelled carbon emissions need not be the preferred investment when cost and another pollutant move in opposite directions. This study develops an ELECTRE-I-style outranking analysis of five cumulative packages from one anonymised energy consultancy report. Total investment, annual CO2 and annual NOx inventories are minimised, normalised by their observed candidate ranges and compared using equal-weight concordance and a maximum-disadvantage discordance test. At a concordance threshold of two thirds and a discordance threshold of 0.5, the graph contains a three-edge chain among the first four packages, while the deepest package remains incomparable with every other option. The unique internally independent, out-dominating kernel is {P2, P4, P5}; the set of packages without incoming edges is instead {P4, P5}. Increasing discordance tolerance changes the kernel to {P1, P4} at 0.75 and {P4} at 1. A separate hypothetical eligibility rule prohibiting an increase in baseline NOx excludes P5 and leaves the base-case kernel {P2, P4}. These results demonstrate why pairwise vetoes, hard eligibility conditions and graph-based shortlists must not be conflated. The method provides a transparent account of unresolved choices, not an objective universal ranking. Its thresholds are analytical scenarios rather than elicited client preferences or statutory limits. The paper contributes a reproducible small-case decision audit with explicit scope, numerical checks and a clear distinction between retained alternatives and recommendations.