Identifying Contingency Liability from P3 Contracts Using Rule-Based NLP

ABSTRACT

The COVID-19 pandemic has increased contractual concerns under contingencies for public- private partnership (P3) projects. Conventional manual contract extraction is time-consuming and error-prone. Devising a method for automatic contract extraction can support contract management in this aspect. This research proposes a rule-based natural language processing (NLP) approach to extracting contingency liabilities allocated between the public sector and the private sector in the contract. The model consists of a domain-specific lexicon developed based on 21 US transportation P3 concession agreements and a set of matching rules to identify target sentences which fall into five classes, namely remedy entitlement, remedy obligation, liability waiver, mitigation, and termination. This automatic process can reduce the time and cost of the contract review process and help identify issues that the contracting parties should consider going forward in drafting new contracts or in amending existing contracts to avoid potential disputes, in response to consequences of contingencies, including the COVID-19 pandemic.