Could transaction-based financial benchmarks be susceptible to collusive behavior?

Lilian Muchimba

Research output: Contribution to journalArticlepeer-review

8 Downloads (Pure)


Prior to the series of manipulation scandals, financial benchmarks were perceived as a competitive and objective reflection of underlying money markets. For example, the manipulation of the London Interbank Offered Rate (LIBOR), underpinning financial contracts worth trillions of dollars was unthinkable. To prevent manipulation, financial market regulators around the world have recommended a paradigm shift from estimation-based to transaction-based financial benchmarks. This shift is based on the mainstream economic view that financial benchmarks anchored on actual transactions are not susceptible to anticompetitive behavior. However, unlike auction markets, underlying interbank money markets have unique features. As most activity takes place over the counter, they are opaque and are governed by conventions, trust, and reciprocity. This complicates the achievement of competitive pricing. Using a novel dataset from Bank of Zambia, this article makes an empirical investigation into transaction-based benchmarks’ susceptibility to anticompetitive behavior. Additionally, it contributes to the theoretical understanding of transaction-based financial market benchmarks. The study reflects on financial market regulators’ recommendation to transit from estimation-based to transaction-based financial market benchmarks. Further, the study is of interest to central bankers, as short-term interbank rates are the first stage of the monetary transmission mechanism.
Original languageEnglish
Pages (from-to)362-370
JournalJournal of Economic Issues
Issue number2
Publication statusPublished - 1 Jun 2022


  • Bank of Zambia
  • banks
  • collusion
  • monetary transmission mechanism
  • reference rates


Dive into the research topics of 'Could transaction-based financial benchmarks be susceptible to collusive behavior?'. Together they form a unique fingerprint.

Cite this