Central bank digital currencies in open economies: A new Keynesian DSGE approach

Abstract

Central bank digital currencies (CBDCs) are admittedly a hot topic worldwide, and rightfully so as they potentially embody the innovative payment rails of tomorrow’s monetary system, thereby helping in making money and payments fit for decades to come. Globally, over 90 percent of central banks are cautiously exploring such financial innovation alongside all stakeholders and a dozen guardians of stability have already spearheaded the (r)evolution. Such endeavors are often motivated by public good objectives, from safeguarding public trust in money, ensuring its safe access and financial inclusion to strengthening financial stability and monetary policy altogether. While its wide-ranging impacts are certainly far from being understood, a consensus is growing that the financial stability risks stemming from bank disintermediation are manageable. However, research proves to be far thinner regarding the implications of CBDCs once available to foreign citizens. This work aims to investigate such intricacies in open economies based on New Keynesian DSGE models. We contribute to the exponentially growing literature in several ways. At first, we provide the lateststance of research concerning the financial and monetary implications of the digital monies. Secondly, we develop tractable models including CBDCs in its most simplistic form. Our two-economy model proves to be most appropriate to study the case of Sweden, with its world oldest central bank expected to lead the change in the medium term.In line with existing literature, we found the existence of real and financial international spillovers,most apparent in light of CBDCs. Unsurprisingly, we also unravelled that its demand is left unchanged when offered at a lower remuneration rate than bonds. At last, we conclude that our simplistic inclusion of CBDCs is likely to have limited impact on the conduct and transmission of monetary policy, in accordance with initial central banks’ efforts.Our results depend crucially on the models’ underlying assumptions, and in particular in regard to the CBDC parameters upon which there is currently little benchmark, and which heavily influence the responses. Numerous avenues for future work are also outlined