Sovereign money: A challenge for science

Hans Gersbach 31 October 2018



The Chicago proposal, the Global Crisis, and in particular the recent vote on the sovereign money initiative in Switzerland have brought up a central question to the public.1 Is the current monetary architecture the best possible one or might other structures yield better results with respect to price stability, the stability of the financial system, or the level of GDP?

Though the Swiss initiative was clearly dismissed, the central questions remain unanswered, and thus require a comprehensive analysis by researchers. There exists an important literature on that subject (e.g. Bacchetta 2018, Benes and Kumhof 2012, Birchler and Rochet 2017), but various key issues are still unresolved.

The current system in a nutshell

To outline the main issues, it is important to recall that the current monetary architecture is built on four pillars (e.g. Faure and Gersbach 2016). 

  • The creation of money is organised hierarchically. The central bank issues banknotes and coins (physical central bank money) which serve as legal tender.
  • Commercial banks have the right to issue deposits when they grant loans or purchase assets. Deposits themselves are claims on banknotes.
  • Commercial banks face a set of rules such as capital requirements or deposit insurance schemes. However, they are not (or only to a small extent) required to hold central bank money as reserves for their deposits.
  • Interbank liabilities – for example, created through the millions of payments in the economy – are usually settled with reserves (electronic central bank money). Only commercial banks have access to reserves.

A sovereign money architecture

Given the four pillars that define the prevailing monetary architecture, a sovereign money architecture would demand that banks can only issue claims on banknotes if they have acquired the same amount of central bank money before, most likely in electronic form. This requirement can be stipulated as a 100% reserve requirement or as a direct requirement to first borrow central bank money before lending can take place. The central bank would lend central bank money to commercial banks through short-term or longer-term loans or through more sophisticated arrangements such as repo contracts. The central bank can allocate central bank money based on quantity restrictions, or it can set interest rates connected with collateral requirements to borrow central bank money. The central bank could also inject central bank money into the economy through financial markets, which could then be deposited at commercial banks and used for lending to firms or for buying securities. 

Hence, the scientific question is how does such a system perform and, in particular, how does it perform in comparison with the current architecture? The following three issues are thereby of particular importance. 

Three core issues

Issue 1: Under what circumstances do the two systems (current system and sovereign money system) yield the same outcomes? 

This question might be surprising, but it is fundamental. We can only comprehend the differences between these two systems if we understand the simplest setup in which both perform equally well. A first answer has been given in my earlier work (Faure and Gersbach 2016) in a simple framework with coexisting bank and bond financing. 

Equivalence proposition

If prices are flexible and there are no bank defaults and no financial frictions, and if the central bank only uses interest rate policy in both architectures, both systems yield the same outcomes.

There is a simple intuition for this result. If the central bank only uses the interest rate as a policy instrument, it does not matter whether a bank creates loans and deposits and later refinances its interbank liabilities – created in the payment process – at the central bank, or directly borrows central bank money in a sovereign money architecture and lends this money to the private sector. 

Issue 2: What are the core arguments for why the current system is strictly better than a sovereign money architecture?

Immediate arguments can answer this question. For instance, it seems obvious that a sovereign money architecture requires much more (and too much) information from the central bank – acting like a central planner for money creation. Moreover, it may prevent banks from acting when they are offered attractive lending opportunities but lack central bank money. Hence, the elasticity of bank lending in the real economy tends to be higher in the current system than in the sovereign money architecture. A thorough scientific analysis of these potential advantages is an important endeavour for a clear assessment of the potential merits of the current monetary architecture. 

Issue 3: Are there particular advantages of a sovereign money architecture in comparison with the current architecture?

The two most obvious potential advantages are greater financial stability, in particular the avoidance of bank runs by retail depositors, and higher seigniorage revenues for the government. Whether bank runs are avoided in a sovereign money architecture depends on whether deposits, as claims on central bank money, are kept on the balance sheet of commercial banks (Faure and Gersbach 2018). If deposits remain off the balance sheets, or are even kept as accounts at the central bank and guaranteed by it, runs are avoided. However, even if runs by retail customers on commercial banks are avoided, we may simply end up with runs from institutional investors, since commercial banks may end up financing themselves from these sources. Again, a clear understanding of financial stability concerns in a sovereign money architecture and an assessment which regulatory tools are adequate to deal with them would be highly welcome. The same applies for the potential to generate higher contributions from money creation to government revenues. Striking an appropriate balance between building up reserves at the central bank and distributing money to the government also remains relevant for a sovereign money architecture. 


The Chicago proposal, the Global Crisis, and the Swiss sovereign money initiative have challenged policymakers to examine alternative monetary architectures. These events have challenged economists to at least the same extent. 


Bacchetta, P (2018), “The sovereign money initiative in Switzerland: An economic assessment”, Swiss Journal of Economics and Statistics 154(3):1-16.

Benes, J. and M Kumhof (2012), “The Chicago plan revisited”, IMF working paper 12/202. 

Birchler, U and J-C Rochet (2017), “Die Vollgeld-Initiative – ein Leitfaden für jedermann”, Institut für Banking und Finance, University of Zurich.

Faure, S and H Gersbach (2016), “On the money creation approach to banking”, CEPR Discussion Paper 11368.

Faure, S and H Gersbach (2018), “Money creation in different architectures”, CEPR Discussion Paper 13156.

Gersbach, H (2018), “Vollgeld – Eine Herausforderung für die Wissenschaft”, ifo Schnelldienst 71(16): 15-16.


[1] For a discussion of the key issues of sovereign money, see Gersbach (2018, German only)



Professor at CER-ETH - Center of Economic Research at ETH Zurich and CEPR Research Fellow


CEPR Policy Research