The raison d'être of central banks is to safeguard the stability of the monetary system. That is, to protect the value of money and to ensure the smooth operation of the payment system, the means by which we transfer money. Throughout history, Germany and German ingenuity have played no small part in reshaping this system. The Deutschmark was an anchor of monetary stability. And inventions by German engineers in the 1960s paved the way for today's smartcards.

While we have seen bursts of innovation to the monetary system before, this time feels different. Innovation is rampant and entrepreneurs are trying to improve not only the way we pay but also money itself. With a noticeable reduction in the use of cash in some countries, interest in and experimentation with digital currencies is growing. As the debate intensifies, some suggest that central banks should issue them sooner rather than later. Yet, as the monetary system is the backbone of the financial system and wider economy, we need to understand the full consequences of any new developments. One does not rip out the old infrastructure before knowing how the new mode of transport works in all circumstances. Adopting untried technology that ultimately proves unreliable could seriously endanger public trust in the currency.

Central banks are leading from the front when it comes to the future of the monetary system. In payments, central banks are embracing new technologies to meet the challenges of today and tomorrow. Key trends are the need for speed and globalisation, while still maintaining reliability.

A delay of a day or more used to be acceptable in payments; today, it seems like an eternity. On their mobile phones, consumers now get instant satisfaction or feedback from email and social media. Accordingly, mobile payments solutions and faster systems for retail payments are emerging in response to demand. They generally allow people to receive funds within seconds, anytime and anywhere. One example is the Eurosystem's new TARGET Instant Payment Settlement service, or TIPS. The system allows retail payment service providers to offer funds transfers in real time around the clock. For large-value and high-priority payments, so-called real-time gross settlement systems have been speeding up transfers since the 1980s. The Bundesbank's Elektronische Abrechnung Frankfurt, or Electronic Access Frankfurt, was an early version of these systems.

Globalisation is increasing the demand for better and faster cross-border payments. Most of these flows still rely on an intricate web of bilateral relationships between commercial banks, known as correspondent banking. But multicurrency systems are emerging. They provide settlement of more than one currency within one jurisdiction or across multiple jurisdictions.

Turning to money, economics textbooks say that money is what money does. Money serves as a means of payment, a unit of account and a store of value. But not all monies are equal. Cash is public money, issued by the central bank. But when I pay for my groceries with a credit or debit card and, increasingly, with my mobile phone, I use private money. That is, the liabilities of a commercial bank, and increasingly in some countries, the phone company or a big tech firm.

This distinction is important for the current debate about a future monetary system with central bank digital currencies. This debate is not really about convenience and digitalisation. Convenient digital payment options already abound. Many people already use one or more cashless payment systems through their mobile phones, paying with private money. Indeed, this is the cutting edge of innovation.

The debate is partly about the potential decline in the use of cash, and what central banks should do about it. In some countries, demand for cash has dropped as consumers and retailers have embraced electronic means. Two examples are Sweden and Denmark, where stores and restaurants are increasingly reluctant to accept paper money. Instant mobile payment solutions are gaining ground rapidly. But this can make it harder for some groups in society to pay.

But for most countries, cash is still in high demand. The amount of cash in circulation has actually increased over the last decade in tandem with the volume of electronic payments. In Germany, nearly half of goods and services are still paid for in cash, according to a Bundesbank study, although cards are now slightly more popular for everyday purchases. Consumers trust cash and appreciate its simplicity. They find it useful to educate children about how to handle money, they like the way it allows them to control their finances and they worry that the elderly, in particular, may struggle in a world without cash. In the short term, there is no urgency to come up with a substitute for cash. Things may change in the future, however, and central banks want to be prepared.

A report last year by two of the central bank committees based at the BIS identified two main varieties of central bank digital currencies. A wholesale variety would be restricted to financial institutions and used for interbank payments and other settlement transactions. A retail variety would be accessible to everyone. Both types can be based either on digital tokens or on accounts.

A retail digital currency might mean that anyone could open a bank account directly with the Deutsche Bundesbank, as part of the Eurosystem. This would let ordinary people and businesses make payments electronically using central bank money. Or they could deposit money directly in the central bank, and use debit cards issued by the central bank itself.

In terms of technology, it would be easier to replicate the attributes of cash - if so desired - with digital tokens than with accounts. Tokens have a long monetary history, from stones and shells through to cigarettes. But the technology behind digital tokens is still broadly untested, whereas the technology for an account-based central bank digital currency has been available for decades. So far, central banks have generally chosen not to provide such accounts. Why not? The answer lies in one of the other major issues underlying the debate: the impact on the financial system.

The current system has two tiers. The customer-facing banking system is one tier, and the central bank is the other. The two tiers work together. If a customer buys something from a shop, the bank debits the customer's account and credits the shop's account. When the customer and the shop have different banks, the two banks settle the payment through the central bank. The central bank debits the account of the customer's bank and credits the account of the shop's bank.

At the moment, these settlement accounts are the only form of electronic central bank money. Only commercial banks have access to them. The debate is whether to widen access beyond the current circle of commercial banks. There are several aspects here.

Banks play an important role as provider of financial services to citizens and businesses. Imagine that the Bundesbank (or, for that matter, the ECB) were to offer deposit accounts to everyone and then issue debit cards and mobile phone apps to make payments with. In such a scenario, the central bank would be taking on the customer-facing business, opening up a new line of work.

Safety could be an important reason to deposit money in the central bank. In times of uncertainty, more customers would prefer to have deposit accounts at central banks, and fewer at commercial banks. A shift of funds from commercial banks to the central bank could be gradual at first. But the trickle could turn into a flood.

If bank deposits shift to the central bank, lending would need to shift as well. So, in addition to the deposit and payments businesses, the central bank would be taking on the lending business. The central bank would need to meet business owners, interview them about why they need a loan, and decide on how much each should receive. Another new business line.

Is this the kind of financial system that we would like to have? Perhaps this thought experiment has gone too far. For instance, the central bank could make do without a lending operation if it sends customer deposits to the commercial banks by opening central bank accounts at commercial banks. In effect, the central bank lends to commercial banks so that they could lend on to customers.

However, the bigger issue has to do with the division of labour between commercial and central banks. The central bank is a public institution charged with ensuring that inflation is under control, the economy runs smoothly and the financial system is sound. Commercial banks are private businesses that thrive by attracting and serving customers. Making profits is a key motivation. They tend to have more staff than central banks because serving customers is resource-intensive.

There are historical instances of one-tier systems where the central bank did everything. In the socialist economies before the fall of the Berlin Wall, the central bank was also the commercial bank. But we cannot hold up that system as an example of better customer service. Less dramatically, publicly owned banks in many economies are hardly paragons of efficient allocation of funds or of good service.

A token-based digital currency may be less prone to this structural shift from the commercial banking sector, as the outstanding amount of currency can be fixed. However, there would then be the question of whether these tokens would command a premium over bank deposits. Would such a premium fluctuate over time with shifts in uncertainty and financial conditions? Offering higher interest rates on commercial bank deposits may be enough to hold funds during quiet times. But it's uncertain whether it would work during periods of tumult and the inevitable "flight to safety", as I mentioned earlier.

We know from historical experience that in times of financial stress, money moves away from banks that are perceived as risky towards banks that are perceived as safer. Money flows from privately owned banks to publicly owned ones, in emerging markets from domestically owned banks to foreign-owned ones, and generally from weakly capitalised banks to strongly capitalised ones. In such scenarios, imagine that depositors also have the choice of putting their money in a digital currency of the central bank or in a central bank deposit account directly. A premium could open up, where one euro of deposits in the commercial bank buys less than one euro's worth of central bank digital currency.

Last but definitely not least, the introduction of central bank digital currencies would change the environment in which central banks conduct monetary policy. This is, of course, the main tool that central banks have to influence the economy. The basic mechanics would stay the same: the central bank would still use its balance sheet to control short-term interest rates. But digital currencies would change the demand for base money and its composition in unpredictable ways. They might also modify the sensitivity of the demand for money to changes in interest rates.

Furthermore, if a digital currency is in demand, it would mean a larger central bank balance sheet. This may require the central bank to hold extra assets such as government securities, loans to commercial banks or international reserves. Purchasing these could interfere with the functioning of key markets or dry up liquidity. At least in a transitional period, these changes have the potential to completely up-end the way that monetary policy affects the economy. This is not something that central banks take lightly.

The Committee on Payments and Markets Infrastructures (CPMI) at the BIS last year surveyed central banks to take stock of current work and thinking on issuing digital currencies. More than 60 central banks participated, representing countries covering 80% of the world's population. Seventy per cent of central banks are working on this issue. Most are looking at both retail and wholesale varieties of digital currencies.

But only about half have moved on to the next stage of actively testing the idea. These central banks are examining the benefits, risks and challenges of potential issuance from a conceptual perspective. Only a couple are experimenting with the different possible technologies, in "proofs of concept" or pilot projects.

If we go one step further and ask central banks whether they plan to issue a digital currency, the picture is quite telling. Very few central banks think it is likely that they will go there in the short to medium term, be it retail or wholesale. Research and experimentation have so far failed to put forward a convincing case. In sum, central banks don't see today the value of venturing into uncharted territory.

 As the saying goes: first weigh up, then weigh in. Central banks are proceeding cautiously and weighing up all relevant issues. We will flash the warning light if needed. Central banks embrace innovation; in the few cases where they put the brakes on, it's for a good reason. We have to make sure that innovations set the right course for the economy, for businesses, for citizens, for society as a whole. This is what we are doing now.

Translation of a column by Mr Agustín Carstens, General Manager of the BIS, in the Frankfurter Allgemeine Zeitung (FAZ), 14 June 2019.

¹ Based on the 2019 Whitaker Lecture, delivered at the Central Bank of Ireland, Dublin, 22 March 2019.