For the vast majority of the time, trust in the financial sphere rested on tangible and visible elements. Money (preferably in the form of precious-metal coins) was stored in safes, receipts were issued on paper, and their authenticity was verified by people whose positions followed from their place within the institution and their physical presence. Trust was earned locally, directly, and, truth be told, rather slowly. A bank branch was not simply another service point; it was living proof that the money was stored somewhere, protected by thick walls and steel doors with solid locks. And that its employees carefully checked who could be paid and who should be refused. Visibility was therefore the basis of trust.
When money and financial operations took on a digital form, the foundations of trust had to change as well. We arrived there gradually, and the digitisation of finance began much earlier than the appearance of smartphones and mass-market applications. This transformation did not attract public attention; it took place quietly as part of the optimisation of banks’ internal processes. Personal assessment and physical proof were slowly replaced by technical procedures subject to standardisation, repeatability and scalability. Decisions that had previously required personal involvement and judgment were codified as rules that could be triggered without human intervention.
Long before clients gained access to digital finance, banks were already using digitised internal operations. Mainframe computers replaced the paper general ledger, allowing banking institutions to centralise account records, automatically reconcile balances and process transactions in batches rather than one by one. From the outside, everything looked much the same: a bank remained a bank, a solid institution rooted in reality. Inside, however, money began to be understood more as an entry in a system than as something physical, a coin or a banknote.
This change was of key importance. Trust began to shift away from the concrete institutional participants in financial operations, for example, bank officials, towards organised processes. Trust was dispersed in at least three directions:
- the institution standing behind the operation,
- the procedure that had taken over decision-making,
- and the interface that enabled access to that procedure.
Accuracy and consistency were no longer ensured solely by personal supervision over the process, but also by software, control procedures and the possibility of reviewing the event log.
Of course, errors were not eliminated; they merely changed their character. When something went wrong, identifying responsibility became more difficult because the process was carried out across different systems. A failed execution of a batch of transfers or a duplicated entry on an account, for example, were problems whose causes could be difficult to establish. Responsibility became more closely linked to the design of the process than to organisational hierarchy.
At that stage, the emerging digital system remained inside the institution. Clients still came to branches, instructed transfers through bank staff, and used paper statements. The changes were noticed only by those who designed and used the internal systems directly. The ordinary citizen first had a direct encounter with this “delegated trust” when cash machines first appeared.
We can treat that innovation as a kind of large-scale experiment in trust. For the first time, a bank proposed that clients use a machine to perform an action previously carried out exclusively by a human. And not just any action: withdrawing money. The risk was considerable. Every error could undermine not only trust in the device but also in the institution that made it available.
The success of ATMs lies in changing habits. Users had to accept that a machine could recognise them, knew their account balance, and functioned properly without direct supervision. That trust did not result from knowledge of how the machine worked. It appeared together with repeatability: ATMs functioned correctly often enough for their use to become an accepted norm.
With the development of ATM networks, however, a practical problem appeared: cards issued by one bank ought to work in the ATMs of other banks. Solving this problem required standardising card formats, authentication methods, transaction messages, and settlement methods. These early, seemingly simple technical requirements carried much more serious consequences.
When these requirements became standard, decisions began to be fully automated. A card could only be accepted or rejected; there was no possibility of negotiation or appeal against that decision. There was a subtle shift in decision-making power to the system designers, who define the rules for card use and govern the process. The ordinary user never came into direct contact with them.
From the user’s point of view, this change could look like simplification: using the bank became faster, more predictable, independent of branch opening hours and personal relations. The system, in turn, became much more complicated. A single ATM withdrawal involves a complex combination of identity verification, interbank settlement, balance reconciliation, and connectivity. The complexity of the background processes remained hidden from users, or users had no reason to become aware of them.
One may, however, get the impression that this principle applied at every subsequent stage of the digitisation of finance: complexity was moved into infrastructure, access to services was mediated by carefully designed interfaces, and explanation was replaced by confirmation.
The branch banking model relied on visibility (high street branches), which sustained trust. The first wave of automation (ATMs) built trust through consistency and repeatability. But both models assumed a limited scale. With the growth in the number of transactions and the scale of access to financial services, visibility and physicality ceased to matter in the building of trust. No institution could ensure personal supervision over millions of daily operations. Digitisation, obviously, was the answer to scale.
Internet banking began its triumphant development on very solid foundations: clients were already accustomed to using interfaces (in ATMs) instead of contact with the institution, and to accepting confirmation instead of explanations.
By the end of the twentieth century, the foundations of contemporary digital finance were already fully in place. Money functioned primarily as data, as an entry in the system. We placed our trust in those systems, not in people. Decisions were built into procedures, and interfaces provided us with ready-made answers but lacked justification.
Understanding digital finance begins with recognising that trust has fully shifted from the sphere of human relations to machines.
Sources:
- Nigel Dodd, The Social Life of Money (Princeton: Princeton University Press, 2014)
- Bernardo Batiz-Lazo, Cash and Dash: How ATMs and Computers Changed Banking (Oxford: Oxford University Press, 2018)
- Bank for International Settlements, Payment Systems in the Modern Economy, Annual Economic Report (Basel: BIS, 2001)
- Bank for International Settlements, Sound Practices: Implications of Fintech Developments for Banks and Bank Supervisors (Basel: BIS, 2018)