The breaking point came in 2016, when Germany introduced a new law on electronic invoicing (E-Rechnung). Elster’s implementation was characteristically rigorous: it required invoices to be encoded in a specific, little-used XML dialect (UBL 2.1) with mandatory timestamping via a government-issued certificate. The result was chaos. Thousands of small contractors found they could not submit invoices at all. A plumber who could fix a boiler in thirty minutes might spend two hours fighting Elster’s validation logic. Local tax offices, stripped of their paper-based discretion, could do nothing but point users to the error logs.
In a rare public rebuke, the German Federal Court of Auditors reported that Elster’s precision had actually increased the administrative burden, because citizens now had to hire IT consultants to navigate the software, rather than tax advisors to interpret the law. The machine had not replaced the bureaucrat; it had created a new, more expensive layer of middlemen. elster software
This was not user hostility; it was a logical consequence of the company’s founding philosophy. Elster had built a perfect mirror of the law, only to discover that the law was not, in itself, user-friendly. The software had become a bureaucratic straitjacket, punishing ambiguity and edge cases with digital silence. The breaking point came in 2016, when Germany
The lesson for modern engineers is uncomfortable. We are now building large language models and automated decision systems that promise to replace human judgment. Elster reminds us that the real world is fuzzy, contradictory, and full of exceptions. A system that is 99% precise but 0% tolerant is not a tool—it is a barrier. Elster did not fail because it was poorly coded. It failed because it succeeded in coding the law so perfectly that it forgot the law is, at its heart, a human institution meant to be interpreted, not executed. Thousands of small contractors found they could not
For a decade, Elster was hailed as a triumph of e-government. Its software was free, secure, and ruthlessly efficient. The company’s engineers, many recruited from the same technical universities that fed Deutsche Bahn and Siemens, believed in a philosophy they called Perfektion durch Zwang (Perfection through Compulsion). If a user made a mistake, the software would not simply warn them—it would refuse to proceed. This was not a bug; it was a feature.
Elster Software was dismantled in 2018, its assets nationalized and its team dispersed. But its ghost haunts every conversation about AI, automation, and governance today. Elster’s failure was a textbook case of Goodhart’s law applied to software: when a metric (strict schema validation) becomes the target, it ceases to be a good metric. By eliminating all ambiguity, Elster eliminated all discretion, and without discretion, a bureaucratic system cannot function.
For professional tax advisors and large corporations—users who understood the system—Elster was a powerful tool. But for small business owners, freelancers, and ordinary citizens, it became a nightmare. The software’s refusal to accept “close enough” answers meant that a single misplaced decimal or a missing auxiliary form would freeze the entire submission. Unlike a human clerk, who could exercise discretion or request additional documentation, Elster offered only a cryptic error code: “Validation failed on field 42.3 (Betriebsausgaben).”