“We save everything as PDF/A, then we are GoBD-compliant.” This is often said, and it is half right. PDF/A solves one archiving problem very well, but not the one most tax audits focus on.
What PDF/A does
PDF/A is a format for long-term archiving, defined in ISO 19005. Its purpose is narrow: A document should look exactly the same in twenty years as it does today.
To achieve this, it primarily requires that the file contains everything needed for display. Fonts are fully embedded rather than referring to those on the viewer’s device. References to external content are prohibited, as are executable components and encryption. Color behavior is defined unambiguously.
The side effect is that the file becomes larger, often substantially so. Embedded fonts take up space. That is the price of ensuring that no one has to obtain a missing font later.
- PDF/A-1b has the strictest requirements and is explicitly required by some bodies.
- PDF/A-2b is the usual standard and the right choice for most purposes.
- PDF/A-3b additionally allows embedded files, which is the technical basis for ZUGFeRD invoices, where an XML file is contained in the PDF.
The conversion itself is straightforward: PDF to PDF/A takes an existing file and rewrites it in the archive format.
What the GoBD require
The GoBD, German rules issued by the Bundesfinanzministerium, Germany’s Federal Ministry of Finance, set out principles for proper accounting with electronic records. They impose requirements of which a file format covers only a small part.
- Immutability. Once a record has been captured, it must no longer be possible to change it in a way that makes the original version impossible to determine. This is a property of the storage system, not the file. A PDF/A in an ordinary folder on a network drive does not meet this requirement, anyone can replace the file there.
- Traceability and completeness. Every accounting entry must be traceable back to the record and vice versa. Nothing may be missing, and nothing may appear twice.
- Timely capture. Records must be captured promptly, not collected at the end of the year.
- Machine readability. The tax authorities must be able to access and evaluate the data.
- Process documentation. The process must be documented, from the creation of the record to its storage. If it is missing, that is a separate deficiency in an audit.
- Retention period. Invoices must be retained; the period begins at the end of the calendar year in which the record was created.
Process documentation is the point most often missing, and it takes the least effort as long as you complete it while the process is still fresh. It describes who captures records and how, where they go, how they are named, and who has access.
The most common misconception
Scanning paper records and throwing away the paper is permitted, substitution scanning is explicitly provided for. However, it requires that the procedure be described in the process documentation and that the result visually matches the original.
If you scan, you should then apply text recognition. Not because of the GoBD, but for a practical reason: A scan without a text layer is not searchable. If an audit asks for a particular invoice, the difference between a search and leafing through two thousand images is significant.
What about electronic invoices?
Structured formats are gaining ground for invoice exchange between businesses, ZUGFeRD and XRechnung. A ZUGFeRD invoice is a PDF/A-3 with an embedded XML file: A person sees the invoice, a program reads the data.
Important: A PDF invoice does not become a ZUGFeRD invoice afterward. The XML file must be embedded when it is created, by the invoice issuer. If you need the data from existing PDF invoices, you need to extract it, our extract invoice data tool writes the fields to a CSV file. This does not replace a structured invoice, but it saves you from typing everything in.
In brief
- PDF/A solves the question of readability in ten years. Nothing more.
- Immutability is the responsibility of the storage system.
- Without process documentation, even the best technology is of little use.
- Scans need a text layer, otherwise the archive is not searchable.
- ZUGFeRD is created by the invoice issuer, not the recipient.
This text provides general guidance, not tax advice. How exactly you must store your records depends on your business, discuss this with your tax advisor.