IMO File formats are not the real problem. Microsoft's binary word processor and spreadsheet formats were reverse engineered years ago and have been pretty stable since 2000. OOXML is XML based and even has some documentation available on how to read it.
The real problem is that office documents blur the line between input and output and this makes them fundamentally fragile. An office document is input to a layout (in the case of a word processor) or calculation (in the case of a spreadsheet) engine but the user always looks at the output of that engine. Especially with word processors since the user is always looking at the output they aren't thinking about the structure of the input, they just bash things arround (holding down the space bar or enter key for example or dragging boxes around with no idea if their position is text-relative or page-relative)
So I don't think this will solve anything, even if MS implements ODF and even if the UK government gets it's employees to start using it as their main format for storing files (good luck) I would expect loading a document from office into libreoffice to still have similar results to today. The input (text typed, pictures included, user-specified values in spreadsheet cells) will probablly carry across fine but in some cases it will result in noticably different output (different and possiblly unreadable layout for word processed documents, different rounding of results for spreadsheets). Especially for large badly structure docuements.