It's a useful budgeting tool. It's not like there is one "government bank account" that every agency and bureaucrat shares.
In the US, agencies can end up with surplus cash (either taxes were higher than needed, or budgets were off, etc). Agencies will buy T bonds with this excess cash to transfer it back to the Treasuries general fund, so it can be reallocated. The T bonds that the agency now holds are far from "meaningless".
From a thousand foot view, the government is one homogenous entity. But that perspective breaks down pretty quickly in the real world. It's pretty easy to find Intragovernmental debt stats and subtract that from the total, if you are so inclined.
Usually, people are concern with the total liability of a government. You can't dismiss intragovernmental debt in this calculation.
I understand that it is an accounting tool. But the function in the economy is totally different.
The case you mention is basically an ability for different departments to earn some interest on unspent budgets and that's fine. Its basically some bureaucracy level budget adjustment mechanism.
However on a state budget and state central bank level the interest paying part of the Bond is not relevant.
And when comparing how large governments debt% is, to have a useful metric for practical purposes those bonds should be removed.
The Central Bank would need to sell that debt on the open market to quell inflation so it is still relevant. If the debt is high enough that it constrains the central banks ability to manage inflation without causing default either high inflation or massive private debt default are inevitable in the long run.
The only “out” is a miraculous increase in future productivity or an increasing percentage of working age adults that increase GDP to make it sustainable. Of course in the long run we are all dead.
Then its meaningless. It should just be subtracted form total debt.