Proprietary software business models have led to two troubling and expensive trends.
First, to win government contracts, proprietary software vendors often submit loss-making proposals for providing services. They make up for that initial loss over the duration of their contract using one aspect of their solution completely in their control: they make their solution so proprietary that converting the data files and other artifacts created by their solution to any competitor's solution - while retaining any sort of continuity of capability and data retention - prohibitively expensive.
Second, most government customers fail to allocate the "switching cost" created by the incumbent suppliers' lock-in efforts appropriately to the incumbent supplier and instead associate that cost with the competitor's proposed solution. This has the tendency to give incumbent vendors long dynasties, and the lock-in gets progressively more painful to break. In the odd case where a competitor does prevail, they often need to take a big up-front loss to win the contract, only increasing their incentive to similarly lock-in the government customer to maximise their return.
This situation is not good for government, nor is it good for the taxpayer who ultimately foots the bill.