One can build a very good predictive model of government agency behavior if one assumes the main purpose of the agency is to maximize its budget and staff count. Yes, many in the organization are there because they support the agency's public mission (e.g. protecting the environment at the EPA), but I can tell you from long experience that preservation of their staff and budget will almost always come ahead of their public mission if push comes to shove.
The way, then, to punish an agency is to take away some staff and budget. Nothing else will get their attention. Unfortunately, in most scandals where an agency proves itself to be incompetent or corrupt or both (e.g. IRS, the VA, more recently with OPM and their data breaches) the tendency is to believe the "fix" involves sending the agency more resources. Certainly the agency and its supporters will scream "lack of resources" as an excuse for any problem.
And that is how nearly every failing government agency is rewarded for their failure, rather than punished. Which is why our agencies fail so much.
Note that organizations in the private world are not immune to similar incentives. A company's marketing staff will work hard to get more people and resources for marketing, and in good times their staff and budget may balloon. The difference is that in the private world, there is competition. Other companies are trying to sell similar products and services. And if the marketing department is screwing up a lot, or those resources spent on it are not being used productively, the company is going to lose sales and thus resources. To survive, massive changes will be made, including likely some deep cuts and large restructurings in marketing.
It is frustrating to work in corporations that seem to lurch from growth periods to cutbacks in an endless cycle. But it beats the alternative where the organization always grows and never is forced to confront the value of how it spends its resources.