For many politicians, GSE reform is a shorthand term, which translates into their plan to shut down the government-sponsored enterprises in favor of an untested system of housing finance. Though they speak of the urgency of reform, they often sidestep a basic question. Reform what? What was it about the Fannie Mae and Freddie Mac that caused problems in the first place? Their answers seem slippery, disconnected from empirical data.

Do they want to reform GSE underwriting? That’s a tough case to make in the business world, because risk taking never exists in a vacuum. And by any metric of loan performance, the GSEs’ underwriting standards have proved to be vastly superior to those of any other segment of the market. This lopsided disparity—concerning rates of delinquency, foreclosure and loss severity—extends back for decades up to the present day.

So why, then, did Fannie and Freddie become insolvent? And why did the government need to bail them out for $187 billion? Fortunately, five years of financial data provides us with a clear and cogent answer. But for three unusual factors—the elimination of deferred tax assets ($93 billion), government-mandated 10% dividends ($46 billion), and losses that predated legal recoveries on mortgages and securities sold under false pretenses ($43 billion)—almost all of the bailout funds would have been unnecessary. These funding obligations had almost nothing to do with the GSE business model.

Before walking through the numbers, let’s note something that is obvious to many bankers. There’s a big difference between insolvency and a proverbial run on the bank. AIG, Lehman, Bear Stearns and others faced the type of imminent liquidity crisis that can force an immediate shutdown of operations. Banks enmeshed with CDOs and credit default swaps had an immediate need for government funds, whereas the primary trigger for the government takeover of Fannie and Freddie was a revised assumption about timing differences under GAAP.

Specifically, two weeks after it told Fannie that it was adequately capitalized, the GSE regulator reversed its position, primarily because it believed that neither Fannie nor Freddie would ever earn taxable income again. Under GAAP all banks book loan loss provisions, and tax benefits, when they see signs of trouble. But under the Internal Revenue Code, banks can only take tax deductions much later, when the distressed loans are finally liquidated off the books. Until that point, banks post those deferred tax benefits as an asset, which bolsters net worth.

The elimination of deferred tax assets contributed to $93 billion in reported losses at Fannie and Freddie between 2008 and 2011, and the restoration of deferred tax assets contributed to $93 billion in reported profits over the past year. The original $93 billion in booked losses triggered a $93 billion cash infusion by the government, which will now be fully repaid. So, while it was appropriate for accountants to err on the side of caution in 2008, we now know with 20/20 hindsight that they erred nonetheless.

Another chunk of the $187 billion bailout was used to fund $46 billionin senior preferred dividends to the government, which mandated that a 10% dividend on all Treasury draws must be paid in perpetuity. More recently, Treasury declared that it will sweep all GSE income as dividends to the government, thereby preempting the companies’ ability to restore their capital bases.

Finally, we know that a big chunk of reported losses were tied to the sale of defective mortgages to the GSEs and the sale of private label residential mortgage securities under false pretenses. So far, the GSEs have recovered about $18 billion from banks that sold defective mortgages.

To get a flavor of the evidence assembled by the GSEs against 17 bank underwriters, check out any complaint and look for the table that compares the stated percentage of mortgages with a loan-to-value in excess of 100%, and the actual percentage of mortgages that were underwater at the time of closing. They’re all alike. You can scan tables showing hundreds and hundreds of RMBS deals, which all say that there were zero mortgages in the pool with an LTV higher than 100%.  And, based on a review of 1,000 loan files from each of those deals, you see that every single transaction was stuffed with mortgages with LTVs higher than 100%. There’s plenty of other incriminating evidence, so it seems hard to believe that a bank would persuade a judge or jury that it met its legal duty to validate the accuracy and completeness of a prospectus filed with the SEC Bloomberg BNA estimates that the GSEs will recover about $25 billion from these securities suits.

Of course, the GSEs incurred unprecedented credit losses over the past five years, as did anyone else who financed mortgages that soon went underwater. But in terms of cause and effect, it seems that a lot of those GSE losses are traceable to the contagion effect of foreclosures tied to private-label RMBS, which realized about 3.5 times the credit losses of the GSEs.

David Fiderer has previously worked in energy banking for more than 20 years. He is currently working on a book about the ratings agencies.