According to Faber, the projected spending for the 2002 and 2008 farm bills significantly underestimated actual costs. (In fairness, errors in the other direction could reduce costs.)
Consider the farm bills as a public relations exercise. To make subsidies more acceptable, Congress is repackaging them.
"Direct payments" to farmers are ending, because they seem (and are) a straightforward giveaway.
Instead, "crop insurance" - which seems prudent protection against droughts and other misfortunes - is being expanded.
In reality, crop insurance resembles "a farm income support program" more than standard insurance, writes economist Bruce Babcock of Iowa State University in a report for the EWG.
Farmers' premiums cover only 40 percent of costs; taxpayers pick up 60 percent. With premiums subsidized, farmers buy generous coverage that produces payouts even in many good years.
The CBO puts the 10-year cost at $89 billion.
The survival of farm subsidies is emblematic of a larger problem: Government is biased toward the past.
Old programs, tax breaks and regulatory practices develop strong constituencies and mindsets that frustrate change, even when earlier justifications for their existence have been overtaken by events.
It's no longer possible to argue that ag subsidies will prevent the loss of small family farms, because millions have already disappeared.
It's no longer possible to argue that subsidies are needed for food production, because one major agricultural sector - meat production - lacks subsidies and meat is still produced.
The larger lesson must be discouraging.
Among other qualities, good government requires the capacity to adapt to change.
It needs to discard what doesn't work or is no longer necessary.
It needs to devote its limited resources - in time, skill and money - to the problems where it might do some good.
In the best of circumstances, this is difficult. But routine politics compounds the difficulty, as the immortal farm subsidies and endless debates over budget deficits attest.