It sucks because they're called on to provide solutions to deep problems, with only the skills of enthusiastic amateurs. Manufacturing solved a lot of problems. It's completely reasonable for an amateur to look at history for past solutions to current problems.
For reference, the us congress has a panel of 700 or so experts in all sorts of fields to help write policy. This is in addition to each congresspersons's staff. There's something on the order of 2000 people available to write legislation. And about 10000 pieces of legislation written. So each potential law gets a few months of effort.
It boils down to a congressperson (probably a lawyer), a few enthusiastic political scientists, and a few economists to shape national policy.
It's not that they don't produce good work. They're smart hardworking people. A trillion dollar budget is probably as difficult to produce as a decent compiler. That's the kind of thing we'll see congress try to slap together right after the election. It's not much different than a team of undergrads trying to get their compiler to pass test cases.
It's not inherently bad, it's that sometimes problems are really hard, and they should probably be handled by experts. We don't have a good way of collectively picking experts, so we pick amateurs and hope for the best.