The incoherence of the compromise in policy

I'd been drafting a different next post in my own piecemeal way over the past few weeks (more of a long term reflections note) but as they say: things happen.

Lots can be said about the bill Congress just passed. I leave most of that to the vast wealth of commentators out there.

But there's one quite small provision in the law I want to pull out as instructive for those of us who think a great deal about implementation of policy, its failings and foibles.

Section 10105 of the One Big Beautiful Bill Act deals with SNAP, and can be summarized roughly as "makes states pay a portion of the cost of SNAP benefits starting in 2028, depending on their payment error rate."

But that summary doesn't capture it. What it more specifically does is the following:

- For states with payment error rates *less than 6%*, there is no state cost share

- For states with payment error rates *6%-13.3%*, there is a sliding scale of 5-15% of SNAP benefit costs those states must pay starting in fall of 2027

- And for states with payment error rates *higher than 13.3%*... there is no state cost sharing until late 2029

If this doesn't make sense to you, be not afraid! It doesn't make much sense without knowing the legislative history.

This final policy design can be explained with the following context:

- Alaska has the highest SNAP payment error rate in the country (24.66%)

- Senator Murkowski of Alaska was a final holdout because of the drastically negative impact this provision (among others) would have on her state

- Initially, Alaska was proposed to be exempted from this — but the Senate Parliamentarian apparently said the rules would not allow such an ad hoc provision in a reconciliation bill

- And so the bill was amended to say give an out for many states with *high* payment error rates (not just Alaska)

The policy design, we have to say, does not... make sense. Or more specifically, it is incoherent because you can't take the above provisions and really create a sensible policy objective out of them. It just, well, it is what it is.

What is the effect of this provision's policy design? Bizarrely, it incentivizes most state agencies to increase their payment error rate as doing so could save them up to hundreds of millions of dollars in state fiscal liabilities. That is most certainly not what anyone — least of all the Republican coalition — was pursuing as a broad outcome here!

And this gets to a phrase a friend of mine coined many years back: the incoherence of the compromise.

Policy (and particularly law) arises from processes of compromise in many, if not most, cases. (This idea rhymes with the notion of "Everything Bagel liberalism," though that speaks more to the specifically high-compromising nature of the Democratic party coalition.)

But if you, like me, are more of an implementer, you have to stand back and say: the policies you are implementing are not by necessity coherent — and you should not think of yourself as implementing some implicit coherence.

A systems thinking fan might put another way: the purpose of a policy is what it does. And nothing more.

What does this mean for implementers? I tend to think it means this: there is no such thing as 'policy loopholes' — the laws and policy as written (its opportunities and constraints) that one is given to implement are the raw material one works with.

The idea that implementation itself is solely pursuit of some accepted higher order goals or intents somewhat of a fallacy. Policies just... are. And those who implement said laws face constraints, but not implicit constraints somewhere "under" the rules as written.

Implementers themselves are always making choices that align to some outcome or objective, whether they do so intentionally or not. (In fact, acting like you're *not* is itself a choice!)

There’s no grand takeaway here, other than I do appreciate having read Roland Barthes in college.