The vote to hike was unanimous, two months before the midterms and after a week of public pressure from the White House.

Last week's 25bp hike proves less than the headlines suggest, because both halves of the Fed's dual mandate pointed the same direction. Inflation is at 3.4%, payrolls beat, and the committee was already leaning hawkish. Independence gets tested when the dual mandate points opposite ways, and nobody can prove which reason won.

After the vote, the president said he wanted 1% or less, and his complaint included that higher rates raise what the government pays on its debt. Anyone who wants cheap money to cut the interest bill will still want it when inflation cools and when the labor market weakens. The argument only gets louder as the debt grows. The direct fix would be to borrow less than the $1.9 trillion CBO projects for this fiscal year, and so far Congress has shown no appetite for it.

And this is one hike. Across the seven tightening episodes since 1983, the median cumulative increase was 300 basis points start to finish, based on St. Louis Fed figures. Fifty by December is a floor, not a forecast. See Exhibit 1, below.

CCM read: What reaches the household is the full cycle, not last week's 25 basis points. This tightening also starts from a supply shock, which the Fed can't fix by making borrowing more expensive. The standard mapping from hike cycles to loss curves is built on demand-driven episodes and may under-predict here. Variable-rate card repricing runs roughly one to two statement cycles, so this lands in fourth quarter remits, not September. Prime is fine. The tail isn't.

Retail sales rose 1.2% in August to $773.9 billion, beating forecasts and running 6.0% above a year ago, according to the Census Bureau.

Census does not adjust these dollars for prices, so where they went matters more than the total. Gasoline stations rose 3.1% on the month and 21.0% over the year, the largest annual increase of any category. Nonstore retailers rose 2.6% and 9.9%. Grocery stores rose 0.5% on the month and 0.5% over the year, which after food inflation means households bought less food than they did last August. Spending did not broaden. It concentrated in the pump.

Census published at 8:30 Wednesday morning. The FOMC voted at 2:00, and Warsh cited a strengthening economy. The Fed will read this as demand. The category detail reads more like cost.

CCM read: Card balances stand at $1.26 trillion, the largest pool of household debt that reprices when the Fed moves. The 50 basis points the Fed will have delivered by December costs roughly $33 a year on an average balance that revolves, which is not the problem. The problem is that the repricing lands on households already paying a fifth more for fuel than they did a year ago, and buying less at the grocery store to do it. The squeeze is arriving from both ends, and only one of them is something the Fed chose.