From the New York Times, of course, not the soporific suburban paper that deceptively puts the word "Seattle" in its banner:

At issue is the way inflation is calculated. The administration’s offer in the fiscal cliff talks — and the approach long advocated by Republicans — calls for using a new measure of inflation, called the “chained” Consumer Price Index, to calculate the COLA.

Unlike the gauge of inflation currently in use, the chained index captures the ability of consumers to adjust their spending across categories as relative prices change — for instance, spending less on fuel as gas prices go up and more on groceries as food prices go down. Such substitution causes the chained C.P.I. to rise more slowly than the current measure, which would result in a lower annual COLA and huge budget savings. The move to a chained C.P.I. would reduce benefits by some $135 billion over 10 years, and far more in later decades because of compounding.

The administration and other proponents of switching to a chained C.P.I. contend that it is a technical fix in the interest of greater accuracy, not a benefit cut per se.

But that claim does not stand up to scrutiny. The chained index is in many ways a better method of tracking price changes for the broad working population, but there is no compelling evidence that it is better for computing the Social Security COLA.

What is known is that elderly households tend to have lower incomes and lower expenditures than younger households, and that more of their purchases are for needs that cannot be met by switching to products and services in unrelated categories. That indicates that they do not have the same flexibility as younger households to respond to price changes while still maintaining their standards of living.

Ohmigod... a newspaper editorial board that actually attempts to understand and explain complex policy issues instead of just bludgeoning readers with regurgitated ideological talking points. How refreshing.

The editorial goes on to suggest that if lawmakers were truly serious about their claimed objective, they would have the Bureau of Labor and Statistics develop a "statistically rigorous index to track inflation as experienced by retirees." Except, if they did that, there's a good chance we might find that the CPI underestimates inflation's impact on retirees. And that would defeat the true purpose of this exercise: cutting benefits for old people.