This article originally was published Aug. 7, 2011.

Talk about a buzz-killing week. It started with an end to the debt-ceiling nightmare that satisfied no one - the best summation: "You don't 'win' a hostage crisis. You resolve it." It ended with reminders, via the stock market and the jobs report, of why no one can just pinch us and wake us all up.

But along the way, two lesser headlines caught my eye.

One was more evidence that the rich are still getting richer: an 8 percent increase in luxury-store sales - the latest echo of a study showing that by the eve of the Great Recession, the top 1 percent of U.S. earners had captured a larger share of the nation's income than any time since 1928.

The other was a remark by House Majority Leader Eric Cantor (R., Va.), who told the Wall Street Journal's Opinion Journal it was time for Americans to "come to grips with the fact that promises have been made that frankly are not going to be kept for many."

Cantor may have been mostly focused on Medicare - in particular, on House Budget Chairman Paul Ryan's proposal, already passed by the House, to dump the current Medicare system for people now under 55. Instead, future seniors would get "premium support" - vouchers to buy private insurance that would, depending on your faith in the health-care marketplace, either force medical costs to moderate or leave less-affluent seniors with increasingly inadequate coverage.

As he has before, Cantor said the goal was to protect those nearing retirement while putting everybody else "on notice." "The rest of us have got ample time to try and plan our lives so that we can adjust to reality here when you look at the numbers. Again the math doesn't lie," Cantor said.

Yup, math doesn't lie. But politicians often use it in misleading ways, especially when they talk about costly, popular programs such as Medicare and Social Security and are mesmerized by the magical notion that, whatever the nation's needs, taxes can only ever be adjusted in one direction: down.

And there may be no better example than the latest simplistic plan - distressingly backed by voices from both sides of the aisle - to "fix" Social Security by changing the way benefits increase over time to account for a rising cost of living.

The idea: Shift annual adjustments downward by basing them on the so-called "chained" Consumer Price Index, a version designed to reflect how consumers shift their behavior when they face higher prices. The idea is that some people who can afford, say, an increase in the price of beef will simply absorb it, but others will instead buy more pork or chicken.

Using the chained CPI is a proposal with a lot of weight behind it from Washington's insider class of "very serious people," such as private-equity billionaire Pete Peterson and the chairs of the Bowles-Simpson fiscal-reform commission. It's even won support from some liberal groups that portray it as a mere "technical change" - albeit one that would save the federal government about $300 billion over the next decade, more than a third of that by reducing Social Security costs, and thus help ensure the program's long-term solvency without what Harry Potter might call the fiscal-adjustment-that-must-not-be-named.

If the change goes through, benefits would rise less over time, because the chained CPI - by definition - grows more slowly than the unchained version. The National Women's Law Center estimates that today's average 65-year-old retiree, now receiving about $16,800 a year, would draw $852 less a year at age 80. At 90, she'd get $1,332 less in 2010 dollars, an 8 percent reduction.

To the affluent, those differences might not seem huge. But tens of millions of elderly Americans rely on Social Security as their major source of income: 52 percent of elderly couples and 73 percent of nonmarried beneficiaries, according to the latest data. And those fractions could rise as traditional pensions vanish and private savings are battered by bad luck, bad decisions, or bad markets.

But the real question should be whether using the chained CPI would really do what its proponents promise: not just save billions, but also better reflect true changes in the cost of living for those on Social Security.

One economist challenging the tide is Dean Baker, of Washington's Center for Economic and Policy Research. Baker points out that not only do elderly consumers tend to buy a different mix of goods and services, but that the very folks behind the Consumer Price Index - the economists at the Bureau of Labor Statistics - have long recognized the issue.

In fact, since the late 1980s, they've been calculating an experimental "CPI-E" to reflect the buying habits of the elderly, who spend more, for instance, on medical care and housing than younger people, but less for clothing and food.

The CPI-E isn't a new index - it's just a reweighting of data collected for the widely cited CPI-U, the index "for all urban consumers," and the CPI-W, the older, narrower index for "urban wage earners and clerical workers" that is, oddly, the one still used to calculate the Social Security COLA.

What does the CPI-E index indicate? "Historically, on average, the CPI-E has risen slightly higher than the U and the W," says BLS economist Kenneth J. Stewart.

But that's not always true. Over the last year, for instance, the elderly index is up 3.2 percent, Stewart says, vs. 3.6 percent for the CPI-U and 4.1 percent for the CPI-W - an index that gives greater weight than its counterparts to gasoline prices.

Baker says rightly that the goal should be to use the most accurate measure possible, not the one that saves the most money. And he notes that a chained version of the CPI-E could be created, with proper study of their buying habits. The elderly may be a little less able, for instance, to drive to Walmart to save.

"If the concern is to make it accurate, we know how to do that," Baker says. "What this is about is reducing benefits."

Amazing what you can do with math.

Contact columnist Jeff Gelles at 215-854-2776 or