The graph above, which portrays Medicare as a percentage of gdp, is from this SSA piece.  In contrast, Matt Yglesias, Kevin Drum, and others have touted a new short essay as evidence for the claim that the Obama health reform plan will succeed on the cost control front, or at least offer a reasonable chance of succeeding, or at least offers some components which will not be reversed.  Here is one key paragraph:

Virtually all of the Medicare cuts enacted in 1990 and 1993, which accounted for a significant portion of the savings in those large deficit-reduction packages, were implemented. And most of the savings enacted in 1997 other than the SGR cuts – nearly four-fifths – were implemented as well.

Given that Medicare spending growth slowed significantly more than was anticipated after 1997 – in 1999, for the first time ever, it was actually lower than the previous year’s level – and the budget was balanced in 1998 for the first time in 28 years, it is surprising that Congress did not scale back even more of the savings enacted in 1997. There is little likelihood that the positive budgetary outlook that encouraged some easing of the 1997 cuts will return in coming years.

See also Box 2 in the piece (which starts slowly, so skip ahead to the meat I am citing).  If you're wondering about discrepancies between these numbers and the SSA graph, the latter is as a percentage of gdp.

My view is this: the aggregate data show that Medicare expenditures, as a percentage of gdp, have expanded at a healthy clip for every medium-run period you can find since 1973.  I don't doubt that the future — like the past — may well show some shorter periods which look better than others but cost control has never worked in the past on anything but a temporary basis.  Citing a bunch of short periods of time doesn't convince me; they didn't stick!  And only one three-year batch of cost controls showed up, as a success, in the aggregate historical data at all.  (Would you believe a worsening alcoholic who pointed to many days or even weeks where his rate of drinking was declining and also mentioned that he drank less for a few years starting in 1993?  Or maybe this reminds you ever so slightly of the debates over recent global cooling and short vs. long-term trends?  Most progressives recognize that a few years of cooling do not contradict the evidence about the long-term trend and yet here is an odd flip of emphasis on a few short-term improvements.)

In Figure D you'll also see that the savings from the 1993-1996 partially period are offset by later, more rapid increases in Medicare spending as a percentage of gdp.

Three additional points are worth consideration:

1. The period of Medicare cost savings, in the early to mid 1990s, coincides roughly with a more general period of cost savings in health care, due to managed care.  This was soundly rejected by the American public, both in their roles as consumers and voters.

2. There will be more and more older voters in the years to come.

3. We should give at least some consideration to a "mean reversion" theory, by which current cost savings increase the pressure for future splurges.  I don't want to push this view too hard, but the aggregate data, as I eyeball them, seem to imply "do not reject" for this hypothesis.

On the other side of the ledger, you might argue, pro-Obama, that the very act of passing the legislation represents a countervailing force against this long-run trend of rising costs.

You can still argue for the bill on this basis: "Congress will increase future spending on Medicare as much as it can.  Any other expenditures in the meantime serve a "stuff the beast" function and slow down the future rate of growth on Medicare expenditure.  We'd rather spend the money on extra coverage now, realizing that the threat of future fiscal crisis will force later Medicare cuts."

That's not my point of view, but it's what I think the debate on cost control boils down to.  The best case scenario for the bill is that it won't much help cost control, may not hurt it, but by pre-emption will result in more money spent on coverage and less money spent on old people.