If you haven't seen it, Slate's John Dickerson has an excellent exchange up this week with with Peter Orszag, the director of the Office of Management and Budget. Dickerson grills Orszag on a variety of health care issues from budget policy to big picture arguments that your "neighbor at the market who is skeptical about health care reform ... needs to hear before he gets distracted by the sale on doughnuts." Orszag, as always, was on the money. The whole piece was quite good, and we thought Orszag's explanation of the difference between scorable savings and game changing reforms that can bend the cost curve, was worth highlighting:

First, let me be clear: The administration is committed to paying in full for the costs of expanding health care coverage over the next decade with hard, scoreable savings—in other words, savings and revenue proposals that are not speculative, but rather have expressed dollar figures attached to them by CBO.

Just making sure that health reform does not add to our budget deficits over the next decade will not be enough to address our long-term fiscal challenge, however, nor will it prevent health care costs from eating up a larger and larger share of people's pocketbooks. Without structural changes to the health system, we would be expanding coverage to millions of people in a fiscally unsustainable way, because we would be bringing them into a system in which we provide more care rather than better care. Then, in no time at all, we would still be facing the high cost growth that has put the current health system on an unsustainable path.

And we do not have to look far to see what a high-quality, lower-cost system looks like because in places all across the country, hospitals, doctors, and other providers are already providing this kind of care. Indeed, the wide variation in how medicine is practiced, how patients fare, and how much it all costs points to a huge opportunity—as much as $700 billion a year, according to some estimates—for reducing overall health care costs in the long run.