Recently, Medicare has been growing more slowly than its historical average. The Congressional Budget Office decided to investigate why, and the answers will probably surprise you.

If you had to guess why Medicare was growing more slowly, you’d probably come up with one of two answers: “Obamacare” or “the recession." (No bonus points for guessing which party prefers which answer.) But the CBO largely rejected both those answers. It found a small role for declining reimbursement rates, and a roughly equal effect for the changing number of seniors who are enrolled only in Part A (since most of these are seniors who are still working, this may be a result of the recession forcing people to work longer in order to prop up their retirement funds).

The largest effect was one I hadn’t expected: the changing age profile of Medicare beneficiaries. The boomers bulge is adding a large influx of relatively young people to the system, dragging down its average age -- which in turn, lowers the average spending per beneficiary.

If you’d ask me what effect the boomers would have on Medicare, I probably would have thought of this eventually, but I wouldn’t have predicted that the effect would be so large. The CBO estimates that this factor alone lowered the annual growth rate of spending per beneficiary by three-tenths of a percentage point.

Of course, that still leaves a good bit to account for, since the gap between the current rate and the historical rate is 3.2 percent. The remainder of that gap is explained by…well, we don’t know. It may be changes in the way that care is delivered (probably driven by insurance companies and Bush-era Medicare changes, not Obamacare, since the broad economy-wide decline in health-care cost inflation started in the middle of the last decade). The gap may be explained by health-care innovations in prescription drugs and medical technology. It might be consumers becoming more price sensitive as health-care costs make up a larger portion of their budgets. Probably, it was some combination of all three; there’s no way to quantify it.

But we can be reasonably sure about what it mostly wasn’t: the two most obvious candidates. Which just goes to show why policy forecasts are so difficult.