Medicare costs and losses

An editorial at the WSJ titled Why it's easy to steal from Medicare says, among other things:

One of the purported benefits of nationalized health care is that it will be more efficient than private insurers since it would lack the profit motive and have lower administrative expenses, like Medicare. But one reason entitlement programs are so easy to defraud is precisely because they don't have those overhead costs -- they automatically pay whatever bills roll in with valid claims numbers.

By contrast, private insurers try to manage care, and that takes money. Not only does administrative spending go toward screening for waste and fraud -- logical, given the return-on-investment incentives -- they also go toward building networks of (honest) doctors and other providers. Medicare doesn't pay for this legwork, so it simply counts fraud losses as more spending.

One of the claims made on behalf of medicare is that it has a very low overhead. I've suspected this overhead is as low as claimed because a number of costs, which count as part of the overhead in the private system, are fobbed off on someone else and not counted as part of Medicare's expenses.

This would be one example. By not spending on preventing fraud before payment is made, overhead is magically converted into product. Masscare reports that Medicare spends 5% of every dollar on administrative costs, while the average private insurer spends 17%. Let's see what happens when we go in to the private insurer and move a penny from fraud prevention to payment of claims.

One question that immediately arises is, how many pennies of fraudulent claims did that penny of administrative expense prevent? Did it prevent two pennies of fraud? Ten? A thousand?

In a free market, insurers are going to allocate expenses based on how much good they yield. An insurer is not going to spend a penny to save less than a penny in fraudulent claims -- it's cheaper to pay less than a penny in fraudulent claims.

We can assume that insurers will pay for fraud prevention up to the point where the savings in fraudulent claims prevented just balances the increase in expenditures. On the margin, a penny spent saves a penny in fraud.

So we take the insurer that spends 17 pennies per dollar on administrative costs, and we move a penny from fraud prevention to paying claims. Since this is a marginal penny, the increase in fraud will not be much greater than the savings realized by the cut.

So administrative expenses decrease by one penny to 16 pennies on the dollar. However, the fraudulent payments that are no longer being prevented are now counted as product. A private insurer which achieves a penny in real savings now has an overhead of 16 cents over 99 cents, or 16.16% – about a 4.9% decrease. However, if cutting that penny results in an increase in payouts by that amount, there's no net savings. The overhead shrinks by 5.9%, but the total spent remains the same. And it will actually be slightly more – the curve is not a straight line. Eventually, a penny of cuts results in two, five, ten, or more pennies of additional payments.

A private insurer can't get away with too much of such "savings", because it has a different incentive than Medicare does. If expenses wind up ballooning, a private firm calls this a disaster, or at best, gross inefficiency. Private insurers get their money from the private sector, which means there are only limited funds available to pay claims. If fraud runs rampant, premiums will have to be increased to cover both real claims and fraudulent ones. And in a free market, subscribers will flee the less efficient company and opt for one with better control over its costs.

A government program cites such a growing budget as evidence of serious need, if not a "crisis". A program with exploding claims can agitate for more public funds to address the "crisis", and the government raises taxes. Or prints money.

Update: Fixed the math, and expanded the last two paragraphs.