One popular idea in healthcare policy has been that providers, such as hospitals, make up for low Medicare prices by charging more for the services they provide to private insurers. This is called cost shifting and hospitals and the insurance industry have based argu-ments against reductions in government reimbursement rates on the concept. But some experts, such as Austin Frakt of The Incidental Economist, claim (incorrectly) that cost shifting is something that almost never happens.

Drawing on a monograph by Michael Morrisey, Frakt reviews the argument against cost shifting in several posts (here and here) at The Incidental Economist. To illustrate the analysis, Frakt uses the standard profit-maximizing model of monopoly, with downward sloping demand, etc., and he demonstrates that as Medicare reimbursement declines, hospitals reduce Medicare capacity and increase their capacity for private patients, which results in lower prices for private insurers.

So rather than raising private sector rates, Frakt purports to show that reductions in Medicare prices would actually lower both rates. But a major problem here is that Frakt assumes that hospitals are able to reduce their capacity for Medicare patients in response to lower reimbursement rates, and his conclusions depend upon this assumption. Of course, assumptions are necessary when using models, but this one is simply untenable.

It’s unlikely that hospitals could refuse service to a significant number of Medicare pa-tients while accepting other Medicare patients. Perhaps a single hospital in the market could get away with this kind of discrimination, but not all hospitals in the market could. At some point, such conduct would bring the wrath of the government down upon every-one. If we reject the assumption, then Frakt’s analysis falls apart and his conclusions about the effects of Medicare price reductions are not valid.

Another problem with the analysis is that private price and capacity, as seen in Frakt’s diagram, are determined at a point where marginal revenue exceeds marginal cost. In a profit maximization hypothesis, which this is supposed to be, we would find the private price level and admissions at the point where marginal revenue equals marginal cost. And this would be true even if marginal revenue from Medicare patients exceeded the margin-al revenue derived from private patients. All that counts is that MR > MC.

So the Frakt/Morrisey diagram is incoherent. Because private admissions would always be at the profit-maximizing level, any attempt to increase them, in response to a reduction in Medicare price, would only reduce profits. And any attempt to reduce Medicare capa-city would also reduce profits, unless the price reduction drove price below marginal cost. Most likely, hospitals would hold firm on admissions for both private and Medicare patients.

How would a Medicare price cut really affect hospitals? For the Medicare segment of its business, each hospital would immediately lose revenues (which would be easy to calcu-late) and because prices are below average costs (experts agree on this), the lost revenues would represent lost income. The lost income would be an additional cost imposed on each hospital, and any analysis of the market for privately insured patients would include them.

In the private market, an increase in costs for each hospital would cause an upward shift in the cost curve for each hospital as well as an upward shift in the supply curve for the entire private market. This means the new equilibrium price for each hospital and, in the aggregate, for the market would be higher. And, of course, supply and demand also sug-gests that capacity for private sector patients would be reduced.

Although a reduction in private capacity would be predicted at the new equilibrium price, it’s unlikely that much reduction would occur in practice. Hospital costs account for approximately one-third of total healthcare costs, so the impact of hospital price increases on the overall price of health plans would be diluted. Demand is relatively inelastic for most health plan purchasers – that is, they would simply eat the increase rather than drop the insurance.

So it appears that those who argue that Medicare price reductions increase the price for private insurers are correct after all. Not only does cost shifting make sense when erro-neous assumptions are rejected, but it would occur when hospitals maximize profits and even when they lack market power.