Is it better to lower prices, or lower costs?

I gave a talk for a class at Wharton last week, and made an offhand comment that preventing an unnecessary laboratory test might lower the amount paid by an insurer,  but since the marginal costs of providing a laboratory test approached zero, this didn’t really lower the cost of health care.


One  class member challenged me on this, and stated that from the purchaser perspective, avoidance of paying for a element of care was highly beneficial (regardless of marginal cost of production) .


I’ve been mulling this over for a few days, and I think that the answer is “it depends.”


If I am purchasing a tank of gasoline for my car, I pick the lowest price per gallon today.  I don’t much care if the gas station increased its efficiency through automation, got a better deal from the distributer, or the oil company paid lower tariffs.  Lower price is what counts to me.  I don’t even care if the gas station loses money.  My dealings with gas stations are purely “one off,” and I can easily go to a competitor the following week.  


On the other hand, when Toyota purchases radios for its automobiles, there is a complicated calculus of quality, service, and price – and Toyota is likely to have an ongoing relationship with the vendor. Therefore, a simple one-time price concession is less valuable than a process improvement that allows long-term lower prices. Even Wal-Mart, which is much-feared by its vendors, makes investments to try to lower vendor resource costs so that there will be more of a “win-win” where Wal-Mart gets a low price and the vendor makes a (small) profit.


Denying payment for services rendered can sometimes transmit an important signal.  For instance, Medicare saves only a few dollars  by no longer paying for “never events” (such as wrong-side surgery),  and the providers do spend real resources delivering such defective service.   In this instance, denying payment sends a signal that never events should not happen, which should lead to process changes that would decrease the number of such egregious errors.   This is not the case with a lab test with no marginal cost; preventing this test would not lead to decreased resource use, and could not lead to long-term sustainable lower prices.


 So to sum it up, it’s most important to reduce the price paid for a one-time purchase, and more valuable to reduce the resources used in a long term relationship.  It’s possible that eliminating reimbursement for a service will lead to decreasing future costs – and in those instances cutting such reimbursements leads to savings in the short and the long term.