… is from page 247 of David Friedman’s excellent 1996 book, Hidden Order :
The result – that price control results in a cost to the consumer, pecuniary plus nonpecuniary, higher than the uncontrolled price – does not depend on the details of the [supply and demand] diagram. Consumers cannot consume more gas than producers produce, so the nonpecuniary cost must be large enough to drive quantity demanded down to quantity supplied. Quantity supplied is lower than without price control, so cost to the consumer must be higher.
In other words, because a government-imposed limit – a “ceiling” – on the monetary price that suppliers are allowed to charge (and that buyers are allowed to pay) causes suppliers to supply less than they would supply without this government-imposed price ceiling, the height of the ‘cost-obstacle’ that stands between consumers and supplies of this good or service must increase. With fewer units available for sale (per period of time), something must deter consumers from actually buying the full number of units that they would have bought were a larger number of units available for sale.
That something is typically (although not always) the nonpecuniary cost of queuing – or, in American parlance, “waiting in line.” The time and effort spent waiting in line is a cost to each person who waits in line no less than is the sacrifice of money paid for the good by each person who buys the good. In each case, the person sacrifices something of value in order to acquire something that has for him or her higher value. Waiting in line to buy the good involves the sacrifice of whatever income or enjoyment would otherwise have been yielded to the person had she spent in some other way – say, working at her office job – the time that she actually spent waiting in line. Spending money on the good involves the sacrifice of whatever goods or services that she would have otherwise bought had she not spent the money on this good.*
While the consumer in this example might be indifferent between these two means of expenditures on the good, the suppliers are most certainly not indifferent between them. (Note: because monetary payments are the most general form of payment, in practice most consumers – even ones with relatively modest incomes – would prefer paying money prices to paying nonpecuniary prices. But explanation of this reality is beyond the scope of this post.) To the extent that consumers acquire the good by paying money to the suppliers of the good, the suppliers get something of value in return – which prompts suppliers to supply the good in sufficient quantities. But to the extent that consumers acquire the good by paying for it in some other way that does not redound to the benefit of suppliers – say, by waiting in long lines – the consumers’ expenditures are not translated into efforts by suppliers to bring sufficient quantities of the good to market. Price controls are a policy of forcing an unnecessarily large portion of the amount of resources that consumers spend to acquire goods to be spent in ways that do nothing to call forth greater quantities supplied of goods.
Again, although price ceilings might be sincerely meant to lower consumers’ cost of acquiring price-ceilinged goods, price ceilings lower only the amount of consumers’ expenditures on the good that are received by suppliers of the good. The result is reduced quantities supplied – which, in turn, inevitably cause the nonpecuniary portion of the cost of the good to rise such that the total cost (pecuniary plus nonpecuniary) is higher than the total cost of the good would be without the price ceiling. In short, price ceilings not only artificially reduce the quantities of the good that consumers get, they raise the cost to consumers of acquiring the good.
* To be more precise: As James Buchanan  explains in his brilliant little 1969 book, Cost and Choice , what is ultimately sacrificed is the subjective utility that the person anticipates she would have gotten had she chosen differently than she actually chose.