11.4 — Money, Prices and Why Things Cost What They Do

Prices look arbitrary until you find the constraint they are solving. Every question on this page has an answer in either a cost, a risk, or a piece of psychology somebody is charging for.

Restaurants

How does an unlimited buffet make money?

By selling you a fixed price and controlling the variables underneath it. The economics are well understood and every one of the mechanisms is deliberate.

The average is what matters, not the maximum. A buffet is priced against what an average customer eats, and consumption is distributed with a long tail on one side and a hard floor on the other — a very hungry person might eat three times the average, but nobody eats zero, and children, older customers and people who came for the company all eat well under. The heavy eaters are subsidised by everybody else at the table, which is why the model works best when parties arrive together.

Food cost is a small fraction of the price. In a typical restaurant, raw ingredients are roughly 25 to 35 per cent of the bill; rent, staff, energy and equipment are the rest. A buffet has an advantage on the labour side: no à la carte kitchen, no waiter taking orders, one production run per dish, and far less plate wastage than a kitchen cooking to order.

The sequence is engineered. Cheap, filling, high-margin items come first and are the easiest to reach — bread, rice, salad, soup, pasta, potatoes. Expensive proteins are placed later, in smaller pans, sometimes carved to order by a person, which slows consumption without refusing anybody. A grill-at-the-table format like Barbeque Nation is doing exactly this: the starters are unlimited and arrive continuously, and they arrive because a table that is full when the main buffet opens eats far less of it.

Drinks are where the profit is. Soft drinks and alcohol carry the highest margins of anything served, are almost always charged separately at a buffet, and are more likely to be ordered by somebody who feels the food was good value.

Plate and cutlery size are chosen. Smaller plates measurably reduce total consumption, which is a well-replicated finding, and so does a longer walk to the counter.

So are they profitable?

Yes, and they are less volatile than à la carte restaurants, which is the real attraction for an operator.

A buffet has predictable revenue per cover, predictable purchasing, simpler staffing, and higher table turnover because customers serve themselves. The risks are on the other side: waste, and footfall. Unsold buffet food at closing is a total loss, so forecasting demand is the whole skill, and a slow evening is expensive in a way that a slow evening in an à la carte kitchen is not.

The failure mode is a location where the average customer eats far more than the model assumed — which is why buffet chains price differently by city and by day, and why lunch is cheaper than dinner by more than the food cost difference.

Luxury

Why are watches so expensive?

Because two completely different products are sold under the same word, and only one of them is telling the time.

A ₹500 quartz watch keeps better time than a ₹5,00,000 mechanical one. That is not an exaggeration or a paradox — a quartz crystal oscillates at 32,768 times a second with excellent stability, and a mechanical escapement runs at 4 to 5 beats a second and is affected by temperature, position and wear. A certified mechanical chronometer is permitted to lose or gain about 4 to 6 seconds a day; a cheap quartz watch does better than that in a month.

So the mechanical watch is not competing on accuracy. Its price is made of:

Labour on parts that cannot be automated. A high-end movement has 200 to 400 components, several of which are finished by hand — polishing bevels on parts nobody will ever see, because the finishing is the product.

Scarcity, enforced deliberately. Production is limited relative to demand, allocation is controlled through authorised dealers, and waiting lists for particular models run for years. That is a supply decision, not a manufacturing constraint.

Brand history, which cannot be bought or rebuilt. A house with a two-century archive and a record of firsts is selling something a new entrant cannot manufacture at any price.

And resale, which is the genuinely unusual part. A small number of models from a small number of makers reliably hold or exceed their retail price on the second-hand market, which converts the purchase from consumption into something closer to an asset — and that expectation is itself part of the demand.

The general principle applies far beyond watches. Once a product is bought for what owning it signals rather than for what it does, the price becomes an input to the value rather than an output of the cost. A cheaper version of the same object would signal less, so lowering the price would reduce demand. Economists call these Veblen goods, and they are the reason luxury brands destroy unsold stock rather than discount it.

Why do diamonds cost what they do?

Covered in 4.5: controlled supply for most of the twentieth century, plus an advertising campaign that invented the convention it was selling into.

The current test of the whole structure is laboratory-grown stones, which are chemically identical and cost a fraction. Prices for lab-grown diamonds have fallen sharply since 2020, and the industry response has been to market natural origin as the product — which is an explicit admission that the physical object was never what was being sold.

Work and time

Why do people retire at sixty?

Because a German chancellor picked a number in 1889 and the world copied the structure rather than the reasoning.

Otto von Bismarck introduced the first state old-age pension in Germany, partly to undercut the political appeal of socialist parties. The original qualifying age was 70, lowered to 65 in 1916.

The number was chosen against the life expectancy of the time, which meant a comparatively small fraction of workers would ever claim it, and those who did would claim for a few years. The pension was designed to be affordable precisely because most people died before receiving it.

Other countries copied 65, and India's colonial and post-independence services set 55 and later 58 or 60 for various categories. In none of these cases was the age derived from evidence about capacity to work.

The arithmetic has since inverted completely. Life expectancy at 60 in most middle and high-income countries is now around 20 to 25 further years, so a pension system designed for a 3-year average payout is funding a 20-year one, with fewer working-age contributors per retiree because birth rates have fallen. That is the entire pension crisis, stated in two sentences, and it is why retirement ages are being raised across Europe and why the argument is politically vicious everywhere it happens.

The physical case for 60 is weak and getting weaker. For most modern work the relevant capacities decline slowly and unevenly, and the strongest predictors of when somebody should stop are occupation and health rather than birth year. Mandatory retirement ages have been abolished or restricted in several countries for exactly that reason.

Why does a service company bill people who have no work?

Because the company is selling capacity, and capacity has to exist before it is needed.

An IT services firm signs contracts requiring a stated number of people with stated skills available on stated dates. Recruiting and training a specialist takes months. So the firm carries a bench — employees between projects — and the cost of that bench is priced into the rate charged for everybody who is billing.

The metric that governs the whole business is utilisation: the fraction of employees who are billable at a given moment. Run it too low and margins collapse. Run it too high and the firm cannot staff a new contract without poaching from an existing one, which risks the client relationship, so most firms target somewhere in the region of 75 to 85 per cent and treat the remainder as a cost of readiness.

The employee's experience of this is genuinely strange — paid, employed, and doing nothing — and it is not generosity. It is an option the company has bought, and the option has value precisely because it is unused.

Advertising

How does a company decide what goes on a billboard?

By computing an expected value, and the inputs are more measurable than most people assume.

Reach is measured, not guessed. Outdoor advertising is priced on audited traffic counts, dwell time at the location, and visibility — a board on a slow-moving approach to a junction is worth more than one on a fast bypass, because the same vehicle count delivers more seconds of attention.

The product chosen is the one with the widest applicable audience and the longest decision cycle. A billboard cannot be targeted, so it is wasted on a niche product. It works for categories where nearly everybody is a potential buyer and where the purchase is considered over weeks — cars, phones, property, insurance, banks, films.

The measurement is the hard part, and the honest answer is that it is estimated rather than known. Techniques include geographic holdouts — running the campaign in some cities and not others and comparing sales — and modelling of aggregate spend against outcome. Brand advertising is one of the least attributable forms of spending in business, which is why the digital advertising industry grew so fast: it could show a click.

Risk is handled by diversification and by testing. No serious advertiser puts a large budget behind an untested creative; concepts are tested in small markets or in digital form first, and the billboard runs the version that already performed.

Where prices come from

Why does the same product cost different amounts in different places?

Because a seller charges what each market will bear, and can only do it when the markets can be kept apart.

Price discrimination is the formal term, and it requires three conditions: the seller has some pricing power, buyers differ in willingness to pay, and resale between them can be prevented. Software licensing by country, cinema tickets by age, airline fares by booking date, and medicines priced differently by market are all the same mechanism.

The airline case is the most elaborate. A seat's price varies by how far ahead it is bought, how flexible it is, whether a Saturday night is included, and how full the flight already is — because the airline is trying to separate the business traveller who must fly on Tuesday from the tourist who will take any week in March, and every restriction in the fare rules exists to stop one buying the other's ticket.

The mechanism collapses when resale is possible, which is why grey markets exist and why manufacturers fight parallel imports.

Why do prices end in 99?

Because it works, measurably, and the reasons are partly stupid and partly rational.

The left-digit effect is well documented: ₹299 is processed as "two hundred and something" rather than as approximately three hundred, because reading is left to right and the first digit anchors the impression. The effect is largest when the leading digit changes — 299 to 300 feels bigger than 289 to 290.

A second reason is historical and about theft. Odd prices force the cashier to open the till to give change, which creates a record of the transaction. Whether this was ever the main motive is disputed, but it is repeatedly cited in early twentieth-century retail manuals.

And 99-endings signal a category. A price of 999 says discount; a price of 1,000 says considered. Which is why luxury goods almost never end in 9, and why the same shop will use round numbers in one department and 99s in another.

What actually determines the price of anything?

Not the cost of making it, which is the most common misunderstanding.

Cost sets a floor below which a seller will not operate for long. Everything above that floor is set by what buyers will pay, which depends on what alternatives exist. A product with many close substitutes is priced near its cost, because any seller charging more loses the sale. A product with no substitute is priced at what the buyer can bear.

That is the whole reason patents, brands and network effects are valuable. Each of them removes substitutes, and removing substitutes is what allows a price above cost.

The clearest demonstration is water and diamonds, which is the oldest puzzle in economics. Water is essential and cheap; diamonds are useless and expensive. The resolution is that price tracks the value of the next unit, not the total value of the category. You already have enough water, so one more litre is worth almost nothing to you — even though water in total is worth your life. Diamonds are scarce, so one more is worth a great deal to whoever wants one.

What comes next

The last Part is the miscellany: food, festivals, animals, the meaning of degrees and job titles, and the leftover facts that did not belong anywhere else and were too good to drop.