, Britain
In January 1988 a housebuilder in Reading had twenty-eight plots with permission, an option over a field two miles further out, and a valuation on the plots comfortably above what he had paid for them eleven months earlier. He had not laid a brick.
That was not unusual. It is what a rising market does to a business whose main asset is land. Prices in the Thames Valley had been climbing for years, the bank was lending against the valuation rather than the cost, and every month he held the site it was worth more.
In May that year base rate was seven and a half per cent. By October of the following year it was fifteen, and it stayed at fifteen for a year.
A housebuilder is two businesses, and only one of them builds
The first business is construction. You buy materials and labour, you put up houses, and the difference between the cost of building and the price of the finished house is the build margin. Estimators have been producing that number for a century and it does not move a great deal.
The second business is land, and that is where the money is made and lost. A builder buys plots, or the right to buy plots, holds them, and sells them with houses on top. The value of that land is not a separate thing from the price of the houses. It is a residual: what the finished houses are worth, less what it costs to build them, less fees and finance and selling costs, less the profit the builder needs in order to take the risk. Land is what is left over.
There are two ways to hold it. Buy it outright, which puts the money out and the asset on the balance sheet. Or take an option: the right to purchase a field if permission is granted, at a price set by formula against open market value when it is exercised. An option keeps cash in the business and it is not free, because the fee is gone whatever happens and the years of promoting the land through the planning system are yours.
Land with permission is the scarce input, which is why builders hold land banks measured in years of output rather than in months.
The boom made every number look better than it was
House prices rose steeply through the second half of the 1980s and fastest in the South East. Credit was easy. Building societies had been freed to compete with the clearing banks for lending, and a household's borrowing capacity was rising alongside its neighbours' selling prices.
For a builder that produced four flattering effects at once.
- Stock revalued while it sat, so being slow stopped looking like a cost.
- The bank lent against the revaluation, so gearing improved without a single house being sold.
- Reservations arrived long before completion, which made the forecast look settled.
- Paying above yesterday's price for a field looked rational, because the finished houses would be worth more than today's.
Three of those four are paper. Only one of them is money.
A builder is not paid when a buyer reserves, and not when contracts are exchanged. He is paid on legal completion, when the buyer's solicitor sends the money. Everything before that is a forecast, and a forecast does not pay a groundworker.
Base rate went from seven and a half per cent to fifteen
Rates started rising in the summer of 1988. From seven and a half per cent in May 1988, base rate reached fifteen per cent by October 1989 and stayed there for a year. What that does to a housebuilder happens in three places, and they arrive in order.
The buyer goes first. A purchaser's borrowing capacity is not set by the price of houses, it is set by the monthly payment their income will stand. Double the interest rate and the same person supports a much smaller loan, and that happens to every buyer at once.
The site goes second. Cash keeps going out. Roads, drains and services go in before a single plot is saleable, site management is there whether the show home sees anybody or not, and a half-built estate cannot be mothballed cheaply.
The facility goes third and finishes it. Development finance is drawn against land and work in progress, repaid out of completions, priced over base and covenanted on value. All four of those terms moved against the builder in the same eighteen months. Borrowing at base plus three went from ten and a half per cent to eighteen without a decision being taken. Completions slowed, so repayments slowed. Valuations fell, so cover fell. A bank looking at falling cover and slowing repayments does what banks do.
Land value falls faster than house prices do
Here is the arithmetic that caught the trade, on round numbers rather than real ones.
Say a finished house sells for £80,000. Build cost, fees, finance and selling costs come to £58,000, and the builder needs £12,000 of profit to justify the risk. The land is worth £10,000 a plot, because that is what is left.
Now take fifteen per cent off the house. It sells for £68,000. Build cost has not fallen by fifteen per cent, because bricklayers and bricks do not reprice in step with the housing market. If the builder still needs his profit the land is worth nothing at all, and if he gives up the profit it is worth very little.
That is why land values collapse in a downturn rather than drift down, and it is why land bought at the top is what finishes housebuilders rather than the houses themselves. The plots sit on the books at cost and cannot be built out profitably at the new prices. A builder can stop and carry the interest, or build and sell at a loss, and most did some of each. Across the trade, land buying stopped almost completely, values were written down, and part exchange and sales incentives appeared because the alternative was an empty site with a facility running on it.
The firms that came through the early 1990s were usually the ones that had let their land banks shorten in 1988 and 1989, taken the cash and looked timid while everybody else was buying. At the top of a boom the correct decision is indistinguishable from a failure of nerve, and you find out which it was about two years later.
Where this shows up in the game
Buy Land, They Said starts in Reading in January 1988 with a housebuilder, Kennet Homes, and runs for three years at hard difficulty. It is scored on the founder's net worth at the end, stated in today's money, against a target of £900,000. Land held at the top and still unsold at the bottom is worth what somebody will pay for it in 1991, not what it cost.
Nothing in it is scripted, which is the point of this one. The scenario is a date and a business that borrows. The real Bank Rate series already carries the rise, so nothing warns you in the summer of 1988 any more than anything warned anybody then. Land you buy is held at cost and sold at the market of the month. The overdraft is priced at base plus three, so the interest bill triples without a decision being taken, and the usual way this one ends badly is the bank calling the facility in 1990.
The seed is fixed at 600007, so everybody attempting it meets the same Britain and the scores mean the same thing. It runs under ironman rules: one save, and what happens happens.