, Britain
On a February morning in 1996 a departure board in the north of England carried a company name that had not existed the year before. The trains were the same trains. The drivers were the same drivers, on the same terms, and the track underneath them had not moved an inch.
What had changed was the paperwork, and the paperwork was the entire business.
The first two passenger rail franchises began in February 1996. Almost everything people argued about for years afterwards was already present in the structure of those two contracts, written down before the first train ran.
One railway became a network of contracts
Under the Railways Act 1993 the single organisation that had run Britain's railway was separated into parts that then had to buy from and sell to each other.
Infrastructure, meaning track, signalling, stations and the timetable, went to a separate owner. Passenger services were let as franchises, for a term of years, by a franchising director acting for the government. Rolling stock was transferred to leasing companies and leased back to the operators. Maintenance and renewal became contracts with prices.
For the person running a train company the effect was total. Almost every relationship that had been an internal memo became an invoice with a penalty schedule attached. If a signal failed, that was no longer an operating problem inside one railway, it was a delay attributed to a party with money moving between them because of it. If you wanted different trains, you did not build them. You negotiated with the company that owned the ones you had.
The case for arranging it that way was that separating the parts makes the costs visible and lets competition happen where competition is possible. The case against was that a railway is one machine and the interfaces are where things break. Both turned out to be partly right, which is no comfort to an operator in year one.
The bid was a subsidy profile that fell every year
A franchise was won on price, and the price was the subsidy the bidder said it needed.
Bidders submitted a profile: a schedule of payments across the term, year by year. Competitive pressure pushed those profiles down, and it pushed them down hardest in the later years, because a bid promising to need less public money each year looked better and the later years were comfortably far away. Some franchises eventually crossed over into paying a premium rather than receiving support.
So an operator starts year one with a contractual commitment to run substantially the same railway in year four on less money. There are only three ways to meet a falling profile. Carry more passengers, charge more, or spend less.
Charging more was largely closed off. A basket of fares, the season tickets and the cheaper standard fares that most regular passengers actually bought, was regulated by a formula tied to inflation. Unregulated fares could move, but the regulated ones are the bulk of what a commuter railway takes.
Spending less was mostly closed off too, for reasons in the next section.
Which leaves passengers. Underneath all the documentation, franchise bids were ridership forecasts, and that is why an optimistic one was so dangerous. An extra passenger on a train that is running anyway costs almost nothing to carry, so growth falls straight to the bottom line. By exactly the same arithmetic, a shortfall comes straight off it. The trains run either way.
Almost every cost was fixed before the first train ran
Set out the cost base of a regional operator in 1996 and the striking thing is how little of it the operator could touch.
- Track access charges, paid to the infrastructure owner, largely fixed and regulated, payable whether the trains are full or empty.
- Rolling stock leases at fixed rentals, on vehicles the operator did not own, could not sell and could not hand back in the middle of the term. On a regional franchise a good deal of that fleet had been built decades earlier.
- Staff, transferred across on their existing terms.
- A minimum service specification written into the franchise agreement, so the quiet evening train cannot simply be withdrawn to save money.
That is most of the cost of a railway, committed before the doors opened, against revenue that moves with the weather, the local economy and the roadworks on the parallel trunk road. It is operating leverage in an extreme form. A small percentage shortfall in revenue is a very large percentage change in the result, and there is almost nothing to trim in response.
Fleet maintenance is one of the few levers, and it is a trap, because the trains are old and leased and their reliability drives the number that everybody else is reading.
One number gets read before the accounts do
Punctuality on the railway is measured as the share of trains arriving within a threshold of the booked time, five minutes for shorter distance services and ten for longer ones, with cancellations counted separately and treated as worse. It is reported by period rather than by calendar month, it is published, and the department reads it before anything else in the report.
Three things about that measure govern how the trade behaves.
It does not care whose fault it was. A train delayed by a signal failure on infrastructure the operator does not own counts against the operator's figure exactly as a late crew does. Money moves between the infrastructure owner and the operator for attributed delay, so the fault is settled in cash, but the published figure is still yours and the passenger on the platform has no interest whatever in the attribution.
It is a trailing figure, so it moves slowly in both directions. A bad period sits in the number for a long time, and a recovery takes just as long to show.
And it escalates. Sustained underperformance brings a requirement to produce a remedial plan, and continued failure after that puts the contract itself at risk.
The lesson generalises well beyond railways, to anybody trading under a contract that somebody else scores. You are running two businesses: the one that makes money and the one that produces the reported number. They overlap, but they are not the same business, and spending to protect the number is frequently correct even when it earns nothing. A spare unit and a driver standing at the north end of the route earn nothing at all on most days and are the cheapest punctuality anybody can buy.
Where this shows up in the game
Franchise Number One starts in Leeds in February 1996, the month the first franchises began, with a regional train operator, Aire Rail, and runs for three years. It is scored on profitable months, twenty out of thirty-six, at medium difficulty, and you begin with £2,000,000 of borrowing that somebody else agreed to.
The bid you have inherited behaves the way the real ones did. The subsidy profile falls each year of the run, exactly as bid, whatever kind of year it turns out to be. Track access charges and rolling stock lease payments are fixed and payable whatever happens. Infrastructure delays are outside your control and count against your figure anyway, which is the point of the scenario rather than a complaint about it. Two periods below the threshold triggers a remedial plan and four triggers a breach notice, so the ladder is visible from the first month.
The seed is fixed at 600010, so everybody attempting it gets the same signal failures, the same weather and the same Britain, and the scores can be compared directly. It runs under ironman rules: one save, and what happens happens.