Here is a chart illustrating some data. Watch me say the exact opposite

The Adam Smith Institute is trying to make itself right about rail privatisation with this chart:

GBR_rail_passenegers_by_year1

Because it goes up and to the right! And if you forget the whole 14 year bit from 1981 to 1995 where it was going up and to the right, you can claim this is down to railway privatisation in 1996! By 1987, ridership was back to the levels of the late sixties. But surely no-one would engage in such shameless cherry-picking, would they? Yeah, of course they would at the thinktank still rated dead last for transparency. Forget trains, this guy should show us his time machine.

If I had to guess I’d say that it’s no coincidence that the ridership series turns around and starts growing in 1981, as exactly the same thing happened to the population of London at exactly the same time. The lowest population recorded was in the census of 1981. Similarly, the previous peak was the census of 1951, and 1961’s census recorded a significant fall. The peak year for postwar ridership was 1955, i.e. exactly halfway between the two, as if it was actually measuring population growth.

(He does make a feeble effort to disaggregate London from national data, but the data series he uses doesn’t start until 1997/8 so it’s basically irrelevant to this point.)

Interestingly, if you were to extrapolate the trend from the 1980s, in a counterfactual world where Tory Recession 2.0 didn’t happen, it would look spookily similar to the actual, just shifted left by about five years, as if something was bound to happen and was only delayed by the recession (or perhaps the preceding housing bubble). Manchester’s population fell rapidly from 1931, the peak year, to the mid-1970s and levelled off by 1981, before something caused it to fall further (looking at you, ASI), eventually turning around in 1991. Birmingham peaked in 1951, declined, and stabilised around 1981 although again, some sort of event around 1991 seems observable (looking at you, ASI) before it began to grow again. The capital of the North peaked as late as 1971, declined to 1981, and then began growing again. Cardiff had the great bulk of the decline between ’71 and ’81. Glasgow peaked in the late 40s and didn’t turn around until 2001 (and again, ASI…). Bristol peaked in 1971 and bottomed-out in 1981. Belfast peaked in 1951 and declined all the way to 2006 but there were reasons to avoid living there for most of that time, almost 2,600 of them. Edinburgh peaked in the mid-60s, declined, and recovered from…you guessed it, 1981. Clearly, there was some sort of big change regarding suburbs and cities around the time I was born.

The lesson from this chart is surely that we ought to be a bit humble about ascribing consequences to our decisions. There was some conscious policy input to the shrinkage of London in the postwar era via the new towns programme. But there wasn’t a deliberate policy to halt suburbanisation in the 80s (far from it! remember the great car economy?), and weirdly, the drift back to the cities began all over the industrialised world at roughly the same time, rather like the fall in violent crime. Much like the fall in violent crime, nobody really knows why. But it should be absolutely no surprise that railway ridership went up when people flocked back into British cities built mostly around rail transport.

12 Comments on "Here is a chart illustrating some data. Watch me say the exact opposite"


  1. The big growth in rail passenger numbers is in the London and South-east sector, which has a lot to do with the growth of the London economy, high house prices near jobs and the lack of alternative means of transport.

    A secondary reason is that overall people are travelling longer distances more often and rail has captured part of this growth in personal mobility. Rail didn’t benefit from this in the 1960s because there was enough road capacity to allow a switch to personal motor transport and a growth in mobility at the same time. This is not longer the case. Roads are increasingly congested and (most importantly) are very congested in the bigger cities. There is a lot of driving to stations and travelling by train to Edinburgh/Glasgow/Leeds/Manchester etc.

    There has been some improvement in passenger train services in the last 20 – 30 years, but a lot of this is due to there being fewer parcels and freight services so there are more paths for passenger services.

    Reply

  2. Hang on a minute, isn’t that graph fundamentally dishonest? I can see nothing about population, as one of the commenters mentions.
    IN 1911 there were 36.1 million in England and wales, say a couple in scotland so 38 million, which with 1.5 billion passenger journeys is 39.474 journeys per person.
    By 2011 population is 61.1 million, which with the same 1.5 billion passenger journeys is 24.55 journeys per person. So the per head popularity is lower than it was, but the graph gives the first impression that it is as high as it was.
    Or in 1971 the population was around 52.5 million, and say 750 million journeys, meaning 14.368 per person.
    The funny thing is that their list of charts doesn’t prove anything except the need to travel about the country. People are opposed to rail privatisation because it’s expensive and doesn’t benefit the passengers, not because they hate to see more people on the train. All the benefits touted for passengers are exactly as you would expect in any situation with increased investment (e.g. the NHS). Safety improvements are a matter of technology, not privatisation.
    And as for customer satisfaction, I can tell you that here in Scotland we hate Worstrail, or Scotfail, but people don’t see that complaining will make a difference. The staff are alright, but they’d be alright anyway were it a public organisation, the issue is the profits sucked out of it and the pathetic ability of the management to plan ahead for events (A rugby match, who’d guess people would travel by train to it?).

    Reply

    1. Not sure about normalising by population; trains don’t carry passengers as a percentage of population, they just carry passengers. The fact it’s not as many on a per-head basis doesn’t change how much capacity the railways need.

      Reply

      1. I was thinking more simply that if your population is larger you’ve more people wanting to or having to travel, (Not to mention bidding up house prices) which will lead to increased numbers of journeys anyway. If you deported 10 million people from Britain I guarantee that rail journeys would drop, for lots of reasons including just fewer people. And then implicitly comparing 1911 to now as the chart does, when there’s a lot less people to travel and no useful alternative means, isn’t very helpful at all.

        Reply

          1. Of course, so obviously that proves that nationalisation drives down passenger numbers…


  3. the pathetic ability of the management to plan ahead for events (A rugby match, who’d guess people would travel by train to it?).

    I am trying to remember who started a letter of complaint to the railway company with “Sir. Monday morning, although recurring regularly at easily predictable intervals, always seems to catch this railway by surprise…”

    Reply

  4. Yes Ajay, that’s the sort of thing I was thinking of.
    It’s a bit similar with the Royal Mail and christmas, but having done a temp stint in December in a sorting office/ warehouse, I have some sympathy for them because of the massive increase in maill volume of the time. They could buy in enough equipment to deal with it all as fast as normal, but it would be lying useless for 11 months of the year and because the governments don’t like that they can’t spend the money.
    The rugby example is easier though; they usually have shorter trains at the weekend, and it wouldn’t be that hard to increase their length to weekday commuter standards on the ones going too and from the affected town.

    Reply

  5. Interesting thought. There’s a very clear macroeconomic signal in the data; you can see Churchill forcing the UK onto gold in 1925, you can see the Great Depression, you can see the beginning of WW2 and perhaps also the OVERLORD buildup, you can see post-war demobilisation, there’s a kick as we leave austerity behind.

    Then there’s a secular decline which dominates everything else. Until the business cycle comes back. Ridership grows from 1974 to 1978, i.e, following recovery from the 1973/4 recession, and then you can clearly see Tory Recession 1, the 80s HST/Network Southeast boom, Tory Recession 2, and the long bull run.

    I said earlier that if you were to extrapolate the trend of 1981-1989 you’d get something a lot like what actually happened, just 5 years early. But if you were to do the same with 1973-1978, you hit off the late 80s peak pretty precisely and arrive at the time of Hatfield at the same levels of ridership, as if it was following a nice Bass diffusion curve all the way from the oil shock and just got delayed by the recessions.

    You can’t see the GFC in there, although perhaps you might if the chart went beyond 2011. But then you can’t really see economic events in the secular-decline era either, nor in the industrial revolution S-curve; it’s as if it follows the business cycle unless something bigger drives it.

    Reply

    1. Then there’s a secular decline which dominates everything else. Until the business cycle comes back.

      People buying cars, until, essentially, everyone has a car who wants one?

      Reply

    1. Roger also says that InterCity spotted the 1989-92 recession in their data before most of the country was aware – strong growth in long distance rail suddenly stopped.

      The other thing the ASI prats neglect is that London is simultaneously the driver of most of the growth and the least typical bit of the rail network, not least because a fair chunk of it is publicly run or publicly directed (and it’s *Conservative Party policy*, if you can call Boris that, to increase the level of public direction of it by taking more suburban franchises under TfL control).

      Reply

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.