It might be argued that the whole of the Brexit debacle started and ended with the subject of immigration. There are too many people in the country, we heard, too many still arriving, we heard, and we therefore need to ‘take back control of our borders’. So went the cries from Brexiteers.
Politics aside, what concerns me here is what does the data say? If anyone is going to make a decision on the subject, it is a fairly safe to say that having meaningful data would be a fairly important first step. We established PanSensic as way to help enterprises to measure what matters rather than what was merely convenient. As much as anything it has now become a way of innovating in the measurement domain: working out what we really need to know, and then working out how to measure it.
This is pretty much the opposite of the philosophy adopted by others in the measurement business. Some of them seem to have veered so far towards the ‘what’s easiest’ end of the spectrum it’s a wonder they can still see any kind of useful purpose in what they end up delivering. One such organisation here in the UK is the Office Of National Statistics. These are the people tasked by the Government with providing the information upon which policies like Immigration are determined. Being in effect a government agency (they were made ‘independent’ a decade ago), they have their own special version of ‘what’s easiest’ and that is, ‘whatever way we’ve always used’. Change, in other words, does not make up a significant part of the DNA of the people that work there. I know this from having had the occasional opportunity to work with their Senior Leadership Team.
‘Whatever way we’ve always used’ usually turns out to either employing an army of people to conduct telephone surveys (this is how a lot of VAT collection calculations are made, for example) or people with clipboards meeting members of the public face to face. This is the strategy used, it was revealed last week, to gather the country’s immigration data. One of the main characteristics of either of these methods is they are very expensive. Another is that, because people are generally speaking sick to death of being cold called or accosted by people with clipboards in the street, they tend to, ahem, not tell the truth, the whole truth and nothing but the truth. (See an earlier blog on our ‘5G’ model.) So, what we end up with is a method that is simultaneously the most expensive and the most shit. Excellent.
Not only are the UK’s immigration figures based on measurements taken by clip-board carrying ONS personnel, it also now transpires that they make the measurements at precisely two airports and zero other entry points on our borders. The two airports are Heathrow and Gatwick. And then somehow – by mathematical chi-squared unicorn-dust – they are able to extrapolate from these two clipboards-worth of data to produce national figures. To six significant figures. More excellent.
If you were setting out to make the least accurate, least meaningful way of calculating immigration, I suspect you would be hard pushed to do it. Unless you were taking Class-A hallucinogenic drugs. Which to my mind I’m not totally sure can be precluded from the discussion in this instance.
Having had this survey method ‘revealed’ last week, the UK Government has been forced to respond by downgrading the quality of the results produced to ‘experimental’. Class-A drugs might also have been involved here too in so far as the committee that opted for that word don’t appear to have had much more of a clue what they were doing than the clipboard bearers.
Experimental implies that, at the very least, you sat down and thought about the problem and then designed something. Key words: ‘thought’ and ‘designed’. For a start, thinking about the 28 airports and 11 ports that receive international passengers, and the destinations that people are able to travel from, and then thinking about the questions you might ask people to elicit a true and meaningful answer rather than the one they think you would like to hear.
But then, I think, pretty soon you’d conclude that, hmm, this all sounds like a lot of hard work. So then you start to wonder whether there might be other ways to work out how many people are living in or leaving the country. Something far easier. Things you’re already measuring, maybe. On Social Media, for example, or mobile phone signals. All the time knowing that you don’t need to calculate absolute numbers (the census will do that for you), merely relative ones. When I sat down and spent half an hour thinking about it, I concluded I could – with only the public domain data available to a shlub like me – design, build and commission a more accurate system than the current ONS effort in about three weeks. Step One: sack all Oxbridge-educated ONS statisticians (about 90% of the total my BS-detector suspects); Step Two: teach the rest Complex Systems 101 and Human First Principles 101; Step Three: sack them too. Step Four: Phone Google.