The Office for National Statistics is working to improve its website which is great. One of the big advantages the American blogosphere has is the St Louis Federal Reserve’s data website FRED. It is much easier for them to find useful data. For example, I just spent 30 seconds making this graph. It’s the rate at which people quit their job:
This ease of data access makes it easy to tell stories. You can see the recession marked in grey and the recovery. But you can also see that for all the talk of success in the United States people are still more scared to quit their jobs than even the most pessimistic point of the mid-2000s expansion.
Easy access to data makes it easier to tell stories and to stop politicians and the media misleading you. If anyone says “optimism has returned to the US” I can point to this graph and say “eh, I don’t think so.” The ONS’s website is not easy to use but they’re working on making it easier. I wasn’t sure if I could make the above graph from the ONS data so I tried. Come on a journey with me in which I fail and end up in the pub.
I searched for “Quits” on the ONS which sent me to a 2003 pdf of Job Separations. The term “Voluntary job separation” is the ONS term for quits so I searched for that which sent me to this 2009 pdf on “Economic and Labour Market Review” from which I got this graph:
That’s not quite what I want. So I kept digging. About 10 minutes in now. But I’ve got a source Labour Force Survey, great! I google “Labour Market Survey” and go to this page:
Which isn’t particularly useful. If you search Labour Force Survey on the ONS website you get to a long list of search results but eventually get to this page titled “Labour Market Statistics, November 2013” and this page of 66 different data tables across (for no particular reason) three pages.]
I find the labour force flows tables under supplementary tables. It must have been 30 minutes now. I can’t find quits but I do have separations. Well I say that. What I really have is a combination of people moving from employment to unemployment and inactivity from one quarter to another and a bodge to turn that into this roughly analogous graph. It’s not quits, it’s a separations proxy, but it’s all I’ve produced.
Chris has a good post using the same data set here, showing that it’s completely doable to use the ONS website. It’s just not very much fun. What can I tell you from my graph? Not a lot, I can’t be bothered after all that work. I’m going to the pub.