Today we publish Riders on the Storm our new report on middle income households. But you know their story, right? The operative verb is to squeeze, or to be squeezed. Yet the data we've used - a panel survey conducted by researchers from the University of Essex for the past 20 years - shows something different.
The middle 20 per cent of working age households in 2011-12 (the latest survey data available) had on average the same income in real terms as four years ago. In other words, even in the teeth of the greatest recession in a century, their income wasn't squeezed, though it did stop increasing. It gets better, literally. Start at the other end in 2007-08 and two fifths of the middle income households moved up into the 40 per cent of the distribution that was above them. Around the same number stayed where they were in the middle 20 per cent. The rest moved down.
But how can this be? It's been proved definitively that real incomes are falling. The reason is that most work on these issues compares snapshots taken at different times except they are snapshots without the same people in them. We, too, found that the middle in 2011 had lower incomes than the middle in 2007. But those two snapshots of the middle don't contain the same people. As I said in the paragraph before, only around two fifths of the middle stayed in the middle over those four years.
To compare the two different snapshots is like looking at family photos and saying "Wow, auntie has really changed", except in the time between when the two photos were taken your uncle has divorced and remarried. This is why we chose to look at what happens if you keep the same people in the snapshot.
There are other problems with the typical approaches in this area too. Take this series of incomes: 3, 3, 5, 7, 20. Some of the work done by others focuses on the median income. The median in this series is 5. Imagine that we leave this series alone for four years and then come back to it. The person on 7 has retired, to be replaced by a much younger person earning 4. The new series is: 3, 3, 4, 5, 20. Suddenly the median has dropped from 5 to 4 even though no one got poorer. And, by the way, the median falls even if the person earning 5 before has popped up to 6, 7 or even 19.
Obviously it's still significant that the median is lower. But it's lower in my example because of a demographic change not because the middle was squeezed.
Right, enough of the thought experiment. Back to our findings. Track the same people and you find that their incomes haven't been squeezed. Yet they are stagnant. Despite being four years' older and despite, as our data also shows, slightly more of them being in work. So we're not getting carried away. The other reason not to put out the bunting is that the bottom 20 per cent, in particular, had a really bad time. Their level of employment had fallen. As many as a quarter were behind on their rent or mortgage.
Squeezed middle? Perhaps not. Smacked bottom? Yes, certainly.
Emran Mian is director of the Social Market Foundation