With hundreds of thousands of people employed in part-time and zero-hour contract jobs, along with many more registered as self-employed and others in various kinds of training schemes, the words 'employment' and 'unemployment' are misleading.
In truth, the situation in any country or locality is that there are people with various kinds of employability and needs. To calculate the real state of 'unemployment' some other system would provide a true snapshot or change across time. To do that, you have to detach 'employment' from individuals, and think of it as a total amount of hours.
I don't have the data but to do that, but you would have to
1. Calculate the total number of hours available in a given population, given present employment laws.
2. I guess that would be taken by multiplying the number of people in the population aged between 16 and 65 by 40 hours. This would give a total 'Hours'.
3. Subtract from 'Hours', the number of those in full-time education, disabled, not needing to work because they have some other method of supporting themselves other than by being employed and multiplying that by 40 hours.
4. Take all those on part-time and zero-hour contracts who would want to work full-time and calculate the aggregate hours they are not working and subtract that from 'Hours'.
5. Take all those who are self-employed, calculate how many more earning hours they would want to work and subtract that from 'Hours'.
6. Any other hours I've left out from this list (!) and subtract that from 'Hours'.
Now you would have two figures - total theoretical number of 'Hours' that people could work. And the actual 'Hours-not-being-worked' - let's call that 'Hours nbw'.
Over time, these two figures would change but there is always a ratio between them.
A true gauge of 'unemployment' would tell us whether 'Hours nbw' was going up or down.
I notice that people like Danny Blanchflower are mocked by the Right because he predicted that unemployment would go up with austerity. 'It hasn't, it hasn't!' cry the pro-austerity folks. But they can only do that if unemployment is calculated as attached to individuals 'in work'. But hundreds of thousands of people are not 'fully' in work and want or need to work more hours.
So, perhaps - we don't know - Blanchflower et al were and are right. My 'Hours nbw' might have gone up over the last 7 years.
Does anyone know? Has anyone done these sums? Do we have a true picture of what's been going on in the last 7 years?