In Defence of Liberty

Driven by data; ridden with liberty.

Statistics and Lampposts XI: Zero Sense

Evidence UK has claimed that there are one million more people on zero-hours contracts since 2010. (Photo: Simon Cummingham)

Evidence UK has claimed that there are one million more people on zero-hours contracts since 2010. (Photo: Simon Cummingham)

It stands that I must issue an immediate correction. I wrongly believed Evidence UK’s claim that there are one million more people on zero-hours contracts since 2010 was due to flawed comparisons. I was wrong: it is so much worse than that.

The author behind Evidence UK – most likely Dr Eoin Clarke – wrote:

The march [sic] publication by the ONS suggested just 168,000 workers were on Zero Hours Contracts in the last year of a Labour government (2010) but this data was drawn from workers self reporting [sic]. If we assume that the same ratio of under reporting occurred under Labour as the Conservatives, then we can perhaps roughly speculate that up to 409,000+ workers were on Zero Hours Contracts in the last year of a Labour government. Either way, the data shows the number of Zero Hours contracts have tripled under the Tories and grown by more than 1 million since 2010.

To say that these statistics are butchered is to insult the considerable care, skill and rigour of butchers. Firstly, the author conflates between the number of people on zero-hours contracts and the estimate of active no guaranteed-hours contracts, recently published by the Office for National Statistics (ONS).

Secondly, by sheer coincidence, the calculation for this scaling factor yields the lowest possible result. There was huge rise in employees self-reporting that they are on zero-hours contracts in 2013. The attributed figure leapt from 250,000 in 2012 to 583,000 in 2013. Dividing 1,420,000 by 583,000 obtains 2.44, the scaling factor. The author says the assumption that the scaling factor remains constant is “big”, but a better word would be untenable.

If this scaling factor remains constant, multiplying each year from 2010 to 2013 through by this scaling factor, this has two main consequences: it suggests there was no advancement in the number of workers on zero-hours contracts between 2013 and 2014, and it suggests the increase in such employees between 2012 and 2013 was about 810,000. The June 2013 edition of ONS Labour Market Statistics states employment grew by 432,000 between February-April 2012 and the same months in 2013. Evidence UK’s method implies the surge in zero-hours contracted employment alone in 2013 was nearly double that of the rise in all employment. This seems wrong.

It is acknowledged by the ONS that the massive attributive uplift between 2012 and 2013 may have been due to the media attention on zero-hours contracts:

However, ONS recognises that this question depends on employees correctly knowing their terms of employment, and that increased awareness of ZHCs among employees may have affected how people respond following increased media coverage in the latter half of 2013.

Lastly, the author confuses upper and lower estimates, as it is the upper estimate that is required to produce a confident estimate of the increase of these contracts from 2010.

It cannot be said that the “data shows” an upsurge of over one million of these contracts, nor is the source of this claim the Office for National Statistics.

Advertisements

Information

This entry was posted on June 13, 2014 by in Statistics and tagged , .
%d bloggers like this: