Both right on Fair Tax Mark

Last night, at the Mazars Fair Tax Mark debate, I challenged Richard Murphy on the way his original mark calculated an average tax rate, and asked for confirmation of how the new one worked.

The example I gave (altered  a little here to make the point more clearly) was what average tax rate he would calculate for a company that paid no tax on £1m profits in each of the first five years, and then £150 tax on £100 profit in the sixth year.

Would he say that it was £100 tax on just over £6m profits, and so as close to zero as made no difference, or would he average the six tax rates of 0%, 0%, 0%, 0%, 0% and 150% to come up with 25%?

He said the former, and that the original Fair Tax Mark (which was launched last year and then withdrawn for consultation) had also used this method. I questioned that, knowing that it had been one of the main issues I and others had had with the original mark.

The Fair Tax Mark site no longer has the original results, but the one for Thorntons plc is still available on Ben Saunders' site. I would argue that it shows we were both right, but that I was righter than Richard...

Richard is right that the average tax rate is calculated by taking the average current tax charge over the average profits earned. This is then compared with the average headline rate for the six years and the difference calculated. In Thorntons case, it paid an average of £1.4m on £4.4m profits, an average rate of 33.5%. This is compared to an average headline rate of 28.1% giving a difference of -5.4%: Thorntons paid more tax than expected.

However, I am "righter" than Richard because in most cases that was not the rate that was used to assess the tax rate. Instead, a weighted average of the differences was used.

The difference from the headline rate for each year was calculated (shown on the line "Difference"), these were weighted arbitrarily with the most recent year counting six times more than the oldest, and then the average of these weighted differences calculated.

The line "Weighted rate difference" total of 24.6% is the total of the six weighted figures for the individual years divided by the total of the weights, 21. 

It is also obviously meaningless, implying the company had substantially underpaid tax when it had not. So in this case the old Fair Tax Mark used the average tax rate rather than the weighted average tax rate. But in most of the cases it looked at, it used the weighted average tax rate.

The weighted average tax rate in my example would have been 6 x 150%/21, 42.9%, for a company that has actually paid £150 on just over £6m profits.

So the methodology previously used was wrong, not just because of the weighting but also because it was an "average of the averages".

Fortunately, the abandoning of the weighting in the new tax mark also means that the averrage tax rate is the average tax over the average profit, as it should be.