Goldman Sachs is considering whether to defer bonuses for its employees into the new tax year, starting 6 April, in order to avoid the 50p tax rate.

The Guardian's Jill Treanor reports [2]:

A number of banks are known to have considered whether to make the move, which would save their top employees thousands of pounds. But City sources believe many of them have rejected the idea to avoid any negative publicity in the wake of the row surrounding corporation tax paid by Starbucks in the UK.

The Wall Street firm – which publishes its full year results on Wednesday and tells staff their bonuses for 2012 shortly afterwards – is not thought to be considering changing the way the bonuses for 2012 are handed out. The proposal being considered would benefit parts of bonuses deferred from the years 2009, 2010 and 2011, which are due to be handed to staff this year in the form of shares.

The move underlines the lack of evidence available that the 50p rate actually hurt revenues. HMRC's analysis in March last year determined that the optimal tax rate was 48 per cent [3], a figure which 1) didn't justify cutting the rate to 45 per cent and 2) was only derived due to a specific statistic – TIE, taxable income elasticity – being given a value of 0.45. Given studies cited by HMRC for TIE showed it being anywhere from -0.6 to 2.75, there's rather a lot of uncertainty in that analysis.

But the bigger problem for evidence of the tax cut's effects is that, in cutting the 50p rate so rapidly, the Chancellor destroyed the possibility that we might actually get some useful data. The most effective way to avoid the tax is to shift income forward or backward. As a result, the first year it was in operation revealed that £6.6bn of taxable income had been shifted forward by a year; and we now know that this year, the last it will be in operation, a significant chink of income will be shifted back to the 2013/14 tax year.

Add in the fact that even HMRC assumed that some income in 2011/12 will have been declared in 2009/10, and some more will have been coincidentally forestalled to 2012/13 (when it could be forestalled further to 2013/14), and it is clear: there has not been a single year when a "normal" amount of tax was paid at the 50p rate. Every year it was in operation will have resulted in an artificially depressed take.

Similarly, the 45p rate will, for the first few years of its operation, have an artificially boosted take. It will look far more effective at discouraging tax avoidance than it actually is.

Consider this a warning, then: 2014 will see a lot of attempts to misuse data to prove a point. Don't take it at face value.