Computer models mess up Congress -

The US Congress has decided to blame the fact it is as useful as a chocolate teapot when it comes to predicting the economic impact of its policies, because it relies on out of date computer models.

For decades the official forecasters at the Congressional Budget Office and Joint Committee on Taxation have assumed that changes in tax rates have little impact on how businesses and households behave or on the competitiveness of the US economy.

We are not sure how it reached this conclusion, as this is the sort of things they teach in sixth form economics in most civilised countries.

However, in the US computers were told that people work nearly as much at a 60 percent income tax rate as they do with a 30 percent rate. If capital gains taxes are raised to 30 percent from 15 percent, investors would not care at all.

According to the Wall Street Journal, all this led to crazy results. In 2003 the computers spat out a prediction that capital gains revenues would be $68 billion in 2006 and $73 billion in 2007.

In May that year Congress cut the capital gains tax rate to 15 percent from 20 percent yet the computers insisted that the rate cut would reduce revenues to $65 billion in 2006 and $69 billion in 2007.

Actual capital gains revenue rose despite the lower tax rate to $109 billion in 2006 and $126 billion in 2007. The fact that capital gains tax was reduced boosted economic growth and meant that there was a greater incentive for investors to cash in their gains at the lower rate.

Now it has been decided to change the former to something a little more dynamic.

The knowledge that it was all wrong has been around for ages. It was just a matter of understanding if a "dynamic response" to the economy was worth factoring in.

In 2005, Harvard economists Greg Mankiw and Matthew Weinzierl looked at the revenue feedback of tax cuts and found that "the dynamic response of the economy to tax changes is too large to be ignored". They pointed out that tax cuts were partly self-financing and this needed to be part of the formulas in Senate computers.

If they can play around with these sorts of figures on the computer they might be less likely to make such howling mistakes. For example Republicans like Eric Cantor, the House Majority Leader have been claiming that tax credits for having children that cost the most in lost revenue do the least for economic growth.

In fact the computers would now tell him otherwise.

Still, both the Democrats and the Republics have voted to change the computer model formulas so in the future he will have no excuse.