In the course of a discussion on a previous diary on budget deficits, I mentioned a specific CBO failure to predict increased revenues as a result of a tax reduction. The CBO has generally low-balled its revenue forecasts for the last four years. I thought this specific case of prediction failure may shed light on what is going on with these things, and as a warning to take all such predictions with a sack or two of salt.

This particular case is interesting as it can be traced to a specific provision in a specific piece of legislation, the reduction in the Long-Term Capital Gains tax rate to 15% (down from 20%), which is, interestingly, set to expire in 2008, part of the Jobs and Growth Tax Relief and Reconciliation Act (JGTRRA) of 2003. I think there will be some argument about this soon enough.

Specifically,

Jan 2004 Capital Gains Tax revenue projections, in billions

http://www.cbo.gov/showdoc.cfm?index=4985&sequence=5

2002 - 57
2003 - 45 - Change in rates, May 2003
2004 - 44
2005 - 49
2006 - 54

They were predicting an actual reduction in Capital Gains tax revenues as a result of the tax cut, various numbers being touted by various sources, from $5.4B quoted by some, to perhaps $45B 2003-2006, by the CBO, based on its earlier 2003 forecast, to some rather incredible figures in the JCT "Blue Book" - http://www.house.gov/jct/s-8-03.pdf , see p.22.

Note also this summary, by some leftish activists, who seem eager to swallow such estimates -

http://www.cbpp.org/3-10-05tax.htm

"The Joint Committee on Taxation estimates that the dividend and capital gains tax cuts, which are scheduled to be in effect through 2008, will cost $148 billion. Making these tax cuts permanent would reduce revenues by an additional $148 billion through 2015, according to Congressional Budget Office estimates. Because this lost revenue increases the deficit, an additional $110 billion in costs will be incurred through 2015 for higher interest payments on the debt. In total, these tax cuts, if extended, thus would add $405 billion to the debt from the time they were enacted in 2003 through 2015. "

Actuals -

http://mirror1.cbo.gov/ftpdocs/77xx/doc7731/01-24-BudgetOutlook.pdf

2002 - 58
2003 - 50 - Change in rates, May 2003
2004 - 61
2005 - 84
2006 - 103

i.e., by my take Capital Gains taxes yielded $101B more through 2006 than the CBO initially predicted post-tax reduction, or $66B more than the CBO predicted earlier in 2003 for a no-cut scenario.

This low prediction was, I am speculating, but with some confidence, based on a conventional wisdom that tax cuts do not yield revenue increases approaching the revenue reductions, i.e., that "clawback" is limited to maybe 25%. In this case "clawback" proved to be more like 224%.

Mr. Gabriel was good enough to post one such study, or rather a thought experiment in the application of various revenue models - http://www.cbo.gov/ftpdocs/69xx/doc6908/12-01-10PercentTaxCut.pdf . This one is interesting as it includes Capital Gains in the analysis.

Some other people seem to have noticed -

http://www.americanshareholders.com/blog/2006/01/cap-gains-tax-cut-more-revenue.php

Note that they quote numbers from various dates, their latest numbers are higher than mine, perhaps they have later data.

Somebody should be eating crow. And it would seem rather difficult to make a case for Capital Gains rates to be raised in 2009, as they seem to be closer to a revenue-maximizing optimum at this rate than the previous one, as collections are approaching the FY2000 "bubble" period though we are not really experiencing an equity bubble.

It would be interesting to see the effect of the other major provision in the JGTRRA, on treatment of dividends, if this can be extracted. I suspect we may have another winner here also.
--