The House of Representatives approved a resolution changing the House Rules to require dynamic scoring for large tax and spending bills.  The resolution contains a disturbing provision that may well transform the Congressional Budget Office (CBO), long the last bastion of independent public policy analysis for the federal government, into a hapless tool of the House Leadership and a few committee chairpersons.

The resolution requires dynamic scoring for all tax and spending bills greater than 0.25% of US GDP.  US GDP stands just south of $17 trillion.  Thus, the CBO will be required to estimate the economic feedbacks for all bills with a direct impact greater than $42.5 billion.  While I remain unconvinced this is the proper way to analyze the fiscal impact of federal legislation, this provision alone would not be that onerous.  In fact, the principal advocates of dynamic scoring should be careful what they wish for.

Problems will arise due to a provision in the resolution that will inherently yield fraudulent scoring in the aggregate.  The provision requires dynamic scoring on smaller bills with fiscal impact if they are deemed important by the Chairmen of the Joint Committee on Taxation (JCT) and the House Budget Committee, both of which are now controlled by a single party.

First, it’s important to understand what dynamic scoring is.  Dynamic scoring, as applied in this resolution, requires the economic feedbacks of certain legislation to be counted.  The CBO already takes into account behavioral changes triggered by legislation.  For example, if the tax on cigarettes is increased, it takes into account a reduction in consumption.  Recent controversial examples of this were the observed when the CBO took into account direct behavioral impacts associated with the Affordable Care Act (ACA), which pleased Democrats, and the direct behavioral impacts associated with Minimum Wage proposals, which pleased Republicans.  Similarly, Iowa’s CBO equivalent, the Legislative Services Agency (LSA), takes into account the impact caused by federal income tax bills due to the fact that federal taxes are deductible from Iowa taxable income.

However, neither the CBO nor the LSA takes into account the derivative impacts caused by the spending/saving/investing behavior that inevitably occur when consumers have more or less money in their pocket.  The CBO did not score the impact of the legislation under both Bush and Obama related to the Troubled Asset Relief Program (TARP) and other counter-cyclical bills.  Had they done so, the bills would surely have had a negative cost.  The CBO would have been charged not just with considering the economic impact of the “base case” (eg. 25% unemployment), but with directly estimating the impact on federal spending (which would have been in much higher demand) and federal revenues (which would have plummeted).  Any given reader may not agree with that assessment, but I assure you any independent analyst at the CBO would have seen it that way.

Likewise, CBO scoring does not take into account the feedback that occurs when the payroll taxes were reduced.  Consumers had more money in their pockets.  The CBO didn’t estimate how much of that would be spent, thereby increasing corporate profits and improving taxes derived from labor.  They didn’t estimate how much would be saved, which would have led to marginally higher interest/dividend/capital gain income taxes.  Nor does the CBO estimate the derivative impacts from Pell Grants, which find their way into the pockets of professors and university bureaucrats, all of whom pay taxes on that income.  The CBO doesn’t take into account the long run fiscal impact associated with a more educated labor force.

As far as it goes, in theory, dynamic scoring is alluring.  After all, virtually every single spending bill that is not a direct transfer of money from one citizen to another (eg. Social Security) will cost less than projected, right?  That is to say, if people are better educated and professors make more money, then the impact of Pell Grants will be less than the amount originally appropriated.  Road projects have similar high return derivative impacts on GDP and federal revenues.

And if that is true, it is also true that most tax cuts will be scored to appear to cost more than they do.  If 20 cents on the dollar makes its way back to the federal treasury, then a billion dollars of tax cuts will only cost $800 million. Right?

So then what could possibly go wrong?

The answer lies in the base case scenario.  That’s what’s required to analyze according to a “balanced budget” analysis.  In other words, the CBO must assume the money going out in tax cuts or spending does not fall magically from the sky.  It comes from somewhere.  What happens to money that is neither appropriated nor returned by way of tax cuts?  Those funds reduce the deficit.  CBO would then be required to estimate the impact on long term interest rates from higher or lower deficits.  Or perhaps they could take a different approach.  They could hold deficit assumptions constant, and assume that tax cuts would lead to spending reductions or that spending increases would result in higher future taxes.  It’s an endless cycle of iterations that no human being I know is capable of estimating.  It’s far more complex than climate models, which supporters of dynamic scoring tend to loathe.  (To my knowledge, the CBO doesn’t use climate models either.)

This entire discussion thus far is largely academic.  Republicans in the House do not want dynamic scoring.  They want their tax cut bills to be cheaper.  Only 6 agencies in the Federal Government even have a discretionary appropriation over the 0.25% of GDP threshold.  None of the benefits from those agencies will be dynamically scored.  Not from HUD spending, Agriculture, Transportation, NASA, or 7 other agencies.  The taxes paid by those federal employees will not be counted as coming back into the treasury.  Their spending in the economy, and the derivative impacts that spending has on federal tax revenues will not be counted.  They will be scored as to their direct cost.  Even Pell Grants only come in at half the size required to trigger dynamic scoring.

But think back to the special provision in the House Resolution.  CBO will also be required to provide dynamic scoring on any bill where the House/Senate Leadership asks for it.  If the leadership asked for dynamic scoring on such things as Pell Grants, K-12 spending, infrastructure improvements, and energy efficiency spending, we would find them all to be less expensive than the static approach used for the past 40 years.  Will they do that?  This House resolution is not about making sure that competing ideas are scored in a manner that yields the best economic results.  It’s designed to ensure that tax cuts are scored preferentially to all other competing ideas.  Even worse, it’s designed to ensure that any spending or tax cuts preferred by two members of Congress costs less than any other spending.

This would be just as bad an idea if Democrats controlled both chambers.  Tax increases do indeed have negative consequences.  Tax cuts have positive economic impacts.  Education and infrastructure spending also have positive impacts, but they shouldn’t be scored in a manner that distorts the benefits relative to their cost.  That’s what the CBO was created to prevent.

CBO has been staffed by career public policy analysts and economists working in an honest and bipartisan manner to act as a check against a single powerful person with parochial political interests.  It will be interesting to see how they continue to serve that central purpose.  I don’t see how it will be possible.  Either measure economic feedbacks on all bills with fiscal impact, or do so on none of them.  To dynamically score only the bills the Leadership wants to pass is shameless, fraudulent, and completely unsurprising.