Should All Tax Cuts Be Paid For?

Click here for a printer-friendly version of this article

Should extensions of current tax policy have to be paid for or offset? Congressional Republicans don’t think so and point to the revenue growth after the enactment of previous tax relief to make their case. Democrats and their allies mock the notion that tax cuts “pay for themselves” by showing that the revenue growth after enactment of tax cuts lags the growth rate that would have otherwise occurred. The revenue lost from extending expiring tax relief needs to be offset, progressives argue, because it leaves the government with less money than it would have otherwise collected. Let’s dive a bit deeper into this debate.

First, it is important to point out the basic disconnect between budgetary rules that apply to tax relief relative to those that apply to what’s known as mandatory spending. As the Congressional Budget Office (CBO) explains, its baseline projections “assume continuance of a number of [mandatory] programs whose authorization expires within the current projection period.” The practical effect is that extensions of expiring spending programs do not have to be offset because CBO builds their extension into its baseline.

• For example, when Congress reauthorized the Farm Bill in 2008, it did not need to offset the entire $710 billion ten-year cost of the program. It only had to offset the incremental $9.6 billion increase in spending above the amount already assumed in the budget baseline.

Conversely, when tax cuts expire, CBO and the Joint Committee on Taxation (JCT) restore the revenues to the baseline in the years past the expiration date. This means that the entire cost of extending tax cuts must be offset. Had the “Farm Bill” been a $710 billion extension of an expiring tax cut package, it would have required $710 billion of offsets rather than just $9.6 billion.

Defenders of treating tax and spending policies differently note that the cost of tax relief can be artificially reduced thorough the use of expiration dates at the time the policy is first enacted. But, let’s use a tax relief example and then compare the treatment of adding a new direct spending program to the budget.

• If you assume that a new tax cut expires after five years, you have effectively halved the ten-year cost of the policy.
• Conversely, setting an expiration date in year five would not reduce the ten-year cost of new direct spending because the baseline would assume that the program would be reauthorized (in the five years following that expiration).
• The health care bill demonstrated that the same budgetary impact can be achieved by delaying the implementation date for the new spending. By having the policy start or take effect in year five, the 10-year cost of the new direct spending can be halved – and built into the future baseline at the same time.

The second theme to examine is that the progressivity of the current tax system means revenue grows faster than the economy over the long term. From 1946-2008, revenues averaged 17.8% of GDP.

Under CBO’s extended baseline scenario, federal tax revenues will rise from 14.9% of GDP in 2010 to 20.7% in 2020 and 23.3% in 2035 (assuming current law remains in place). In other words, if no tax relief is enacted, taxation will increase by more than 30% as a share of the economy (compared to the historical average) in less than 25 years. If you look at the change relative to today’s very low share (14.9%), the jump is more than 50%.

Although current policy discussions center on whether or not to extend current tax rates, the long-run revenue path is not much altered by the expiration of the 2001 and 2003 tax cuts. Even if every provision of the Bush tax cuts was permanently extended, tax collection as a share of the economy would still reach record levels. Donald Marron, using the same CBO tax revenue projections, explains why today’s low receipts-to-GDP ratio is projected to quickly spike back up, and approach the highest level on record (which was 20.9% of GDP back in 1944):

That rapid growth [that’s on the horizon for federal tax receipts] reflects six factors. First, the economy will recover, lifting revenues from currently depressed levels. Second, the 2001 and 2003 tax cuts will expire, as will tax cuts enacted in the 2009 stimulus. Third, the Alternative Minimum Tax, which is not indexed for inflation, will boost taxes for millions more taxpayers. Fourth, the new taxes that helped pay for the recent health legislation will go into effect. Fifth, retiring baby boomers will make more taxable withdrawals from tax-deferred retirement accounts. Finally, in a phenomenon known as bracket creep, growing incomes will push taxpayers into higher brackets and reduce their eligibility for various credits.

Together, those six factors will increase tax revenues by 8.4 percentage points of GDP over the next 25 years, according to CBO. About a third of that increase (2.7 percentage points) comes from expiring individual income tax provisions and the expansion of the AMT. Another third (2.6 percentage points) is due to real bracket creep and reduced credits. And about one-seventh (1.2 percentage points) results from the tax increases in the health legislation. The other factors account for the remainder.

It’s interesting to note that the expiring Bush tax cuts have about the same impact on projected tax revenues as bracket creep, where real growth in income pushes people into higher income tax brackets over time. Since the tax code is also indexed for inflation, any wage growth in excess of inflation results in a higher effective tax rate on labor income.

• For example, a household in the middle of the income distribution might face a 15% marginal income tax rate today, but would pay taxes at the 25% marginal rate in the future even if the 2001 tax relief was made permanent.

Tax revenues as a share of GDP have historically remained within a 16% to 20.9% range and have demonstrated no obvious trend over time. At a certain point, the projected tax increases under current law become unrealistic because revenues would rise well above the historic range, which assumes future generations would agree to endure a tax burden far in excess of that which past generations have been willing to bear. For this reason, prior to Peter Orszag’s tenure as Director, CBO’s annual Long Run Budget Outlook assumed revenues remained at historic averages (see page 43). CBO would estimate the path of revenues under current policies and then “assume enactment of a series of legislative changes that would keep receipts close to their historical average share of GDP.” In the long-run, whether the Bush tax cuts are extended or not, some future tax relief has to be assumed to keep revenues within the range of the past 60 years.

In 2000, federal revenues reached 20.6% of GDP, a level only exceeded at the height of World War II. The tax relief enacted in 2001 reduced the level of tax receipts and also slowed their growth. But tax revenues still grew substantially; leaving total tax receipts 27% higher in 2007 than they were in 2000 in nominal terms. After adjusting for inflation, revenues in 2007 were 5.3% greater than revenues in 2000 – a fact rarely mentioned in the current debates about tax policy.

The chart above uses actual tax receipts from 2000 to 2008 to measure the deficit that would have resulted in each of those years under three spending scenarios:

1. What actually occurred (blue line);
2. Real government spending remained constant starting in 2000 (red line); and
3. Real government spending remained constant starting in 2002 (green line).

Had government spending grown at just the rate of inflation over this period, the federal government would have run a $414 billion surplus in 2007. Allowing government spending to grow at three times the inflation rate from 2000 to 2002 (as actually occurred) and then slowing spending growth to the rate of inflation thereafter would have resulted in a $250 billion surplus in 2007. So what’s the lesson here?

• From 2003 (when the tax cuts were fully phased-in) to 2007, tax revenues grew 18% more on a cumulative basis than GDP. The result was a 2.3% increase in the receipts-to-GDP ratio to 18.5%, or 0.7% above the post-1945 average.

This makes discussions about whether tax cuts “pay for themselves” a bit more nuanced. If a tax cut “pays for itself” only when revenues remain constant as a share of GDP, then this is an almost impossible standard to meet.

After all, the effect of a marginal rate cut is to reduce the amount of revenue collected for a given level of income (which thereby flattens revenue growth). However, if a tax cut pays for itself when it results in higher real revenues, then the 2001 and 2003 tax cuts met and exceeded this standard (prior to the collapse of revenues during the recession).

Current budget rules disadvantage extensions of current tax policy relative to extensions of current spending. This is unfortunate because the exponential growth rate of tax receipts means that stabilizing taxes as a share of the economy requires a permanent extension of current tax policy and the enactment of future tax cuts.

Yet, despite claims that the 2001 and 2003 tax cuts starved the government of revenue, revenues still increased above their post-war average as a share of the economy (up to the recent crisis) and would have resulted in record surpluses had spending growth simply been tied to inflation.

As long as the economy grows, the federal tax system will generate revenues in the historic range of 16% to 20.9% of GDP. Current budgetary rules are skewed towards larger government in that they make it much more difficult to keep future tax revenues in this range. Rather than debate arcane budget rules about what should and shouldn’t be “paid for,” policymakers should have a frank discussion about whether the tax burden should rise well in excess of historic norms.

Christopher Papagianis is the Managing Director of e21 and was Special Assistant for Domestic Policy to President George W. Bush.