Back to podcast

Intro: This is DC Dynamics, a podcast about what’s coming up in US tax policy, with a look to the past as our guide. I’m your host, Ray Beeman, from Washington Council EY, and the topic today – Demystifying the Revenue Estimating Process – is one that is near and dear to my heart. The Joint Committee on Taxation (or what we’ll call JCT) is where I first landed in the Washington tax policy world 20 years ago, and I can attest to the importance the revenue estimating process plays in making tax law and the thought that goes into it by really smart people.

I’m a lawyer, not an economist, and it really takes a village at JCT and the Treasury Department to develop estimates of the impact of tax law changes on federal revenues. Economists consult with lawyers and accountants to understand legislative text and produce estimates that – in the case of JCT – can make or break a proposal in Congress and – for Treasury – provides a blueprint for implementing a President’s vision of how to use the tax system to cut taxes, raise revenue, incentivize behavior, and – of importance lately and perhaps even more so now – reduce the deficit. So, what’s really behind the 12-column tables that put a price tag on policy proposals?

Joining me today is my friend and colleague Bob Carroll, an economist who is head of the QUEST group here at EY – QUEST stands for Quantitative Economics and Statistics. Bob previously served as Deputy Assistant Secretary for Tax Analysis at the US Treasury Department in the George W. Bush administration, serving as the Department’s top economist and working on tax policy issues during a very busy time for tax legislation with all that was going on in the early 2000s. Bob oversaw the Department’s work on developing tax cuts for individuals and business tax reforms to improve the global competitiveness of the United States. He also led the Department’s efforts on comprehensive tax reform and, in particular, the 2005 President’s Advisory Panel on Federal Tax Reform. Prior to these roles, Bob spent more than a decade as a revenue estimator in Treasury’s Office of Tax Analysis.

Bob is going to walk us through where it all starts and where it ends. So, Bob, where does it start?

Bob: Well, thanks Ray, it starts with the baseline. For congressional purposes, the JCT baseline is the Congressional Budget Office’s 10-year projection of federal receipts, which assumes that present law remains unchanged during the 10-year budget window. It is an estimate of the federal revenues that will be collected over the next 10 years under current, or present, law.

Treasury uses the OMB baseline. But, in a sense, has more latitude to place estimates into the context of the administration’s policy objective. Estimates presented in the president’s budget are typically estimated relative to the administration’s policy baseline – a baseline that assumes all of the administration’s budget proposals are enacted.

There have also been examples of Treasury building a set of policies into the baseline to provide further context, including the Obama administration assuming permanency of some of the Bush tax cuts and then expansions of the child tax credit and the earned income credit. Just this past year or two we saw the Biden administration budget assume enactment of the Build Back Better Act even though enactment was very much in doubt. The choice of a baseline can have a significant effect on the estimates.

Ray: So, what about dynamic scoring, or macroeconomic analysis to use the technical term – essentially, counting macroeconomic effects of behavioral reactions to tax law changes that would change GDP, or gross domestic product, and account for long-term effects on the economy. Democrats for years criticized Republican efforts to take these effects into consideration, including when I worked for House Ways and Means Committee Chairman Dave Camp on the 2014 tax reform draft, and then said the scoring process was being “rigged” when the Republicans in 2015 required CBO and JCT to incorporate dynamic scoring into their official estimates of major legislation.

Bob: That’s right. In 2015, the House budget rules were changed to require the JCT to provide macroeconomic estimates for any revenue change with a revenue impact greater than 0.25% of GDP in any year during the budget window. It was a very controversial topic not so many years ago but, fast forward to 2021, and the Democrats embraced dynamic scoring for the Build Back Better Act, which in August was enacted in much narrower form as the Inflation Reduction Act.

Just to take a step back, it is important to remember that even conventional JCT revenue estimates include a broad range of behavioral responses by individual and business taxpayers but assume that the overall size of the economy remains unchanged as reflected by GDP. In dynamic scoring by contrast, JCT relaxes the conventional scoring assumption that the size of the economy is fixed and models the impacts of a given proposal on macroeconomic variables like output, employment, and investment. JCT then calculates tax revenue based on the new underlying economic variables consistent with the estimated change in the size and composition of the economy.

JCT staff actually began developing a capacity to model for macroeconomic impacts of tax policy proposals way back in 1996, and macroeconomic analysis has been provided to Congress since 2003 even if they could not be used as official revenue estimates under the rules at the time. The 2015 rule change I mentioned, which was done away with in 2019, required a single estimate of the deficit effect due to the macroeconomic response to proposed tax legislation. This rule was reinstated in the House for the current Congress under the new Republican majority.

Incidentally, the House rules for the current Congress also require that cost estimates include, to the extent practical, a statement estimating the inflationary effects of legislation. A similar rule was in place some four decades ago in the early 1980s.

Ray: And projected economic growth played a pretty big role in getting the 2017 Tax Cuts and Jobs Act over the finish line, when then-Senator Bob Corker finally agreed to a budget that would increase the deficit by $1.5 trillion after he was promised that increased revenue spurred by economic growth from the TCJA would make up that projected shortfall. The macroeconomic analysis that was released as the bill was signed into law said it would reduce the TCJA’s total $1.45 trillion increase in the deficit only by about $451 billion.

Now, let’s give our audience an idea of what’s behind the process of modeling, which to many people might conjure up images of a room-sized 1950s computer with reams of paper pumping out estimates in dot matrix font.

Bob: OK, it isn’t like that anymore at least, but I can see how there’s some mystery because people outside the process don’t really get to see it. For conventional scoring, JCT and Treasury rely on Individual, Corporate, and Estate and Gift Tax Models, and other models, and also large volumes of tax return data provided by the IRS Statistics of Income Division. They also use a variety of other data and economic research to inform their models and analysis.

When measuring the macroeconomic effects of tax legislation, JCT uses three models of the economy that, through differences in structure and behavioral parameters, attempt to capture a range of approaches for reflecting those effects: (1) the macroeconomic growth (or MEG) model, (2) an overlapping generations (or OLG) model, and (3) a dynamic stochastic general equilibrium (or DSGE) model. In short, each model has somewhat different depictions of the economy and how the economy can be expected to evolve over time.

The MEG model forecasts prices adjusting so that demand equals supply in the long run, but not necessarily in the short run. The OLG model assumes that prices adjust to any changes in economic conditions so that supply equals demand in both the short and long run. This model is often described as a full-employment model. The DSGE model accounts for uncertainty about future states of the economy, which can affect decision-making. The DSGE can account for both random variables and the reactions of businesses and individuals to these random variables. Also, the MEG and OLG models are capable of modeling international tax changes.

Ray: So, Bob, what are some hot topics in revenue estimating? We are coming off a pretty busy year for tax legislation, and the TCJA cliffs around R&D amortization and limitations on interest deductibility, as well as the phase-out of bonus deprecation, remain front of mind for potential tax legislation in the current Congress. Of course, the big fiscal cliff with TCJA individual and other provisions does loom at the end of 2025. And there’s still a lot of uncertainty around international tax, which probably isn’t going away any time soon, with the global tax agreement in limbo.

Bob: Well, it’s kind of interesting that a 4-year delay in the 174 R&D amortization provision, which was included in the House-passed Build Back Better Act, was scored as costing a mere $4 billion over 10 years, but repeal is likely to be over $100 billion. The provision was scored in the TCJA as raising $120 billion over 10 years. Short-term extension of provisions that involve primarily timing, such as the R&D amortization provisions and the bonus provisions, typically cost little over 10 years, but can cost a lot over the 10-year budget window.

On both the R&D amortization provisions and the other major TCJA cliff that took effect this year, the stricter EBIT-based interest expense limitation, QUEST recently estimated that both would significantly affect jobs and GDP.

And on the overall economic front, there is a lot of attention being paid to both inflation and job numbers. Again, the House reinstated a close to 40-year-old rule for the JCT to include estimates of the inflationary effects of significant legislation.

Ray: And we’ve also heard from Republican members of Congress who say that Treasury has not provided enough information for JCT to estimate the revenue effects of the OECD-led global tax agreement. In fact, Congressman and Ways and Means Committee member Kevin Hern made this very point in offering a resolution of inquiry directing Treasury to provide documents relating to the impact of the Pillar One agreement. So, I expect that to be another big revenue estimating topic this Congress.

Revenue estimating has often been referred to as the black box of tax policy, but that perhaps should be taken as a compliment to the JCT and Treasury staff who all have PhDs in Economics while the rest of us mere mortals watch from the outside. And with that, I hope we have helped you demystify the revenue estimating process. Thank you, Bob, very much for joining us today and providing your insights and, until next time, I’m Ray Beeman and this has been DC Dynamics.

Back to podcast