With the U.S. Presidential elections around the corner, voters are grappling with issues such as how the candidates’ economic plans will affect jobs and economic growth. The recently expanded Penn Wharton Budget Model offers some help. It uses data analytics and cloud computing to create a non-partisan, interactive, online tool to figure out how various changes in government policies might affect the economy. In this Knowledge at Wharton interview, Wharton professor of business economics and public policy Kent Smetters, who led development of the tool, discusses the new module on tax policy analysis, comparing Hillary Clinton’s plan with Donald Trump’s. For a another perspective on the plans, check out this recent segment that aired on the Knowledge at Wharton radio show on SiriusXM.
Below is an edited version of the conversation:
Knowledge at Wharton: What are the different tax plans that you’re analyzing?
Kent Smetters: The tax policies of the two Presidential candidates, [Hillary] Clinton and [Donald] Trump. There’s also been a large effort by the House GOP, led by Speaker [Paul] Ryan and Chairman Kevin Brady of the House Budget Committee, and we also analyzed their plans.
Knowledge at Wharton: What are the main features of the plans for the average household?
Smetters: For individual households, all the plans make two adjustments. The first is that the marginal tax rates are changing. The marginal tax rate is how much in taxes you pay for the next dollar that you make. Economists generally believe marginal tax rates are much more important for economic activity than the average tax rate. The second feature of the plans is they change the amounts of itemized and standard deductions and thus the income before you’re taxed.
Knowledge at Wharton: What about business taxation?
Smetters: That’s where there are much bigger changes, especially in the Trump and the House GOP plans. First, the corporate tax rate itself is lowered in both the Trump and the House GOP plans. In the Clinton plan, it’s increased a little bit.
The second aspect of the Trump and the House GOP plans, which is the most aggressive, is that they accelerate how companies can depreciate their capital investments. Economists generally believe that that’s much more important for stimulating the economy than the corporate tax rate. The reason is, by accelerating the amount of depreciation — sometimes called expensing — it isolates new investment, whereas, changing the corporate tax rate impacts not just new investment, but existing, installed capital as well.
The third feature of the Trump and Ryan plans is that they affect what’s called the “pass-through rate” taxed on income earned by businesses like partnerships and LLCs that are passed to their individual owners, who add them to their returns. Such income will have a lower tax rate.
“The Trump and the House GOP plans … accelerate [how] companies can depreciate their capital investments.”
Knowledge at Wharton: What are the main differences between those plans?
Smetters: It’s a sweeping generalization, but roughly speaking, the Trump plan is the most aggressive. The Clinton plan is the least aggressive. The House GOP plan is in between, but a bit closer to the Trump plan.
Knowledge at Wharton: When you say aggressive, you mean just the degree of change.
Smetters: Just the overall degree of change.
Knowledge at Wharton: Users can go to your website and look at different scenarios. The budget model for Social Security and immigration looked at how an individual change would change the bottom line. But when it comes to taxes, it’s a little more complicated, because you get into behavioral issues. What will people do with tax savings? How will they spend it?
Smetters: In the case of Social Security and immigration, our simulations allow people to play with their own policy ideas: for example, raising the retirement age or changing the amount of the taxable maximum of income that’s applied to Social Security. Less important in Social Security and immigration are aspects like demographic assumptions. Economists are much more aligned on those.
When it comes to tax policy, the big disagreement is on the behavioral assumptions. So that’s what we highlight in the simulator. Also, the candidates have specified their policies. So we’re focused more on letting people see if they believe one behavioral assumption is more important than the other; they can test their ideas.
After the election, we’ll bring back the tax calculator and the ability for people to design their own tax plans.
Knowledge at Wharton: How is the difference in this behavioral aspect handled? In the end, that’s a judgment call, correct? Do you offer different scenarios where users could make a judgment about how they expect taxpayers to react?
Smetters: Yes. There are four basic controls. The most important is the rate at which international capital will flow in and out of the United States. That is by far the most important because several of the tax plans, in particular those of Trump and the House GOP, will create some deficits along the way. If international capital flows are very aggressive, then that will minimize the negative impact of those deficits on the economy. If, however, we are closer to a closed economy, then those deficits compete for household savings and reduced private capital.
“If international capital flows are very aggressive, then that will minimize the negative impact of those deficits on the economy.”
Knowledge at Wharton: That’s the famous “crowding out” effect, correct?
Smetters: Yes, that’s exactly right.
More international capital flows [will mean] less crowding out. So it’s more favorable. Then we have other assumptions, like, how much do households themselves increase their labor supply or decrease it with respect to a tax change? Also, [we look at] what’s called a saving loss, to see how they respond in changing their savings behaviors.
Historically, academic models have focused on those elasticities — the labor supply elasticity and the savings elasticity. The reason is academic models typically focus on balanced budget experiments, like optimal tax design for a given amount of revenue. But it turns out that for this exercise, international capital flows is by far the most important assumption.
Knowledge at Wharton: How do these things stack up against each other? When you changed behavioral aspects, what did you find?
Smetters: We provided a generous range of different behavioral assumption parameters.
At the same time, we do pick what we think is a reasonable baseline, but allow users to change that. In conjunction with building the simulator, we have conducted empirical exercises and reviews of the literature to narrow down what we think is a reasonable starting point for people. In the case of the tax simulator, there are 256 different combinations that people can play with. If they have different judgments about it, they can decide for themselves.
Knowledge at Wharton: How did it turn out when your group, let’s say, put in what would be the most likely scenario from their point of view?
Smetters: Because of the upcoming election, maybe we’ll just discuss the Clinton versus Trump analysis.
What we viewed as the most likely scenario is, in the Clinton plan, in the short run — it’s fairly neutral on the economy. There’s a small little positive, followed by a small little negative. But it’s fairly neutral, in terms of GDP, jobs, and so forth. [This single part of the overall model is] focused on just the tax policy.
In the long run, because the Clinton plan is overall increasing taxes, it’s leading to lower debt that otherwise would have occurred under current policy. That has a positive effect on jobs. By 2027, we’re projecting that there would be about 600,000 more jobs. By 2040, [the U.S. would have] about two million more jobs than we otherwise would have had in that year. These aren’t enormous changes, but they start out at slightly neutral, and then going to more positive.
For Trump, it’s the opposite. Almost all the bang comes early on, and in the short run we’re projecting that GDP will go up about 1.75%. In an upper-bound calculation, about two-and-a-half million jobs would be produced early on, with his approach.
However, over time, because his plan is unbalanced fiscally, it’s going to produce fairly large deficits. That will have this crowding-out effect. [Government borrowing will] compete with private capital for household savings. We’re projecting that within 10 years — 2027 — we’ll have 700,000 fewer jobs. By 2040, if the debt is as continuously almost spiraling out of control, we’re projecting 11 million fewer jobs.
Knowledge at Wharton: Down the road, when policies are chosen, you could look back and say, “Here’s what the model projected. Here’s what actually happened.” And you’ll be able to tweak it in different ways.
“The Clinton plan is overall increasing taxes, and leading to lower debt that otherwise would have occurred under current policy.”
Smetters: That’s right. As I like to say, all models are wrong. What I mean is, there’s no model that is going to give a perfect crystal ball look into the future. What we get from these models is more of an understanding of the direction of things. Is this likely to be stimulating or contractionary for the economy, relative to where the economy eventually lands? It’s those – what we call deltas — that are more reliable. Essentially, how big are those deltas? Are we talking about a small impact, or a potentially large impact?
Knowledge at Wharton: For someone who wants to dabble in this model, where can they go to see it?
Smetters: They can go to our website. The big picture of this project is consistent with the Wharton School’s overall vision of trying to use data analytics, advances in theoretical modeling as well as cloud computing to simulate public policy.
A webinar explaining more about this model and the tax plans of the presidential candidates, created in conjunction with the Urban-Brookings Tax Policy Center, can be viewed here.