00:00:02
We're here with Kent Smeters. He's a Wharton professor of business economics
00:00:06
and public policy. Welcome, Kent. Good to be here. He's going to talk to us today about the latest iteration of the
00:00:13
Penn Wharton budget model which Kent was instrumental in developing and uh this is a free nonpartisan online interactive
00:00:23
model uh which anyone can can plug into. Uh, and the last time we spoke, we covered social security and immigration,
00:00:31
the effects that they might have on on the budget. And this time, we're going
00:00:34
to talk about tax policy, uh, those policies of the two presidential candidates and also, I think, a couple
00:00:41
of other, uh, leading proposals that are floating around Congress, but I guess we'll focus mostly on the presidential
00:00:47
candidates. So, uh, let's just jump into it. What What are the different tax
00:00:51
plans that you're analyzing? Yeah. So, we uh as you just mentioned the two presidential candidates Clinton and
00:00:57
Trump and there's also been a very large effort by the House GOP led by Speaker
00:01:03
Ryan and chairman Kevin Brady of the House Budget Committee. Um and so we've
00:01:09
also been uh we also analyze their plan. So what are the main features of the plans for the average household? Yeah,
00:01:16
for individual households there's all the plans have basically make two adjustments. The first is simply the
00:01:24
marginal tax rates are changing. And in particular, the marginal tax rate is how
00:01:30
much uh in taxes you pay for the next dollar that that you make. And economists generally believe marginal
00:01:38
tax rates are much more important for economic activity than say the average tax rate that you you you face. And the
00:01:45
second thing that the plans do is they change the amount of um itemized and standard deductions. And that's that's
00:01:52
the amount of income that you can have before you're taxed. And what about for
00:01:56
business taxation? That's where there's a lot bigger amount of changes, especially in the Trump and the House
00:02:03
GOP plan. First, the corporate tax rate itself is lowered in both the Trump and the House GOP plan. And in the Clinton
00:02:12
plan, it's actually increased uh a little bit. And the second thing that the Trump and uh uh House GOP plan in
00:02:20
particular do, which is the the most aggressive thing, is that they accelerate the the rate at which uh uh
00:02:28
companies can depreciate their capital investments. And economists actually generally believe that that's much more
00:02:36
important for stimulating the economy than the corporate tax rate itself. And the reason why is by accelerating the
00:02:43
amount of depreciation sometimes called expensing it only isolates new investment whereas changing the
00:02:51
corporate tax rate is not just impacting new investment but existing installed capital as well. And uh the third thing
00:03:00
that the Trump and uh Ryan plans do is that they affect what's called the pass
00:03:06
through rate. Some capital income is taxed at the household level, some at the corporate level. More of theirs
00:03:11
would be taxed at the corporate level, which will have a lower tax rate. So, what are the main differences between
00:03:18
those plans? Right. I mean, it's a sweeping generalization. Um, but roughly speaking, you know, the Trump
00:03:26
plan is the most aggressive. Uh, the Clinton plan is the least aggressive, and the House GOP is kind of in between.
00:03:34
and um a bit a bit closer uh uh uh to the to the Trump plan. Though when you say aggressive, you mean just the the uh
00:03:42
degree of change, just the overall degree of change. That's right. All right. So, um so users can go to your
00:03:49
website and look at different scenarios. Uh the budget model for social security
00:03:54
and immigration uh looked very much at uh uh specifically how an individual change would change the bottom line. But
00:04:02
when it comes to taxes, it's a little more complicated because you get into behavioral issues. What will people do
00:04:07
with a tax savings? How will they spend it? Where will they put that money? And as you talked about, for example,
00:04:13
accelerated depreciation that's changing a company's business decisions. So talk
00:04:18
about that a little bit, those differences. Yeah. So in the case of social security and immigration, what
00:04:23
our simulas allow people to do is to uh go and play with their own policy ideas of for example um raising the retirement
00:04:33
age or changing the amount of uh the the taxable maximum that's applied of income
00:04:38
that's applied to social security. Um and the and less important in the in social security and immigration are
00:04:45
things like demographic assumptions. Those are the things that economists are much more aligned on. Um, when it comes
00:04:51
to tax policy, where the big disagreement is is is actually on the behavioral assumptions. And so that's
00:04:58
what we're highlighting in the in the simulator itself. And also the candidates themselves, they have
00:05:05
specified what the policies are. And so, um, we're focused more on, uh, giving
00:05:11
letting people see if they believe, um, uh, one behavioral assumption is more important than the other. They can test
00:05:18
their their their ideas. Eventually, after the election is over, we'll we'll
00:05:23
bring back to the tax calculator, um, the ability for people to design their own tax plans. So, this difference u in
00:05:31
in this behavioral aspect, how how is that handled? Because in the end, that's
00:05:35
a judgment call. Correct. So you offer different scenarios where you can uh where where users could make a judgment
00:05:42
about how they expect taxpayers to react and Yes. Yes. And so in particular there's really four basic um controls.
00:05:52
The one that's the most important is the is the rate at which international capital will flow in and out of the
00:05:59
United States. And that one is by far the most important um one because uh several of the tax plans in particular
00:06:08
Trump and the House GOP will create some deficits along the way. And if international capital flows are very
00:06:15
aggressive then that will um minimize the negative impact of those deficits on the economy. If uh however, if we're
00:06:25
closer to a closed economy than those deficits that compete for household saving and reduce private capital um but
00:06:32
then we have other assumptions. So that's that's the famous crowding out
00:06:35
effect. Correct. Yes, that's that's exactly right. Okay. And so more international capital flows, less
00:06:40
crowding out um and so it's more favorable. Um uh then we have other assumptions like how much do households
00:06:47
themselves increase their labor supply or decrease it with respect to a tax change and uh this uh what's called a
00:06:54
saving elasticity. How much do they respond in changing or saving? Historically academic models have really
00:07:01
focused on those elasticities the labor supply elasticity and the saving elasticity. And the reason why is
00:07:06
academic models typically focus on balanced budget um experiments uh kind of like optimal tax design type stuff
00:07:13
for a given amount of revenue. But it actually turns out for this exercise um the international capital flows is by
00:07:21
far the most important assumption. Uh and so how do these things stack up against each other? I assume I'm
00:07:28
assuming that you made assumptions that your your best idea about how things were likely to go when you were choosing
00:07:34
these behavioral aspects. Uh what did you find when you did that? Yeah, I mean so when we um we provide a a very
00:07:43
generous uh range of those different behav behavioral assumption parameters and so it's not necessarily that I
00:07:52
personally believe that range is that wide. You plural, but yeah, that's right. And uh but at the same time we do
00:07:58
pick what we think is a reasonable baseline but then allow users to change that. And so when it comes to uh how do
00:08:06
we come up with a reasonable baseline is that we've uh in conjunction of with uh
00:08:11
building out the simulator have conducted empirical exercises and reviews of the literature to uh narrow
00:08:19
down what we think is a reasonable kind of starting point for people. But in the
00:08:23
case of the tax simulator, there's 256 different combinations that people can
00:08:27
play with, you know, so that they um if they have different judgments about it, they they can they can decide for
00:08:34
themselves. But how did it turn out when you uh when uh your group let's say put
00:08:41
in what would be the most likely scenario from their point of view, right? So we why don't we because the
00:08:46
upcoming election maybe just discuss the the Clinton versus you know Trump analysis and so what we viewed as the
00:08:54
most likely scenario um is uh for in the Clinton plan in the short run it's fairly neutral on the
00:09:03
economy. There's a small little you know positive followed by a small little uh
00:09:08
negative but it's it's it's fairly neutral in terms of uh GDP jobs and so
00:09:14
forth. And this would be just tax policy, not accounting. For example, I mean, both candidates, for example, are
00:09:19
talking about uh fiscal spending, right? They're talking about, you know, spending on infrastructure and that sort
00:09:26
of thing. So, this is just looking this is isolating tax. That's right. It's
00:09:29
it's focused on on just the tax policy in the long run because uh the Clinton
00:09:34
plan is overall increasing taxes. it's actually leading to lower debt um that
00:09:42
that otherwise would have occurred under current policy and that's actually having a positive effect on jobs. So by
00:09:50
2027 uh we're projecting there will be about 600,000 more jobs by 2040 about 2
00:09:57
million more jobs um uh than otherwise would have had in that year. So these aren't enormous changes, but they um
00:10:05
start out at, you know, slightly neutral and then going to more more positive. For Trump, it's just basically the
00:10:12
opposite. Almost all the bang comes early on. And in the in the short run, we're projecting that GDP will go up
00:10:19
about 1.75%. And a real upperbound calculation on how many jobs would be calc uh would be uh
00:10:29
produced um is about two two and a half million jobs would be produced early on with his approach. However, over time,
00:10:38
because his plan is very unbalanced uh fiscally, it's going to produce fairly
00:10:44
large deficits, and that's going to have this crowding out effect. It's going to
00:10:48
be competing for with uh uh for with private capital for household saving. And so, we're projecting by within 10
00:10:56
years, 2027, um he uh we'll actually have 700,000 uh fewer jobs. And by 2040, if the debt
00:11:06
is just continues to kind of almost spiral out of control, uh we're actually
00:11:11
projecting 11 million fewer jobs. So it be very interesting down the road when policies are chosen and so forth
00:11:18
for you to look back and say here's what the model projected, here's what
00:11:21
actually happened and you'll be able to tweak it in and in in different ways.
00:11:25
That's right. And and that's why I like to say all models are wrong. I mean
00:11:29
there's no model that is going to have you know perfect crystal ball in into
00:11:34
the future. What we really get from these models is more understanding kind of the direction of things. Is this
00:11:41
likely to be stimulating or contractionary for that economy relative to where the economy eventually lands.
00:11:49
Um it's those what we call deltas that we think are more reliable. And then essentially how big are those deltas?
00:11:55
Are we talking about a small impact and potentially large large impact? And for someone who wants to dabble in this
00:12:02
model, where can they go to see it? Yeah, they can go to um simply our our website wi which is uh uh www um
00:12:12
budgetmodel.warton.upen.edu. Okay. Anything else you'd like to add that we haven't covered about this? No,
00:12:19
I think um you know the big picture is this whole u um you know project is really consistent with you know the
00:12:26
Wharton school's overall vision you know of trying to use data analytics advances in theoretical modeling as well
00:12:34
as cloud computing all these things that we're we use um for things in you know
00:12:40
like marketing and business analytics we're now applying this to public policy
00:12:44
and so this is you know I think an exciting venture for us and and the Wharton School. Thanks for coming in
00:12:50
today. Thanks for having me. [Music]