I applaud progressives in their aims. Their methods are not always very smart, but the idea that the government should protect those unable to fend for themselves very well is an attractive one. Doing this efficiently is wonderful.
Of course the problem is that there are always more problems to fix than there is money to use to fix it, and politicians don’t like to compromise. Let’s walk through the current Federal budget to understand why America’s deficits aren’t going down anytime soon, meaning that the country will be on an unsustainable path and will likely end up in a crisis that puts this one to shame at some point in our lifetimes.
The Federal budget in 2008, as laid out by the CBO: http://economix.blogs.nytimes.com/2009/12/07/where-do-your-tax-dollars-go/#more-43763
Defense and Security: 21%
Social Security: 21%
Medicare, Medicaid and CHIP: 20%
Safety Net Programs: 11%
Interest on Debt: 8%
Benefits for Federal Retirees and Veterans: 6%
Scientific and Medical Research: 3%
Transportation Infrastructure: 3%
Non-security International (aid, etc): 1%
All other: 5%
Here are the sources of revenue:
45% Individual income tax (levied on the top approx. 52% of earners)
36% Payroll tax (flat tax with a ceiling to pay for social security and medicare)
12% Corporate tax
3% Excise taxes
Of course, just looking at this, we could just say “fine, double taxes, and we don’t have deficits anymore”. As any good economist knows, you can’t really do that. This requires a simple model of the economy. For those of you not too numbers oriented, skip the model and look at the conclusion.
Choosing the macro model is the hardest part, because every macro model has problems, and I’d like this post to be reasonably simple and illustrative, instead of technically correct.
Let Y sub n equal GDP in period n
C sub n equals consumption in period n
I sub n equals investment in period n
G equals government spending in period n
t sub n is the tax rate in period n
s sub n is the savings rate in period n
A is the return on savings invested in the prior period, which is consumed this period.
Start with the basic IS-LM equation (I know, it’s basic, but it should get the orders of magnitude correct enough for what we’re going to use it for), except we adjust for the fact that the tax elasticity of pretax income isn’t 1. As taxes go up, pretax income goes down. We call this factor z. It will equal 1/(1+e), where e is the elasticity.
Yn = Cn + In + Gn
Cn = (1-tn)(1-sn)Yn + AIn-1
I = sn(1-tn)Yn
G = ztnYn
This simplifies into:
Yn = Asn-1(1-tn-1)Yn-1 / [tn(1-z)]
Which means economic growth, period to period, is equal to:
Asn-1(1-tn-1) / [tn(1-z)]
You can write this out to look simpler:
(Asn-1 – Atn-1sn-1) / (tn – ztn)
I sincerely hope that the n’s and n-1’s came out subscripts; if they did not, I apologize, I haven’t figured out blogger’s technical nuances yet.
You’ll note that as savings goes up, this period’s consumption goes down, but next period’s goes up. Duh.
As tax rates increase, growth decreases if z is less than 1.
So given all of this, let’s set constraints on our little budget exercise.
Because we don’t have info for individual income groups, we’ll use Chetty’s determination of a 0.5 tax elasticity of pretax income (In other words, when you multiply tax rates by (1+x), you have to multiply pretax income by (1-0.5x) (I use this terminology to avoid confusion about how big a shift in tax rates from 50% to 51% is… is that 1% or 2%? By the definition of elasticity, that’s 2%, but that gets confusing). This sets z at 1/ (1 + 0.5) = 2/3.
Capital investment by businesses in 2006 was 1.31 trillion on a GDP of 13.06 trillion. Thus, I set a savings rate of 10%. I’ll also arbitrarily assume that in the long run, there’s an equilibrium savings rate, so last period’s savings equals this period’s savings.
Historically, I believe A has been in the vicinity of 1.1, so let’s keep it at that (that’s certainly what most people in finance use for the weighted average cost of capital). In reality, the savings rate isn’t exogenous, so A won’t be exogenous, but for simplicity, we’re keeping them both exogenous, especially since the other inputs are similar to recent times (this helps the case of those who want to see a European welfare state, doesn’t hurt it).
I will similarly make the assumption that you can’t keep fueling growth with tax cuts forever, so prior t must equal present t.
This simplifies things into
(1.1*.10 – 1.1*t*.10) / (t- (2/3)t)
Which means growth is
(.33 - .33t)/t
This breaks down at extremes because z changes. (it’s unlikely you’d see 0 growth at 100% taxation – it’d almost certainly be negative, because z would get very, very negative - whatever side you think we're on of the Laffer curve, a 100% taxation would almost certainly be on the wrong side of it).
However, if we’re targeting the same growth we’ve had for the last 50 years (about 2.5% real GDP growth), it’s much more reasonable to assume the elasticity remains the same. To account for some level of acceptable slowdown but still leave a margin of safety, let’s say that we need to clear 2% annual GDP growth. If we’re not growing at all, then you’ll end up with some very bad externalities caused by massive gains in geopolitical power by developing countries with unstable governments. There will always be exogenous shocks (wars, resource crises, etc). With 2.5% GDP growth, the US’s share of world GDP budged only slightly. Asking us to maintain that growth rate is tough (and even maintaining that growth rate, we’ll lose a world GDP share), but for the US to remain strong, we need to grow. So 2% is the arbitrary limit. I’d be very scared to go much below that.
Setting (.33-.33t)/t = 1.02, this means that taxation cannot be higher than 24.44%.
Government spending as a percentage of GDP was 45% this year. This is artificially high because of the bailouts. It’s expected to be a little over 42% next year, so we’ll use that as a baseline. http://www.usgovernmentspending.com/us_20th_century_chart.html
About half of that was debt, indicating the current tax load is about 21%, normalized. You can thus only increase tax revenues by about 16% and still have a stable country.
This is supported by the tax share data. Here are the appropriate tax shares with the threshold for hitting that salary. What this means is that the top 1% of earners in the country pay 40.42% of all income taxes. The bottom 50% of earners pay 2.89% of all income taxes.
99%+: 40.42% ($410,096)
95-99%: 20.21% ($160,041)
90-95%: 10.59% ($113,018)
75-90%: 15.37% ($66,532)
50-75%: 10.52% ($32,879)
For reference, the shares by the top 1%, 5%, and 10% are the highest shares of any developed country, indicating that the tax burden is more progressive than elsewhere. This lends further credence to the idea that you can’t really add that much more in terms of taxation, because eventually, the wealthy will just leave the country (meaning everyone is in big trouble), and the poor don’t have that much to give.
So let’s say we can increase Federal revenue by 16% through clever tax increases. This means that taxes can support 58% of the current Federal budget. Let’s round up - it’s not unreasonable to assume there may be a few “free lunches” out there, like cigarette taxes, or infrastructure spending, which actually boost Y as they go up, and provide government revenue. It also acknowledges that this model is imperfect. Finally, the fact that we’re growing at 2% a year in this model means that if debt grows slower, we’re improving things. So let’s say that the revenue side and efficiency side can support 2/3 of all government spending.
To balance the budget, we need to eliminate about 1/3 of all Federal spending. Because I posted the percentages much higher on the page, I repost them here. I’m converting percentages into points to make it less confusing to subtract.
Points of spending
Defense and Security: 21
Social Security: 21
Medicare, Medicaid and CHIP: 20
Safety Net Programs: 11
Interest on Debt: 8
Benefits for Federal Retirees and Veterans: 6
Scientific and Medical Research: 3
Transportation Infrastructure: 3
Non-security International (aid, etc): 1
All other: 5
How would you cut 33 points from that budget?
A few other notes:
About 8 points of the 21 Defense and Security points are revenue positive (technology and patents), and another approximately 5 points are uncuttable - we still need a military, and we need to support the rest of the world. So you can take 8 points of defense and security.
You cannot touch the interest on debt, and scientific and medical research is revenue positive. At least some component of education and transportation infrastructure is revenue positive, and they’re small anyway.
So cut 25 points from:
Social Security: 21
Medicare, Medicaid and SCHIP: 20
Safety net programs: 11
Benefits for Federal Retirees and veterans: 6
Non-security International (aid, etc): 1
All other: 5
(and for those of you wondering, almost everyone thinks that the healthcare bill will increase Medicare’s portion, not decrease it. Things must be cut.)
Does this make it any easier to understand why rapid expansion of government has occurred? And does it provide any sort of insight into some of my opposition to socially beneficial programs? There’s an opportunity cost to everything, we must cut if we want the country to stay afloat, and so some socially beneficial programs can’t be done.