Taxpayers "on the hook for $59 trillion." Wait; is that a big number, or a small one?

USA Today's reporter Dennis Cauchon knows how to sell newspapers: just put a huge-looking number like "$59 trillion" in a headline about how much US taxpayers will have to come up with to cover social insurance programs.  Here's the article that got so much attention yesterday. 

Conspicuously absent was any mention about how much time taxpayers would have to come up with the money, or the estimated tax receipts that would flow into the government over that time period. 

I hunted and hunted in Cauchon's article for a reference or link to those (and other) important assumptions behind the $59 trillion; no luck.  Then I hunted and hunted for Dennis Cauchon's email address or telephone number, so I could ask him directly; no luck there, either.

I hope he sees this article, so he can send me an email explaining those important little tidbits behind their analysis.  Reason: Just as any mortgage holder plans to pay the mortgage gradually over a period of years out of future income—instead of coming up with the money tomorrow—so does the federal government plan to pay the social insurance costs gradually over a period of years out of future tax receipts—instead of coming up with the money tomorrow.  The scary-sounding "unfunded promises" really means "future costs to be funded out of future tax receipts"—although the article doesn't explain that. 

Just for fun, I added up how much tax revenue the federal government would collect between now and 2052 (the approximate timeframe it looks like USA Today's analysis covered), for various productivity growth scenarios.  After verifying the timeframe, all I'll need to know is the discount factor USA Today used on the way to their $59 trillion result.  After that, we can see what portion of future tax receipts will be required to pay the future social insurance costs.  For reference, it takes 15% of tax receipts today. 

If anyone knows Dennis Cauchon's email address or telephone number, please send it to me via private email.  My gut feel is that $59 trillion will look small compared to tax receipts over the same time period, using reasonable assumptions about productivity growth, but I can't confirm or disprove that without some help from him. 

Until then, I'll have to file this story in the folder I titled "Single Entry Accounting."

Comments

[mark]

Anyone familiar with a fiat money system knows the Treasury can make any payment. The real question should be will the economy produce the extra $59 trillion worth of goods and services. And the only way to produce more goods and services with a smaller work force is to increase productivity. As you pointed out Steve, many economist across the political spectrum deliver the productivity message, unfortunately it just doesn't sell newspapers or win elections.

[Bob]

Feinstein and Domenici in the accompanying photo. Both well versed in finance and economics, of course.

[Buddy Kidd]

Go to USAToday.com and look up that article on line or look up any article he has written. It has a place there to e-mail him.

[Matt]

If the "future costs to be funded out of future tax receipts" concept is truly viable, then why do we not permit corporations to operate under this principle? Whenever corporations incur a certain liability, like a future retirement/medical benefit for an employee, they are required to account for it on a yearly basis and plan ahead. The government has to account for it when it happens down the road. Thus, the government is free to overspend on retirement benefits because these future expenses are not going on the books. Which accounting practice seems more responsible?

[Stephen Reed]

I don't think Steve is disagreeing with the liability, but thinks the asset (future tax receipts) should also be recorded. Do I understand correctly?

[Steve]

Stephen:
That's correct. Debt is on the liabilities side of the balance sheet, and assets = liabilities. To make a fair judgment about those liabilities hogging all the headlines, we'd need to compare the value of future liabilities with the value of future tax receipts -- and that requires someone to estimate future GDP growth, future tax policies, etc. No wonder those assumptions tend to get buried; who would want to defend a guess about future tax rates and economic growth when the intended focus is on the scary liabilities side?

Further, companies have the market value of the stock to help gauge what investors think of the company's future. But the government doesn't sell equity shares, so the closest parallel is the interest rate on the long bonds. Judging from the low and steady interest rates on Treasury notes, investors apparently think our country's long term prospects are bright. Unfortunately, that inference doesn't get any mention in the doomsday articles.

[marmico]

The USA reporter is probably using the infinite-future basis for the discounted, present value of the liabilities.

For instance, the social security present value debt is ~$5T measured on the intermediate cost 75 year basis but ~$13T measured on the infinite-future basis.

Anyway, matt nails the issue.

[smess]

Using Google, Dennis Cauchon's USA Today telephone is 703/854-6508 and the email address is dcauchon@usatoday.com.

[Yevgeny]

The GAO issues an annual report on the state of US government finances.
The latest report can be found at

www.gao.gov/new.items/d07607t.pdf

At the top of page 9 (page 11 in the Acrobat reader) of the document it states that the NPV of total US gov liabilities is $61 trillion.

[Steve]

marmico:
I actually don't care one way or the other which accounting method the government selects. I care about the rhetorical equivalent of single entry accounting misleading the public into thinking tax hikes and benefit cuts are the only alternative, when in fact growth-enhancing policies are a superior (but unmentioned) alternative.

smess:
Thanks; I had already emailed the questions to USAT, but forwarded it to that address, too. I'll post a followup if anything interesting develops.

Yevgeny:
Thank you, that's what I suspected. I read it; they seem to "know" what GDP will be each year for the next several decades, and that's my fundamental complaint: they don't know, and that calls for testing various scenarios of GDP growth (or non-growth). For example, on pages 8-9, they say that if public debt is held at the current %GDP, huge tax hikes or spending cuts would be needed; again no mention of the GDP growth rate they assumed. (I believe it was around 2.5%, by the way, but am waiting for confirmation if I can get it.) In any case, if GDP grows at a faster pace, the outcome improves -- as I showed in the productivity article.

[marmico]

The GDP growth rate doesn't really matter for testing (modeling)social security solvency (actuarial deficit). The following table from the 2007 Report lays out the demographic and economic assumptions:

http://www.ssa.gov/OACT/TR/TR07/II_assump.html#wp95325

The plan is in balance or slightly overfunded in the low cost scenario. IMO, funding is not an immediate problem. Ten years ago, the Trustees Report projected insolvency in 2031. Now it projects the same in 2041.

However, medicare is a "crisis" problem.

[Steve]

marmico:
On the contrary, the GDP growth rate matters significantly in the social insurance solvency question. One day, the situation we have today will reverse, and the system will begin "borrowing" from the general fund. (In other words, instead of the government's left pocket borrowing from its right pocket, the two pockets will switch roles.)

The government is taxing 18% of GDP away from the public (all forms of taxation, regardless of how the government accounts for transactions between its own pockets). Let's assume that 18% is sustainable as a taxation policy indefinitely. Next question then becomes: Would 18%, plus a percentage point or two of borrowing, pay for all government spending we will incur in the future?

The answer to that depends on the magnitude of GDP in the future -- and GDP depends on growth rates, and growth rates depend on our commitment, or lack of commitment, to growth-friendly policies, starting now. And that has been my point all along.

[Kevin]

Hmmm, they should have included one of those italicized sub-heads to read:
"National Income During Same Period Expected to be $1,250 Trillion".

But I guess they'd lose the punch if they did that.

[Kevin]

Marmico & Steve, you're both kind of right. I see the OASDI table of variables does not include GDP but does include productivity. But the productivity variable is a function of GDP and population, and so necessarily has a GDP growth assumption built into it that they're not showing. The high productivity guess = high GDP growth.

[marmico]

I found the GDP numbers in this table:

http://www.ssa.gov/OACT/TR/TR07/VI_OASDHI_GDP.html#wp103985

Looks like ~5% nominal GDP growth.

I understand that tax revenues are fungible but most people see a FICA tax taken from their pay cheque in exchange for a future benefit and expect that promised benefit to be funded from the taxes withheld not from income taxes.

Sure, GDP increases but taxable payroll as a percentage of GDP is declining. And the FICA tax is based on taxable payroll.

[Jim Glass]

"IMO, [Social Security] funding is not an immediate problem. Ten years ago, the Trustees Report projected insolvency in 2031. Now it projects the same in 2041."

Alas, the Trust Fund that is supposed to last to 2041 is a *liability* of the gov't, not an asset. I don't know how so many people get this backward. By the end of the Trust Fund's life a 20% across-the-board income tax increase (personal and business) will be needed to pay off its bonds.

That *will* be a problem, especially in light of the matching tax hike needed for Medicare -- and the near matching one *also* needed to pay unfunded federal gov't and military pensions and health benefits.

"However, medicare is a 'crisis' problem"

A popular myth. The huge projected liabilities for Medicare are in the far out years -- but we're never going to get that far with a >50% income tax increase needed just to get though the Trust Fund years. During that period the cost of SS and Medicare are comparable.

It's like in Butch Cassidy and the Sundance Kid: "I can't swim." "Don't worry, the fall is going to kill you."

75-year actuarial balance for these programs means nothing when cash flow is going to kill them as we know them in the next 25 years.

It is true that Medicare reform will be much more difficult than SS reform. SS, after all, is a mere cash transfer program operating a checkbook. Medicare is trying to run a hugely complex industry. But cash-cost wise they are in the same ballpark in relevant coming years.

My next comment has the cash flow numbers.

[Jim Glass]

The "2006 Financial Report of the US" gives net liabilities over projected tax revenues as $53,062 trillion current value for 2006, on a 75-year basis. This is a big number. (The open ended number is much bigger. )

That's at
http://www.gao.gov/financial/fy2006financialreport.html

See the "Management's Discussion and Analysis" section, Table 10, for the bottom line number. There's much more detail in the other sections.

"they seem to 'know' what GDP will be each year for the next several decades"

The Trustees use the historic long-term average of the components for GDP growth -- note that with the decline of the working-age pct of the population following the retirement of the baby boomers, this indicates GDP growth will *fall* -- then run thousands of variant scenarios through Monte Carlo simulations to come up with their projections of GDP growth.

After that, SS's cost projections are basically just arithmetic. Medicare's are more "iffy" -- but many experts worry they are low because they are below historical experience.

Cash flow is the killer, and it will kill the current state of these programs as we know them in 30 years, probably 20 -- *long before* the SS Trust fund is spent or the big Medicare costs kick in.

The Trustees project cash flow cost increases in GDP terms from 2007, in the middle case scenario, at http://www.ssa.gov/OACT/TRSUM/images/LD_ChartE.html

On current law these shortages must be covered by income taxes. In 2007 income taxes are 11.1% of GDP (expected)

These columns give by year the increase in cost in terms of GDP, and the resulting across-the-board income tax increase (GDP terms) from 2007 to cover them:

year ... pts GDP/11.1 = tax increase

2020 ... 2.10 = 19% income tax increase
2030 ... 4.58 = 41%
2040 ... 6.09 = 55%

...etc., increasing forever, but by 2040 the law will be very different so it is pointless to project further.

BTW, such tax increases will further reduce growth, since the deadweight economic cost of taxes rises by the *square* of the increase in the rate. And that sets up a potential viscous circle.

This is why Stadard and Poor's projects the US credit rating as dropping from AAA to "junk" in just the 10 years from 2017 to 2027, under current law -- and 2027 is just the *start* of the cost increases.

Now 2017 is only 10 years away, that doesn't leave a lot of time for huge increases in economic growth to change the situation -- it's not the 75-year liabilities that are going to crash things, it's the 20-year liabilities that will.

I'm 100% with Steve that the more economic growth the faster the better!!

But wishing aways the size of these coming costs is not smart, IMHO. To the contrary, when I argue for faster growth policies I cite these coming costs as the #1 reason *why* these policies are needed ... and needed *now*.

[marmico]

"By the end of the Trust Fund's life a 20% across-the-board income tax increase (personal and business) will be needed to pay off its bonds."

Not necessarily. The intragovernmental debt can be rolled into public debt, starting today if the Treasury is so inclined. In other words, the off budget debt becomes on-budget debt.

[Jim Glass]

"Not necessarily. The intragovernmental debt can be rolled into public debt, starting today if the Treasury is so inclined. In other words, the off budget debt becomes on-budget debt."

Sorry, but the *cash* for paying SS benefits has to come from "somewhere", this cost rises by >2 points of GDP annually in latter years of the Trust Fund era ... and "somewhere" in the end is taxes, nowhere else (unless Congress decides to start selling off the national parks).

Rolling over the TF debt into the public debt doesn't save a penny of this cash flow cost to the gov't, obviously, as it doesn't reduce benefits to be paid -- nor does it save a penny of taxes.

Note that incurring $1 of public debt to avoid collecting $1 of tax in fact costs you $1 of tax -- because the cash flow cost of servicing the interest on $1 of debt discounts to a current value of $1, and ultimately the only way the gov't can get that $1 is through taxes (or selling off assets). So the tax cost remains exactly the same, you just deferred it by adding interest to it.

The "just increase the public debt" scenario causes the annual deficit to exceed 20% of GDP -- more than the size of the entire federal gov't today -- in the 2030s, due to the compounding interest cost. Clearly, an annual deficit >20% of GDP (and rocketing straight up through compounding) is impossible to sustain. This is why GAO projected "the end of government" by then, and S&P projects "junk" status for Treasuries by then, under current law, which requires exactly the running up of such debt.

There is no way to escape the cost of cash flow. Cash flow rules all.

[marmico]

Hello Jim, presumably you were directing me to your blog posted here:

http://www.scrivener.net/2004/12/social-security-reform-simple-reason.html

Some thought provoking notes contained therein. Is there anything in the actuary.org reform proposals that strikes your fancy?

http://www.actuary.org/pdf/socialsecurity/reform_07.pdf

"So the tax cost remains exactly the same, you just deferred it by adding interest to it."

It seems to me that if nominal GDP grows faster (say 6%) than the cost of capital (interest rate on the debt at say 4.5%) and if tax revenues are fixed as a percentage of GDP (say 18%), then debt servicing costs declines and the present value of the future tax cost is lower than the tax cost today.

However, I agree with you in principle that it is best to be prudent and begin to pre-fund the entitlements today rather than wait for tomorrow.

[Ken]

Me?

I think its ludicrous to think one can look out 50 years in the future and predict what the economy will be. Its really just a wild guess.

One may make assumptions about growth rate, productivity, population and other things, but the probability of any one particular scenario occuring is low. With the rumblings that Hillary Clinton is making about her "shared responsibility" society, I can easily see a replay of the 70's economy occuring. Factor that into your assumptions.

[jomama]

This is a wake up call, folks.

There are going to be numbers tossed
far and wide with many disinfo
artists promoting 'feel good'
responses.

Read this and weep:

http://djomama.blogspot.com/2006/12/first-world-government-junk-bonds-on.html

The disease is pandemic.

Then see this:

http://sfgate.com/cgi-bin/article.cgi?file=/c/a/2004/09/12/MNG2S8NOI21.DTL

[BP]

Found this in less than 1 minute using Google.

Want to email specific reporters at USA Today? For those that have email addresses, the following formula works most of the time: first-initial-last-name@usatoday.com; for example, dlieberman@usatoday.com (for David Lieberman, media reporter in the Money section).