I don't make any claims to expertise in government finance. I do admit to being greatly frustrated by the subject for my entire career as a financial writer. That's why I was eager to coauthor "The Coming Generational Storm " with Larry Kotlikoff. He is one of the creators of Generational accounting, its prime mover, and its most tireless advocate. More important, generational accounting is the only tool I have seen in over 30 years of reporting that makes sense of government finance.

Generational Accounting isn't a complicated idea.

It begins with a simple concept, an algebraic identity.  The present value of all taxes to be collected must be equal to the present value of all benefits and spending plus all current debt.

Note that this is not a straight jacket. Its demand is logical: over a very long period of time, all forms of income must equal all forms of outgo.

If collections fail to meet promises, current beneficiaries (and taxpayers) are leaving a debt to future payers of taxes. I'll leave the fine-tuning to the model builders.

Needless to say, our government doesn't use this concept in its accounting. While any business entity in America larger than a Kool-Aid stand uses accrual accounting to try to sort out the impact of promises and commitments over time, our government uses cash accounting. It declares its books balanced when the money that comes in during the year is equal to the money that goes out during the year.

Our government is not alone. Virtually every government in the world uses cash accounting to report its operations. If you are a cynic you will say they do this because it allows them to sell assets and count the money as current income. This is a habit of governments around the world.

If you have a historical bent you might be more sympathetic. Cash accounting was reasonable until 1935 because our government made few promises that involved the future.

But in 1935 our government became a de facto insurance company. A very large insurance company.  The passage of Social Security put our government in the position of collecting current premiums (the payroll tax) in exchange for a promise of future benefits. Over time, the promises were enlarged, e.g. early retirement benefits at 62 instead of 65; the creation of Medicare in the mid-sixties; and the passage of the Prescription Drug bill last year.

In each instance, premiums are collected from a population that expects (and depends on) future benefits. Any private insurance company that operated on a cash basis would (a) be laughed at, until (b) the officers and directors were carted off to jail.

The problem of operating an insurance business on a cash basis has received little attention from Congress. I suspect they like it that way, regardless of party. One Senator (the late Leverett Saltonstall, MA) did manage to pass a bill in the mid-seventies that required a regular presentation of government finance according to GAAP. I have a collection of these reports in my Library of Obscure Government Documents.

They receive little attention. The report has been done for a quarter century. But it remains a little known exercise.  Like corporate annual reports, these reports are subject to assumption changes and other forms of manipulation.

Few on Wall Street are aware of these reports. In fact, I think it is safe to say that few in the investment community look very closely at government finances or understand them. The best recent example is the Disappearing Treasury Bond Panic during the Clinton administration.

The announcement of the first federal "surplus" in 1998 got the attention of bond traders. Serious investment memos started to appear, worrying what security would benchmark credit quality when we didn't have, in the manner of Richard Nixon, Treasury obligations to kick around anymore. There was a brief buying panic.

The problem, of course, is government accounting. Like a game of Three Card Monte, our government keeps three budgets. The "On" budget accounts for most government operations.  The "Off" budget accounts for receipts and activities in various government trust funds, dominated by Social Security. The "Unified Budget" puts the two together.

Prior to the Greenspan Commission and the resulting salvation of Social Security in 1983, people in government liked to report the "On budget."  Not that it was in surplus. It wasn't. But the deficit was smaller if they didn't have to add the "Off" budget deficit. There was one from 1976 through 1982. Worse, the Off budget deficit would make the public worry about those Social Security checks.

After 1983, however, the surplus in Social Security receipts over outlays started to grow handsomely. The Off budget became a cash cow.  Suddenly people in government no longer talked about the "On budget." They talked about the Unified budget because its deficit is smaller, reduced by the Social Security surplus.  Here are some sample figures:

Q. Which Budget Will We Use? A. The One That Looks Best
Note that gross government debt rose consistently through the period, regardless of reported surpluses in the Unified Budget.
Year Unified Budget On Budget Off Budget Gross Debt Held by Public
1996 ($107.5) ($174.1) $  66.6 $5,181.5 $3,734.1
1997 ($  22.5) ($103.3) $  81.4 $5,369.2 $3,772.3
1998   $  69.2 ($  30.0) $  99.2 $5,478.2 $3,721.1
1999   $125.6   $   1.9 $123.7 $5,605.5 $3,632.4
2000   $236.4   $  86.6 $149.8 $5,628.7 $3,409.8
2001   $127.4   $  33.4 $160.7 $5,769.9 $3,319.6
2002 ($157.8 ($317.5) $159.7 $6,198.4 $3,540.4
2003 ($375.3) ($536.1) $160.8 $6,760.0 $3,913.5
2004e ($520.7) ($674.8) $154.0 $7,486.4 $4,420.8
Source: Economic Indicators, March 2004

None of these figures are remotely close to the real liabilities our government has created by over-promising benefits while failing to recognize the true cost of rising life expectancies. The figures below, developed by economist Kent Smetters and Jagadeesh Gokhale, are virtually identical to figures that were supposed to appear in the Presidents Budget for 2004. They were cut at the last moment, days after O'Neill left Washington and Snow arrived.  They subsequently appeared in Congressional testimony.

These figures represent the present value of promises that will not be covered by the present value of expected tax collections. It is important to note that these are based on government projections and estimates. They are not based on any radical departure from conventional expectations about life expectancy, death rates, birth rates, and medical practices. It's also important to note that these calculations were done before the Medicare Prescription drug bill passed. That added another $6 or $7 trillion.  This increases the total unfunded liabilities to about $51 trillion. A very, very big number.

For reference, try our collective net worth. The Federal Reserve computation of the Consumer Balance sheet was $43.4 trillion at the end of 2003. While some are alarmed at the $500 billion official federal deficit, the real increase in obligations is much greater--- well over $1 trillion a year.

What Has Been Promised- Present Values in Billions of Constant 2002 Dollars
  2002 2003 2004 2005 2006   2007 2008
Imbalance in Social Security 7,022 7,204 7,436 7,692 7,967 8,258 8,569
Imbalance in Medicare (Parts A and B) 34,654 35,545 36,629 37,805 39,029 40,307 41,651
Imbalance in the Rest of the Federal Government 550 676 753 864 1,005 1,153 1,310
Debt Held by the Public 3,540 3,793 3,993 4,119 4,231 4,314 4,402
Total Federal Fiscal Imbalance 42,225 43,425 44,817 46,361 48,001 49,718 51,530
Memo Items:
Year-to-Year Change in Fiscal Imbalance -- 1,200 1,392 1,544 1,640 1,717 1,812
Fiscal Imbalance as % of PV of (Uncapped) Payroll 15.9 16.0 16.3 16.5 16.8 17.1 17.4
Fiscal Imbalance as a % of PV of GDP 6.2 6.2 6.3 6.4 6.5 6.6 6.7
Source: www.house.gov/judiciary/smetters030603.htm  (Smetters and Jagadeesh Gokhale Generational Accounting Study)













If putting it all down in a lump bothers you, Smetters and Gokhale also measured the unfunded liability in more conventional ways--- as a percentage of the present value of total payroll and as a percentage of present value of GDP. To add some perspective, federal receipts have typically been a bit under 20 percent of GDP in the entire postwar period. They peaked at 20.9 percent in 2000 and are estimated at 15.7 percent for 2004, reflecting both recession and tax cuts. Compare that to the 6.3 percent of GDP shortfall.

I want to emphasize that these are not "way out" figures. If anything, they are conservative. You can confirm this for yourself if you are willing to dig around a bit in the recently released annual reports from the trustees of Social Security and Medicare. In the report for 2003 the trustees first reported on the "infinite horizon," a method of calculating liabilities that goes beyond the traditional 75-year period. The addition received a great deal of attention because the liabilities of Social Security tripled when the horizon was extended.

A little thought will tell you why. When the Greenspan Commission saved Social Security in 1983 they did it on a 75-year projection basis. Yet the system is in trouble again and it's only 20 years later. This is a recurring event. Why? Because each time the trustees roll forward a year they substitute a new high life expectancy year for an old lower life expectancy year.

In very rough numbers, life expectancy has advanced 2.5 years a decade, a change of nearly 19 years each year rolling forward. As long as expectancies continue to advance, the 75-year horizon will be a "low ball" estimate of future liabilities.

The table below assembles 75 year and infinite horizon estimates that are spread through the two trustee reports for 2004.  As you can see, there is a $32 trillion difference. In other words, officially reported government debt held by the public (including foreign holdings in China, Japan, the Middle East, and Europe) is only one-fifth of government obligations on a 75 year horizon and 1/12th using the more meaningful horizon.

Financial types, who tend to suffer from excess reification, may look at this and say that Treasury Bonds are real but Social Security is only a promise. In fact, they have it backwards, government debt is only a promise made into bonds. If you personally don't understand the difference, visit a Wal-Mart and ask the nearest Greeter which is more real, Social Security benefits or Treasury obligations.

The $32 Trillion Lump Under The Rug
Program 75 Years Infinite Horizon
Social Security $  3.7 $10.4
Hospital Insurance $  8.2 $21.8
Part B (from general revenues) $11.4 $23.2
Total $23.3 $55.4
Source: www.ssa.gov/OACT/TR/  2004
Which brings us to an interesting question. If our finances are really this out of control, why aren't we in a crisis?

One answer is less than reassuring. While the government of the United States is heroically over-promised, we look really good compared to almost every other nation on earth. If others point to the gossamer nature of this emperors clothes, they will have to acknowledge being stark naked themselves. Like Goldilocks visiting the home of the three bears, the porridge in our bowl looks just right compared to the others.

Before you get too depressed, I want you to remember one thing.

This is what some call "a Cadillac Problem"--- what's important is not the ailment our Cadillac has but the fact that we have a Cadillac. Other nations are facing far worse problems. Look around the industrialized world and you'll find a need for millions of those T-shirts from the 70's, the one's that said:

"I can't believe I forgot to have children!"

We've been so fascinated by our growing lifespan that we've forgotten procreation. (Hard to believe, isn't it?) The consequence, for the industrialized world, is the demographic stresses and change pointed out by Martin H. Barnes.  By 2050, he points out; the population of Yemen will exceed the population of Germany. Can you believe it!

Given these realities--- that everyone wants to live a long and healthy life and that we seem determined to have fewer children--- the "compact between generations" isn't a good idea. Whether you are left winged or right winged, financing retirement and health security on the backs of the working young is a tragically misguided idea simply due to rising life expectancies. Add our collective failure to cement the contract enough children and the fulfillment of centuries of human dreams works to doom our society to failure.

That tells me we should be thinking about this differently.

In fact, we need real returns from real--- not virtual--- assets to accommodate the cost of financing longer lives. The table below shows how much of our income we have to save, at different real rates of return, to finance our lengthening retirements.

The Magic Bullet of Investment Return--- and the Need for Real Saving
This chart shows the percentage of income that must be saved to provide a sum equal to the number of years of life expected at 65. In each case, the amount of income to be replaced has been reduced by the amount of income saved. As a consequence, the amount that can be consumed rises rapidly with real return.
Real Return 12.6 years (1935) 17.6 years (2000) 20.6 years (2040)
0% 24% 31% 34%
2% 17% 22% 25%
7%   6%   8%   9%
Source: The Coming Generational Storm, MIT Press, 2004, pg 216
We can't get this from a government program for two reasons. First, transfer payments are a substitute for real assets. Second, the amount of assets required is enormous. And growing.

Ironically, we seem to be moving in the opposite direction from what is required. And we're doing it at all levels.
  • Our government continues to expand the promises of Social Security and Medicare, encouraging the illusion that individuals can save less because they will have the virtual wealth of government benefits.
  • Our corporations have clearly signaled that the Age of Corporate Benefits is over. Defined benefit pensions are being scrapped; health insurance after retirement is being terminated, and financial security benefits in general are being reduced.
  • Households, in the aggregate, are saving less, not more.
It's not a pretty picture and it will change how we invest our money.