Our Grandchildren's Debt Burden in 2050: 540 scenarios

Last weekend I was leaning over the kitchen sink eating a turkey sandwich, and contemplating several gloomy articles I'd just finished reading.  All had mentioned the economy our grandchildren "could" inherit.  All scenarios were bad news, and were, of course, our fault for not being “fiscally responsible” today.  Some were written by people with economic perspectives I highly respect; others by analysts with skills I highly respect; others by journalists or politicians with economic knowledge that commands little to none of my respect.  But the common thread was the terrible scenario that “could” come about. 

The key word is “could.”  The use of that word in all those articles implied that maybe, just maybe, there might be some other scenarios, not so gloomy, that also “could” come about.  So I decided to spend a few evenings testing the possibilities. 

Sure enough: It turns out that in many plausible scenarios, the future’s so bright I gotta wear shades. 

Rather than bore you with the methodology details, I’ll summarize how I did it, then show the results.  First, I combed several articles—some gloomy, some not so gloomy—by people whose perspectives I respect.  Here are just a few examples: • Alan Greenspan (Fed) on fiscal risks; • Brian Riedl (Heritage) on the looming fiscal disaster; • John Ince on the debt time bomb; • Kane & Hederman (Heritage) on growth; • Alan Reynolds on immigration; • Steve Forbes on immigration; • Arnold Kling on growth.

Then I found the US Census Bureau’s projections of population by age group through 2050.  Click on the little thumbnail at right to see a chart I assembled from the data.

Then I set up a sensitivity analysis model, using a “Red-Yellow-Green” matrix to help illustrate, in compact form, how the economic outcomes of a complex economic model shift favorably or unfavorably for hundreds of possible combinations of key inputs.  In this analysis, the model calculates what our economy “could” be like in 2050, based on a range of possible results for several key inputs from those articles I combed.  I chose the Debt-to-GDP ratio as the key output measure for each of hundreds of combinations of the key inputs; there are others we could choose—such as "interest as a percent of tax receipts"—but anyone familiar with this weblog should be very familiar with the Debt-to-GDP statistic by now.  (It's widely accepted by economists as a measure of a nation's debt load, and it compresses a significant amount of economic performance information into one single statistic.)

That’s enough of the boring background; below is the “Debt-to-GDP 2050” matrix that resulted from this week's analysis.  (Quickly: I assumed 3% inflation, 13% GDP on-budget tax receipts, and off-budget tax receipts per worker growing at a smidgen more than inflation; remaining assumptions are explained below the matrix.  See end of this article if you'd like to challenge any of those assumptions.)

It’s immediately obvious that productivity growth (input #4), the key driver behind GDP growth, is of overwhelming importance; look how green it gets as we move to the right in the matrix.

Click to enlarge.

It’s also obvious (...to me, anyway) that the pessimists and political ax-grinders like to dwell on the lower-left, bright red portion of the matrix—all those bad, ugly scenarios that “could” happen (...unless, of course, we vote their political group into power so they can fix things right up for us). 

Let’s be honest: Could one of the scenarios in sections A, B, C, or D come true?  Of course it could, especially if we develop collective amnesia about what made the USA the rich country it is today.  But let’s tell the whole truth for once: Could one of the scenarios in sections E, F, G, H, or J come true?  Of course it could.  It depends on many things, but especially on the growth results our economic system is able to produce.  (Note that if Ray Kurzweil is correct about growth taking off exponentially, as Arnold Kling’s article points out, the economy in 2050 will be about a mile to the right of section “J” on the diagram above.)

Here's one last graphic, comparing the doomsday lower left corner with the happy-day upper right corner:

Note that the best scenario in the analysis, shown above, is an economy 45 years from now that's running a fiscal deficit of $10 trillion, paying interest on the debt of $4 trillion, and carrying a public debt of $101 trillion.  I bet the debt phobes and political ax-grinders will have a field day with those numbers—and I also bet they’ll conveniently forget to mention the $1 quadrillion economy that turns those numbers into chickenfeed.

Comments are open on this post; anyone and everyone willing to mind her manners is welcome to participate.  Same applies to any man willing to mind his manners. 

================

End note, for those who want to dig into the assumptions:

Please send me an email if you have some questions about the assumptions I made, or think any of them could be improved, including productivity growth, tax receipts, incremental immigration, interest on the debt, on-budget spending, or the growth rate of social insurance cost-per-old-age-person.  I am happy to get questions, and I answer every email I receive.

However, please don’t bother trying to convince me that National Security spending should be lower.  [“National Security spending” is defined as the sum of intelligence, diplomacy, defense, homeland security, and civil defense spending.]  In this model, it will remain pegged at 4½% GDP; anyone who disagrees with that should feel free to build a different model on their own, and should be prepared to defend it publicly. 

In this model, National Security spending will remain pegged at 4½% GDP, for two reasons:

(1) I’m assuming we learned our lesson after allowing it to drop to 2.9% during the regrettable late 1990s period of surplus-worship; and

(2) My granddaughter’s security is nonnegotiable.

Comments

[Morgan]

Some of the scenarios are more realistic than others. For example, it is highly unlikely that social insurance per old-timer will rise at an 8% annual clip while productivity growth averaged 2% - that would imply social insurance per person 65+ greater than production per worker by 2050.

I also imagine that interest rates on "debt to the public" will correlate positively with productivity growth rates.

And maybe on-budget expenses as a percent of GDP will fall with increasing growth rates - on the optimistic assumption that, at a certain point, our government will have a hard time finding ways to spend our money.

[angus]

10.7% on budget spend per %GDP is current 2005 figure, but the table maps only for 11% and lower. This pessimist assumes that politicians can always spend more money and sees the table as skewed to the green.

[simon]

EXCELLENT POST!!!

This type of thining is truly value-added. It provides context for the debate we nee to have in our nation. Great Work!

[Joe Doyle]

Very clever visual device. Do you intend to publish your underlying model, so we can eyeball whether it makes sense? You assume 3% inflation per annum - why don't you make that one of the variables? I would swap that for immigration, as the scenarios seem relatively inflexible to variation therein.

I also second Angus' observation that your on budget spending assumptions are optimistically biased. Does Social insurance include Medicare/Medicaid? If so, 8%/annum growth is not so far-fetched, and now you've got me worried!

The inflation variable seems big to me, since as we previously discussed it could go MUCH higher should Debt/GDP become unsustainable. Of course then you'd have to worry about feedback loops between inflation, productivity, and nominal interest rates. Leave it me to complexify a great tool!

[tom]

What a great presentation of data in a way that makes it meaningful.

However, it seems to me that box A (low productivity and high spending on social insurance for the elderly) is a more realistic box than box J (high productivity and low spending for the elderly).

I believe this for two reasons.

First, the jump in productivity over the past decade is directly linked to information technology developments that made much of the transfer of information in our economy more efficient. Information is the essence of technology, and it is not obvious to me that another economic sector (like information) exists that can be made more efficient with further technology developments. (That's not to say that technology will not continue to add to efficiency -- for instance, one of the big reasons health care is so expensive in the United States is because very little health information has been digitized. But, as information is made more efficient, we will see less and less of these productivity gains.)

Second, even if health care is made more efficient, it will become more expensive. We surely will see fantastic developments in health care over the next fifty years -- but these developments, and their corresponding benefits, will be expensive. Drug companies will make life-saving and life-extending products, but they will patent these products and charge more than they do now. I do not envision a federal government unwilling to pay for the continued health of its (voting) senior citizens, even if such costs do annually increase 6-8%.

So, a question: if productivity is low, and health costs are high, does it become more likely that our grandchildren will live in a box that looks like A? Or, are there some other variables that paint a Box-J picture?