Shiller on Trills

In this short piece, Robert Shiller explains one of the basic ideas of his work on macro markets:

The governments of the world should issue shares in their GDPs, securities that pay to investors as dividends a specified fraction of GDP, in perpetuity (or until the government buys them back on the open market). Governments need to end their historic reliance on debt financing: governments issuing shares in GDP is analogous to corporations issuing equity. My Canadian colleague Mark Kamstra and I propose issuing trillionth shares in GDP, and so to call these “Trills.” Last year, a U.S. Trill would have paid $15.09 in dividends, a Canadian Trill C$1.72. The dividends will change every year as GDP is announced, and predicting these changes will certainly interest investors, just as in the stock market. Governments can auction off Trills when current government debt comes due and needs to be refinanced, as part of a debt reduction program.

In this piece, Shiller focuses on the benefits of Trills as opposed to debt:

Substituting Trills for conventional debt helps deleverage the government, something whose importance has become very clear with the debt crisis in Europe.  The payments required of the government by the Trills is connected to the country’s ability to pay, measured by their GDP.

Trills could also be the foundation for many types of insurance products, for example, products that would pay off when GDP was down helping to alleviate business cycle issues. A market in Trills could also be used to make predictions and to judge policies (see Gurkaynak and Wolfers for an early test). Which policies will most increased the value of future trills? Similarly, by looking at how the market for trills changes as the Iowa Political Markets change we could identify which politicians are best for GDP (not just the equity and bond markets).

I featured Shiller’s work on macro markets in my book Entrepreneurial Economics: Bright Ideas from the Dismal Science. I think of this body of work as his most visionary and deserving of the Nobel.


"we could identify which politicians are best for GDP." You mean the luckiest ones?

I expect strong demand for these assets from Lil Wayne

How is this really better than debt financing? Equity interest in most entities means you get whatever's left in the event of a bankruptcy, but how does this really stop the government from defaulting? Do you get to repossess the country? At some point, you'd run into the same issue of x% of GDP being simply too much for a government and economy to bear and there would be a default (in some form).

And who gets to measure GDP?

It's interesting as a thought experiment, but this doesn't come close to solving the issues of governments taking on too much debt. It's just a new debt package.

Yes. Basically we already do this. It's what the dollar is.

Whichever economist runs the department that determines GDP would have money and women like nobody else.

This x10. Like all metrics, GDP is imperfect and subject to much manipulation. The more importance you attach to a metric, the harder people will work to manipulate it to their advantage.

Amen! My experience at a company having quality problems in the 1990s was this: while the metrics were instituted and run by the operations people, actual quality improved (in part because the operations people tended to be compensated on a different metric, e.g. client satisfaction surveys). The year the CEO announced that his bonus would be dependent on a certain percentage increase in the quality metrics, the system became corrupted in short order -- the metrics increased, of course, but became unreliable.

With regards to GDP, consider the case of Ghana: 60% change in GDP overnight in 2010.

Wait, they don't already!?

(In all seriousness, a very good point. I like the basic idea but think that there are too many moral hazards/measurement difficulties/[who is paying that money, exactly?] to ever be done properly.)

Maybe. As someone who thinks it is owners of wealth who should naturally think of the $17.25 trillion of federal debt for services rendered as "their" debt, this wouldn't affect my worldview much.

There is only one test: would it arrest spending?

This is an absolutely terrible idea. The instruments Shiller proposes are nothing at all like equity securities. First, company shares pay out based on the profit that a company makes on its own. A country's GDP isn't produced by its government, so any money for the payouts on these instruments would have to be appropriated through taxes. Second, the earnings on company shares are net of expenses, while the earnings of Shiller's trills are reckoned on gross output. Third, even after deducting expenses, a company almost never pays out 100% of its net earnings on dividends, while trills instruments allocate all of the "earnings" attributed to the instrument to dividends. Fourth, equity shares can be diluted, but trills cannot; a trill promises a trillionth of GDP, no matter how many other trills have been issued.

And what would the effect of these trills be? Shiller touts their ability to ease the burden of servicing debt when GDP drops in a recession, but that's about all they have to commend them. They also increase the burden when GDP grows. In fact, because they are indexed to GDP and perpetual, the country who issued them will always have a debt burden of GDP * x / 1 trillion, where x is the total number of trills issued over its history. Thus, you cannot outgrow them the way you can with a fixed nominal debt. A country that gets into debt trouble by issuing too many trills literally has no way out because even if they manage to get their economy growing again, the growing burden of the trills sucks up all that surplus without giving them any opportunity to pay down the principal.

One final effect of the non-dilutability of trills would be that there would be a practical limit on the total number of them that a country could issue over its entire history, determined that the fraction of the country's GDP that it can feasibly devote to debt service. Beyond that point a country would be forced into a balanced budget. Shiller is a smart guy, so I doubt this consequence has escaped him, and perhaps he even considers it a feature, not a bug, but he carefully avoids mentioning it, which is a bit dishonest. In practice, however, countries would blow right through this cap because it's not a hard boundary, but a fuzzy one. The temptation will always be to push it a little further down the slippery slope because, if we're managing now, then surely another 0.01% of GDP for debt service won't kill us, right? Once that happens, see previous paragraph.

In summary, if I were trying to design a policy to guarantee a country's fiscal collapse and eventual default, it would look a lot like Shiller's proposal. I don't know what his game is, but the only winning move is not to play.

Good points, but, like bitcoin, this idea might be useful since it has a cap and it would limit the amount of 'debt' (or equity) a government could issue. With debt, there's no limit (even the Republicans will admit that).

There is the same practical cap. Outside investors either buy them or don't. There "should" be a hard cap where payouts would be 100% of GDP, but that is irrelevant as there "should" be the exact same cap for debt issuance where coupon payments hit 100% of GDP. The real cap will always be related to the markets' interest in buying the debt and there is little reason to think this changes it significantly.

Additionally there seems to be little difference in the value of the trill between 2% growth and 0% growth because the payout is based on GDP, not GDP growth (basing it on GDP growth has its own problems). A payout of $15 per share in 2014 with 0% growth and a payout of $15.30 with 2% growth is not proportional to the difference in damage an economy takes- and the difference in the tax base- between a 0% and 2% change. You absolutely have to go to a flat tax in this scenario or you end up bankrupt as your ability to pay and obligations diverge very quickly.

But the equity price would collapse, and the government issue would raise less revenue for the same payout.

I presume that the equity price is not set by the Federal Reserve.

If you price perpetuities, there is a huge difference between 0% and 2% growth.

There is, but in this case it works to make the situation even worse. When do you want to borrow? When times are bad, like in recessions, right? But the low or even negative growth prospects in bad times mean those are the times when the trill securities will be worth the least. When do you want to retire debt? When times are good, right? But the improved growth prospects mean that trill securities will have seen a lot of capital appreciation since you issued them. Therefore, a government trying to execute responsible counter-cyclic policy will find itself selling into bear markets and buying in bull markets, which is exactly the opposite of what you want.

Really, the closer you look at this proposal, the uglier it gets.

Yes, but the difference of a single year of 0% growth for the governments ability to generate revenue is greater than the 2% decline in expected payout that year. Combine this with the fact that the government can raise less revenue because the price of trills would drop due to the recession and you have a problem.

Re: "They also increase the burden when GDP grows."

This might be a benefit, not a cost: it would be countercyclical: during a boom, there would be less government spending.

It would also create a security that would discourage unfinanceable tax cuts.

during a boom, there would be less government spending

I don't think that's necessarily true. It's true that the expanding expenditures on interest, or dividends, or whatever we're calling these payments would squeeze out other government spending, but that by itself is not enough to make a policy countercyclical. If the government wanted to reduce total expenditures during boom times it would have to reduce the non-interest part of the budget by more than the interest part expands. If anything, that makes it even harder to do a counter-cyclical policy.

Re: "It’s true that the expanding expenditures on interest, or dividends, or whatever we’re calling these payments would squeeze out other government spending, but that by itself is not enough to make a policy countercyclical,"

I don't understand. If the payments squeeze out other parts of the budget, then you decide NOT to invest in other government projects.

You could get the same result by having an infrastructure lend for infrastructure during poor times, and when times are good, and rates for private investment are higher, you don't lend.

"Countercyclical" doesn't mean reducing "investment in other government projects" in good times. It means reducing the amount of borrowing (or increasing net savings, should your government find itself in that happy circumstance). In order to do that, you must either reduce total spending, including interest payments, or increase revenues. Either way, it's harder to do if you have a big surge in interest payments.

Your definition of countercyclical is not mine, and I've never heard of yours before. Countercyclical means reducing spending during the upcycle.

No wonder Bush had problems.

"The temptation will always be to push it a little further down the slippery slope"

Only if those in charge are so short-sighted as to risk the long-term economic viability of their government for short-term gain. In which case, a more rapid "fiscal collapse and eventual default" would be beneficial, particularly if it brings regime change to one with a higher future time orientation. It would bring Darwinian selection to nations; if the more reckless collapse, they will be out-competed and replaced by those who are better able to balance the well-being of future generations against those of the present.

Fiscal collapse and default is never "beneficial" to the people who have to live under it, the overwhelming majority of whom had no hand in bringing it about. In fact, it often happens that the very people who caused the collapse are the most shielded from its ill effects, and that greatly dampens their incentive to avoid disaster. Short-term thinking has been endemic to pretty much every form of government that humanity has managed to come up with so far, and I see no reason to expect that to change any time soon.

Darwinian selection doesn't work well with nations in practice, as even a cursory review of either current events or world history will reveal. Indeed, it's not clear to me that it's even possible for Darwinian selection to operate without heredity as a mechanism for ensuring that bad traits are expunged from the gene pool. Unlike bad genes, bad ideas can rise from their own ashes, and in this case there are a lot of public choice reasons to suppose that they will do just that.

rpl is making the most sense on this thread. I am confused as to what Shiller thinks the benefit would be. There seem to be none for a country like the United States with its own currency. Maybe he is thinking about a place like Greece? But is he expecting counter-cyclic fiscal policy coming out of this new approach? He hasn't really explained how that's supposed to work.

"Fiscal collapse and default is never “beneficial” to the people who have to live under it, the overwhelming majority of whom had no hand in bringing it about."

But neither is a long, drawn out flailing decline, likely followed by fiscal collapse; getting it over with more quickly reduces the total damage, and provides greater opportunity for reform.

"Short-term thinking has been endemic to pretty much every form of government that humanity has managed to come up with so far"

Perhaps, but it's definitely a far bigger problem for some systems than for others; the Ming Dynasty was definitely on the ball, as it were. And aside from the occasional bad result of inbreeding, and an unfortunate religious schizm, the Habsburgs were pretty good at the long game.

In fact, short-termist financial mismanagement seems to be primarily a product of systems where those in charge are only so for a short term, and must pander to the irrational prejudices of the empty-headed mobs to get or stay in charge; those who expect to pass the nation on to their progeny have much better incentives to proper long-term management.

Can we assume that the GDP numbers are as easy to fake as the unemployment numbers? And if so, would we not simply be handing whoever is in power a slush fund to manipulate at their pleasure?

First, I totally thought this was going to be about the economics of symbiotes in Star Trek.

Second, this smells like it has similar problems to Bitcoin: it's a deflationary currency. Someone can just get a bunch of trills and then hide them in his mattress and accumulate value that way instead of investing actual capital into the economy.

it’s a deflationary currency.

Like they used to have in the bad old days of gold and silver? Sounds good to me.
Why would these go under the mattress when you could loan them out instead? Even if they did go under the mattress, wouldn't your reduced consumption (relative to spending them all) make resources available for investment?

> First, I totally thought this was going to be about the economics of symbiotes in Star Trek.

I'm pleased somebody mentioned this. It was my first thought as well.

I think you could start small, and limit the equity to projects where the government gives a subsidy (infrastructure, R&D, etc.) based on the premise that the project would not be successful, but for government support, or that there were measurable values to public goods.

The easiest to do would be highway or public transportation projects, with payment to the stockholder based on usage, reduced delay time, or even tolls which collected revenue. This might also expose some projects, ie, ballparks, tax increment financing projects, etc. as useless because investors would not sign on to the project.

Would you invest in a public school where your payment back for better performance would be a dividend? Could you make your dividend or stock a charitable contribution?

Isn't that how they built the Erie Canal?

I like the thought experiment but shouldn't the metric be federal revenues minus federal expenditures? The government is the relevant unit, not the country; the country is a regional division of several hundred million people, whereas the government is the provider of legislative and violence services to that regional division. Since the government is issuing the 'equity' it should be pegged to the government's performance.

In which case basically every state corporation law would prohibit the US government from ever paying a regular dividend, since it's running at a loss almost all the time. Occasionally it would be able to pay a dividend (particularly a nimble dividend) but for the most part a company cannot pay out profits when it has none. Tax law for earnings and profits would agree that any distribution would not be a dividend but a return of basis (conceptually parceling out the principal asset).

Also, in case it's not immediately obvious, it's horrible to create an investment vehicle that pays more based on excess tax revenues. Sure it creates a constituency to cut spending but it also creates a constituency to raise taxes in order to distribute them to what would almost certainly be mostly wealthy investors, universities, banks, etc.

Yes, but the price they sell for would reflect that.

Surely this would be a significantly more expensive source of financing than Treasuries? The payoff would be more uncertain and payments would be a lower priority than Treasury interest. Alternatively, I would think of the dividend yield as roughly equivalent to that of a large corporation whose earnings are growing at a 2-3% annual rate. Either way the dividend yield would be higher than the risk free rate (maybe 5%?), and so government obligations would be higher unless you had a major depression where GDP fell 30% or more. In that event, sure you'd be able to issue more Trills to meet your obligations, but they'd still be riskier and higher yielding than new debt, right?

I guess it would be more acceptable to selectively default on Trills, so in that sense the government would be paying down debt through a tax on gullible people, but then why not just have a federal lottery and call it a day.

His example is Exxon hiring an unemployed engineer for 100k that finds no oil. GDP is unchanged. But when the federal government hires an unemployed engineer for the same salary that finds no oil, GDP increases by 100k.

GDP is a better measure of the economy than anything I can come up with but you can't solve government spending problems by tying it to government spending.

But the GDP only determines the dividend. The price of the equity is determined by the market participants expectation that there is a difference in Exxon spending that $100k compared to the government. The dividend may rise, but the equity price would drop.

Presumably the idea is to get the equity issue price based on the market evaluation of whether the spending would produce growth or not.

So in 2002 Bush would have had to raise $600 billion for a war. What would have been the price for the equity? TARP in 2008. Stimulus in 2009? $1.5 trillion or close for a couple years after.

In Canada since the deficit cutting days of the 80's and 90's, especially in the provinces, public policy had a direct effect on revenues. With health care spending going up by 6-10% a year, and it being over 50% of expenditures, and the political inability to deficit finance, jurisdictions were forced to put forth policies that maximized revenues. The Laffer curve is very real, and we saw very good government and very good policies for a few years. Bad policy was reflected in health service statistics; waiting time for surgery, etc.

The only way this type of scheme would have the same effect is if we saw equity price fluctuations in response to government policy. The debates on deficit spending multipliers (remember them?) would be the price of the equity.

This feedback is already there, but politicians are stupid and need a two by four across the head before they pay attention.

Debt financing provides the same feedback mechanism in yield. Unless the Federal Reserve in their wisdom buys the bonds to set the price.

There is the question of whether they could raise the money spent on the wars. I found the case for the wars sound, but found that it lead to way higher spending than should have been necessary. Perhaps we could have achieved the same or better results if we had to raise the money in a more constrained manner.

A most thought-provoking read by Prof Shiller. Here are my thoughts: wtf? omg. lol. lmfao. A virgin continent of unactionable securities fraud? Sure, why not? A vast new market in risk-free derivatives? Gotta love it.
Ease the debt burden by taking on more obligations? Counter-intuitive! - must be good. J.G. Wentworth awaits the call from Uncle Sam.

Never underestimate the ability of econ PhDs to be economically illiterate.

One problem, most investment in treasuries is done to get a small return on ensured cashflow. That was the main credit-rating concern over the debt-ceiling debates. The primary credit rating concern is over high levels of spending. But the debt ceiling presented a short-term cashflow risk, which is a huge factor for these investors. They are look for a small return on idle cash, but need to be certain of cashflows to meet fixed obligations, or have them at likely times (eg. to have cash availible to pay a contractor if he meets a milestone).

On the other hand, these would make good retirement account invesments.

This is the purpose I could see for them. It wouldn't be a bad idea for that purpose, honestly. But I'd want to keep it small and use normal debt for the bulk of financing. I've always thought of SS as essentially providing this function, but with more political risk and heavy distributional effects.

I think that it might be good if retirees were paid an amount equal to total amount of FICA taxes collected divided by the number of eligible retirees.

With or without wage increases?

Because in the U.S., GDP seems to be increasing without wage increases, which is the baseline for Social Security. Almost as if someone had discovered a way around the idea of a 'Generationenvertrag.'

Except for thew awkward, usual gap between your Narrative and reality. 11% increase in median real wages since 1990 according to Social Security.

The US government does not own the economy, neither its assets nor its liabilities. It is a horrible idea to create an operational myth that the government owns the GDP. The government's ruling faction of the moment has the ability to tax the economy, impose its idea of a currency on people who reside there, and use force to extract from earners the amount the moment's faction wishes. Extending or encouraging any sense of government ownership of GDP an extremely dangerous game. Fortunately the constitution does not grant to the government of limited powers the right to claim ownership of the GDP.

No wonder we had a massive financial crisis. The currency's value is systematically manipulated in ways said to promote national prosperity but actually designed to assure growth of the financial sector and nominal asset prices. And look where that prosperity went. Hand in hand with the Fed's sense of duty to its votaries, the Legislature feels entitled to commit more and more of your future earnings And you like that? The redistribution of wealth towards financial owners was said by the Fed to insure employment did not shrink. But, as more and more jobs become part time and without benefits, it became obvious that it is the jobs themselves, rather than their absolute number, that shrank. That is a disastrous equation.

Re: "The US government does not own the economy, neither its assets nor its liabilities. It is a horrible idea to create an operational myth...."

This is a synthetic instrument. Ownership of the economy is irrelevant.

The own us...that is why they can force us to pay them as much of our income as they want. Have you learned nothing?


Please explain. I do not understand your comment.

Who exactly is going to make the insurance claims when GDP is down?

Sorry, the number you have dialed has been disconnected.

Just a friendly reminder: there are no existing political futures markets that have the volume you would need to treat them as even vaguely serious. Pretending that the Iowa Electronic Markets can play this role adequately does an enormous disservice to the general cause of prediction markets.

Talk about self-selection problems

UGK and R. Shiller present :

Too Trill: GDP Edition

If the professors are testing us, then I'm flunking, because I don't understand either of them. How could trills become the foundation of insurance products that pay off when GDP is down, when the whole point of trills according to Shiller is that they pay *less* when GDP is down? Is AIG going to be writing trill decline protection?

Governments don't "own" GDP. Linking payments to something that they don't own causes all sorts of mismatching problems. Better to sell equity in the right the government has to collect tax.

Wouldn't paying nominal GDP be an infinite series which might be diverging?
Shouldn't GDP growth rates be used (though floored at zero)

E.g. If USA had nominal GDP of 15 trillion and expected to grow at 3% while nominal interest rates on 10y treasuries are 2% (say) these could end up being massively expensive.

PV = infinite series of NGDP discounted back to today

Comments for this post are closed