Crunch time is here for the Trump administration, and for us all. President Trump’s economic policies, in particular his deregulation and an ongoing rise in interest rates, will if they work as advertised produce a surge in productivity growth that will allow the United States to pay the bills coming due. If the productivity growth does not materialize, or if other less benign Trump policies pull the global economy off course, it is difficult to see how we can avoid economic collapse and universal debt default. Things could genuinely go either way.

Modern societies have tolerated an almost infinite increase in government, in particular in the welfare sector, and have paid for it through boosting productivity at a rapid clip. The exceptionally rapid increase of 2.8% annually in U.S. productivity in the post-war period of 1948-73 paid for various wars and the creation of the Great Society welfare system, which in its early years appeared actuarially sound. Even the lower 1.9% productivity growth rate from 1973 to 2010 allowed for the extension of the welfare state, an explosion of state and local government spending, and an absurd escalation of medical costs, itself exacerbated by foolish government decisions.

Only since 2011, when productivity growth has dropped to 0.6% per annum (and only 0.2% in the latest full year from 2015 to 2016), have the sums definitively failed to add up, with government debt doubling in the last 8 years and no end in sight.

Rapid productivity growth is the only factor that makes today’s bloated government accounts add up. Before the Industrial Revolution, with very low productivity growth, government’s take was more or less capped at around 10% of GDP. Governments that attempted to take more than that – for example the French ancien regime attempting to keep up with British military prowess in the eighteenth century – found themselves consigned rapidly to bankruptcy or revolution. Only during the Napoleonic Wars did Britain push government spending up to the hitherto unprecedented level of 20% of GDP without suffering bankruptcy.

That should have made it clear that the new technologies, which had enabled Britain since the 1780s to enjoy continuous economic growth at hitherto unprecedented and accelerating levels, had also enabled government to grow beyond its previous bounds. However, the fiscally cautious managers of Britain’s austere post-war period instead brought spending back in line with revenue, even including the bloated debt service requirements, and thereby paved the way for a century of small-government growth.

It was only World War I in Britain and the Great Depression in the United States that allowed government-expanding liberals to discover that the old fiscal limitations no longer applied. With productivity growth rising at a rapid clip, they could make long-term promises to voters, such as Social Security and Medicare, and the economy’s inexorable rise in productivity would increase everybody’s wealth and the fiscal take sufficiently for the promises to be kept and the books to be more or less balanced.

This happy capability (except for those of us who believe in small government for its own sake) may now be endangered. It is achingly clear that the annual productivity increases of 0.6% we have seen since 2010 are nowhere near enough to allow the books to remain balanced, or even stably out of kilter. If indeed Robert Gordon is right, and rapid productivity growth has come to an unexpected end, the nation’s and the world’s finances are doomed.

Social security payments are linked only to prices (though there is a hidden partial earnings link before retirement), so the system cannot get too far out of whack except to the extent we live longer, which can be paid for by extending the retirement age. However, Medicare payments are linked to medical costs, which increase faster than inflation, and which escalate as new medical technologies are discovered. An annual productivity increase of 0.6% is nowhere near sufficient to cover the long-term escalation in Medicare/Medicaid costs.

In addition, the U.S. economy has inefficiencies built into it which exacerbate the funding problem. It has relatively high immigration, at a skill level lower than the average of the current population; which means immigrants pay less in taxes and receive more in transfer payments than average, worsening the overall burden. It also has a tendency, at least during Democratic administrations, to burden the economy with additional regulations, creative new forms of disability, and creative new grounds for lawsuits, all of which push the fiscal position further from balance.

Thus, while with 1948-72 productivity growth the nation could add new social programs and pay for them, and with 1973-2010 productivity growth it could just keep existing programs funded, with 2010-16 productivity growth the position is hopeless. Even an immediate return to the 1.9% annual productivity growth of 1973-2010 will not be sufficient to balance the books, because of the additional imbalance caused by the 2010-16 dearth. At this point, the country must achieve at least a decade or so of 1948-73’s robust 2.8% productivity growth to restore the fiscal balance, after which it can relapse to the 1973-2010 level.

If Trump can restore productivity growth to its 1948-73 level, at least while he is in office, and is not succeeded by another hyper-regulatory President Obama, and doesn’t waste more than modest amounts on futile infrastructure projects, then the Social Security and Medicare funding problem will go away fairly painlessly. The additional productivity will increase Federal revenues and Social Security premiums sufficiently that the actuarial deficits in the trust funds will diminish, and any remaining hiccup because of Baby Boomer retirements will be funded painlessly through borrowing.

If Trump reverses the productivity damage of the Obama years, but raises annual productivity growth only to its 1973-2010 average of 1.9%, then the Social Security and Medicare funds’ deficits may be solved, but only through stringent economies elsewhere. The current budget deficit, above $500 billion, would decline initially, but increase again in the next recession, before remaining at a dangerously high level that could be addressed only by tax increases or budget cuts that went far beyond the usual “waste, fraud and abuse.” Even the steps Trump might take to reduce the budget problem, such as limiting low-skill immigration, legal and illegal, would probably not be sufficient to counter the overall drag of ever-escalating medical costs and increasing waste in areas such as infrastructure.

In such a case, Trump would probably not get credit for the partial recovery in economic performance, which would have restored productivity growth fully to its level prevalent over the last 40 years, but would be defeated for re-election in 2020, as voters noticed their pain more than his achievement.

If Trump fails to raise productivity growth from its level of the late Obama years, then the Medicare and Social Security trust funds are bust in the medium term, as is the United States. With current levels of productivity growth, even the Social Security trust fund is due to go bust in the 2030s, with no countervailing thought that “they’ll be richer by then” – without faster productivity growth, they won’t be. The U.S. healthcare system in general, which already absorbs some 18% of GDP, will continue to swell its share of output, boosted only partly by the health travails of the baby boomers’ last years, and the Medicare/Medicaid trust funds will thus predecease Social Security. Finally, the Federal deficit will continue to grow, slowly at first but accelerating rapidly in the next recession.

Full national bankruptcy may be avoidable during a re-elected President Trump’s eight years in office, but it will a close-run thing. Fortunately for Trump, in these circumstances he is very unlikely to be re-elected; the GOP will then have fun blaming national bankruptcy on Elizabeth Warren.

Of the three trajectories outlined above, the middle is the most likely. The first would require truly superior economic management. To get back to pre-1972 rates of productivity growth, we must optimize capital allocation, by raising short-term interest rates sharply, and reduce regulation to its pre-1970 level. Purists might object that environmental and other regulations have done much good since 1970; all I would respond is that it is surely worth suffering a few burning Cuyahoga Rivers to avoid national bankruptcy. In addition, enthusiastic economies would need to be made in the remainder of the Federal Budget, draconian tort reform would need to be implemented, and infrastructure spending would have to be avoided until infrastructure costs are once again sensible. We can wish for such a Nirvana, but should not hope to get it; even Trump cannot Make America that Great Again.

On the other hand, interest rates do seem likely to be raised a bit, bringing capital allocation a little closer to optimal, and Trump is at least partially de-regulating. He is almost certainly going to waste a fair amount of money, and will cause inflation by doing so, but it does not seem unreasonable to suppose that he can raise economic policy to the modestly successful levels of Presidents Nixon, Ford and Clinton, thereby achieving productivity growth at the 1973-2011 level, albeit accompanied by some inflation. Robert Gordon’s productivity Armageddon will be at least postponed.

That suggests that this column will continue to have work to do in the decades ahead; economic policy will remain imperfect, and the U.S. economy will teeter on the edge of national bankruptcy without actually falling over the cliff. That’s a relatively optimistic outlook by this column’s standards; it does not however justify a Dow Jones Industrial share index above 20,000.