Some thoughts on Chinese manufacturing dominance and deglobalisation

Richard Baldwin has multiple blog posts that mine the data to add nuance to the Chinese manufacturing dominance, which, while undoubted and growing, includes some important subplots. 

For one, rapid productivity gains have widened China’s competitiveness gap but also reduced manufacturing employment, though it appears to be reversing slightly or stabilising since 2019, with Xi Jinping’s push toward high-technology industries and the Make in China 2025 campaign. But while manufacturing employment has been declining in the advanced economies, it has been rising in the emerging economies (excl China). 

The rise in emerging economies is especially pronounced since the pandemic. It has also been inching upwards in the advanced countries too.

Putting both together, Baldwin makes the case that since 2013, while China has been shedding manufacturing jobs, losing about an eighth of its base, emerging economies have been gaining and advanced economies have plateaued. However, with 134 million factory workers, China has twice more than the advanced countries combined.

Among the gainers in this giant manufacturing jobs reallocation, Vietnam, Nigeria, and Indonesia lead. Nigeria and Pakistan are surprising candidates. India has been a notable laggard. Together, the developing countries added 22 million. 

China’s export-supported factory employment is down 16% since 2013 and its share of the world total has fallen from 37% to 30%… China’s factory employment is now doing what the advanced economies’ did decades earlier: shrinking as productivity rises and as labour-intensive stages migrate to lower-wage locations.

This decline in Chinese manufacturing employment comes alongside a steeper decline in its manufacturing value added as a share of GDP. Its manufacturing GDP share has fallen over 15 years, from a peak near 32% in the mid-2000s to under 25% today. 

However, declines in manufacturing employment and share of value addition do not mean China’s dominance and vice-like grip on manufacturing is on the way down. Far from it. In fact, China’s manufacturing merchandise exports as a percentage of world GDP have been rising, even as they have declined for the rest of the world. 

In fact, the percentage has nearly doubled in the same period. It has also risen for Vietnam and India. The advanced countries have been the clear losers.

So, what explains this apparently contradictory trend of deindustrialisation in employment and value addition, co-existing with the doubling of global manufacturing export share?

The answer appears to lie in productivity. Over two decades, the price of a unit of China’s manufacturing value added fell by nearly half, which explains the whole of the decline in value addition by manufacturing. 

It reflects the most consequential fact about Chinese manufacturing: Chinese goods are getting cheaper at an astounding pace. Fast enough to distort the whole economy, maybe even the whole world economy. That price collapse is also why the Chinese are such fearsome competitors globally. It is why the rest of the world’s industry feels squeezed, and why politicians are responding with plans to hobble Chinese exports.

… there is a China-specific factor that is driving down manufacturing prices even faster. In China they call this competition “involution.” If you think the competition is tough outside of China, you should see it inside the country. China experts talk about the ruinous internal scramble in which too many domestic firms, egged on by rival local governments, drive down prices until nobody makes money.

This comes out clearly if we map the nominal and real price shares of manufacturing value addition. Since about 2014, the manufacturing sector has grown at the same real rate as the economy as a whole. 

The Chinese economy has made much more with a smaller set of resources. 

There is no doubt that the expansion of the Chinese manufacturing sector has been an era-defining phenomenon. The rise can be seen in the red line in the chart below. Real manufacturing output in 2025 is more than six times its 2005 level. That’s amazing. That’s the China shock 1.0 and 2.0 in a nutshell. Over the same twenty years, the whole economy (the grey line) grew only four-and-a-third times. So, manufacturing did outrun the economy. But only up to the mid-2010s.

Finally, he disputes the oft-cited deglobalisation argument. He points out that while the global manufacturing trade ratio fell by 2.6% in the 2008-22 period, it rose slightly by 0.3% if we exclude China. 

The main contributor was China’s GDP growing faster than its manufacturing exports.

China’s manufacturing exports boomed, but its GDP boomed more. China’s share of world manufacturing exports rose from 11.5% in 2008 to 19.3% in 2022. China’s share of world GDP went from 7.2% to 17.9%. Exports relative to China’s own GDP fell from 24.2% to 16.0% as it morphed into a normal mega-economy and became its own best customer.

Another driver of this was the reshoring of supply chains into China. However, this deglobalisation by China did not mean similar trends elsewhere. The rest of the world expanded the global value chains in their manufacturing. 

China’s domestic value-added share fell from 83% in 1995 to 72% in 2004, as the processing-trade boom stuffed Chinese exports with imported parts. It then climbed steadily to 81% by 2022. From the 2004 trough, that is nine percentage points of genuine input localisation… For the world excluding China, the domestic value-added share moved the other way: 68.7% in 2008 to 66.4% in 2022. Everyone else’s manufacturing exports became slightly more dependent on foreign inputs, not less. If you special-case China, there hasn’t been any localisation of manufacturing. The domestic content share has not risen and seems to be continuing to fall.

It would be useful to disaggregate the value addition destinations of the manufacturing exports of countries excluding China. It will not be surprising that a significant share of that value addition is coming from China. 

So, not only is China increasing the localisation of its own manufacturing, but it is also capturing a greater part of the value addition of its trade partners’ exports. 

Baldwin also points to how global export volumes continue to rise unabated, thereby contradicting talk of deglobalisation. He says that world trade never stopped growing, but only stopped outgrowing the world economy, pointing to a deceleration but not a reversal. 

None of these nuances and qualifications take away from the reality of the world economy’s China problem. Its manufacturing dominance and export dumping is among the biggest problem facing the world economy. 

China inverts the conventional wisdom of a large developing country in its development trajectory being a large importer of goods and services. Instead, even as its own market remains largely walled off, it has become a massive exporter of components and goods spanning the full spectrum of sophistication.

China also contradicts the orthodoxy of developing countries vacating lower-skilled industries and moving up the value chain as they develop. Apart from some limited vacation of space, as discussed above, China has uniquely retained its competitive advantage across the value chain of industries. Its competitive advantage spans across sectors, supply chains, and value chains.

This has also meant that even as the world has globalised supply chains, China has ended up increasing its localisation. As mentioned above, this means that China’s manufacturing dominance is boosted by both the localisation of its own manufacturing and by capturing an increasing share of the rest of the world’s supply chains. 

All this has meant that China neither offers its large consumption market nor vacates any significant part of the manufacturing landscape for its trading partners. Further, it wants to use the rest of the world as a dumping ground for its heavily distorted manufacturing industry. Furthermore, it is now showing an increased willingness to weaponise its manufacturing dominance not only for national security and strategic reasons, but to also prevent efforts to diversify supply chains away from the excessive dependence on China.

Weekend reading links

1. Excellent essay by William Dalrymple about the East India Company, drawing parallels with Big Tech today.

By the end of the 18th century, the East India Company had created a vast and sophisticated administration in India and built much of the London Docklands. Its annual spending in Britain — around 8.5 million pounds — equaled about a quarter of the British government’s total annual expenditure. In India it collected taxes, minted coin, administered justice, ran its own courts and diplomatic service, flew its own flag, negotiated treaties and made war on sovereign states, all in pursuit of increased dividends. An international corporation that began trading spices ended up transforming itself into a colonial superpower... For much of the 18th century it was staffed with as few as 35 people, in a building five windows wide. Yet in India the East India Company owned a private army that by 1803 numbered some 200,000 men — roughly twice the size of the standing British Army. Its stock was a pillar of British public finance, and its profitability and solvency were matters of state...

Its lawyers and lobbyists and parliamentarian shareholders slowly and subtly worked to use its immense wealth to influence and subvert legislation in its favor. In 1693 the company was discovered to be using its shares to buy influence with prominent members of Parliament and ministers. The following parliamentary investigation, the world’s first corporate lobbying scandal, found the East India Company guilty of bribery and insider trading. After that, the company became more subtle and instead backed candidates favorable to its policies. It arguably invented corporate lobbying, that alchemy by which the interests of a company somehow magically become the policies of the state. Around one in 20 members of Parliament sat on the East India Company’s board, and more than a fifth of the company’s directors sat in Parliament at some point; about 40 percent of members of Parliament were shareholders.

2. The Netherlands is Ground Zero for electricity grids that are hitting capacity constraints to evacuate the renewable capacity becoming available. 

Since July 1, the grid operator has frozen new connections in Utrecht to avoid power cuts... Fewer housing projects can be developed, while plans to electrify local industry and install faster chargers for electric vehicles are on hold. Utrecht’s predicament could become the norm across the EU unless governments learn from its example and invest heavily in grids to support the bloc’s shift from fossil fuels to cleaner technologies... The EU’s fifth-largest economy has more solar panels and electric charging points per person than any other country in the region... But its failure to invest fast enough in pylons, cables and substations to support the transition means the queue of companies waiting for a power connection increased from 12,000 to 15,000 last year. Only 700 companies received a connection in 2025... Ember research suggests that in countries such as Austria, Poland, Portugal and Romania, there is enough grid capacity for less than 10 per cent of the renewables projects planned by 2030.

3. The FCNR(B) deposit scheme inflows in perspective.

Gross inflows on the current account are at $1.1 trillion a year. On the capital account, those are $1.7 trillion a year. Put together, we are getting inflows and outflows of about $2.8 trillion a year or about $11 billion a day. A comparison against conditions in the 2013-14 currency defence is instructive. At that time, gross inflows were $0.55 trillion and $0.52 trillion on the current and capital accounts, respectively, i.e. $4.3 billion a day. Today’s India is 2.6 times bigger. The FCNR (B) stratagem relies on using public money to defend the rupee. The state subsidises foreign borrowing to attract dollars. This tool requires large-scale borrowing to make a material difference against a gross external flow of $11 billion a day. A commensurately large fiscal cost falls upon the exchequer. The Ministry of Finance will choose how much it is willing to spend in exchange for this round number.

And its total cost could exceed $10 bn for the RBI

Under the Foreign Currency Non-Resident (Bank), or FCNR(B), program, the central bank agreed to shield banks from losses if the rupee weakens, through a favorable currency-swap facility estimated to cost 3 per cent-3.5 per cent a year. It will also need to absorb some of the extra cash pumped into the banking system as banks exchange the dollars they raised for rupees. The two operations could cost as much as ₹1.2 trillion ($12.7 billion) over five years, according to an analysis by Madhavi Arora, economist with Emkay Global Financial Services.

4. The rise and rise of bond yields.

5. India cash transfer schemes facts.

The Sixteenth Finance Commission estimates that unconditional transfers by states increased from ₹73,099 crore in 2018–19 to ₹2.63 trillion in 2023–24 and are at ₹4.14 trillion in 2025–26. Large-group cash transfers alone are projected at ₹.96 trillion, accounting for 47.4 per cent of all unconditional transfers, up from roughly 16 per cent in 2018–19. Social-security pensions have also risen in absolute terms, from ₹44,453 crore to a budgeted ₹1.58 trillion, but their share of unconditional transfers has fallen from 60.4 per cent to 37.9 per cent; the share going to farmers has similarly declined from about 24 per cent to 14.7 per cent. This implies that states are not merely expanding welfare but are shifting their composition towards broad category-based payments that claim an increasing share of fiscal space.

6. Europeans pull out their gold reserves from the US.

The Dutch central bank has shifted more than 78 tonnes of gold from New York to London in a politically sensitive move, citing “increasing geopolitical unrest”. The transfer follows calls from European politicians and taxpayer lobbyists to repatriate gold reserves from the US, warning that an unreliable American government under President Donald Trump may otherwise seize them amid growing transatlantic tensions... The move follows a similar decision by France, which removed all of its gold from the New York Federal Reserve between July 2025 and January 2026. François Villeroy de Galhau, the governor of the French central bank at the time, said then that the move was not politically motivated. Gold last year overtook US government bonds as the world’s largest reserve asset, according to ECB data, but some central banks are becoming increasingly skittish about storing gold in the US.

7. Interesting trends in gold prices. 

For the first two decades of this century, a 1 percentage point move in US real rates typically coincided with a roughly 14 per cent move in gold in the opposite direction. That relationship ended in February 2022, when western governments froze Russia’s foreign exchange reserves. Reserve and asset managers globally were left confronting a simple question: if $630bn held in Treasuries, Bunds, gilts and other bonds could become inaccessible overnight, what constituted money? Their answer was gold. Emerging market central banks and sovereign funds have since increased gold allocations from 5 to 7 per cent of reserves in 2022 to 11 per cent today, still short of the 26 per cent held by developed-market peers. Between March 2022 and October 2023, US five-year real yields rose more than 4 percentage points. Based on historical relationships, gold should have fallen about 55 per cent; instead, it rose 7 per cent. Over the following two years, real yields fell less than 1 point while gold rallied 110 per cent. Gold is now far more responsive to falling real yields and less sensitive to their rise.

8. Early evidence of job losses in India's IT industry.

9. Shyam Saran describes the efforts being made by China to circumvent the US-controlled SWIFT and other cross-border financial flows management systems. 

The mBridge project is a network of the central banks of China, the Hong Kong Monetary Authority, Malaysia, Thailand, the United Arab Emirates, and Saudi Arabia, who have linked their central bank digital currencies (CBDC) on a blockchain ledger system pioneered by China, to allow virtually instant cross-border financial transactions. The pilot phase of mBridge is now over and its commercial launch is imminent... mBridge does away with the correspondent banking system... mBridge enables escape from sanctions... One should consider mBridge as only one component of the Chinese efforts to achieve the internationalisation of its currency, the yuan. There are parallel institutional and procedural tracks.

One is the Cross-Border Interbank Payment System (Cips), which provides both an inter-bank messaging system like SWIFT and a clearing and settlement system. It does not seek to supplant SWIFT, but provides an efficient alternative, taking advantage of China’s role as the world’s largest trading nation and increasingly as a significant source of investment. It makes sense for partner countries to opt for yuan-designated transactions to avoid exchange risk. Increasingly Cips also avoids sanctions risk since, unlike SWIFT, it is not subject to US or Western regulation and control. This is why both Russia and Iran now use Cips almost exclusively for their cross-border financial transactions.

10. Excellent op-ed by Ajanta Krishnamurthy about the premiumisation of everything we consume. 

A biscuit is no longer a biscuit. It is handcrafted, slow-baked, perhaps even inspired by some grandmother somewhere. The packet is brown, the lettering is tasteful and the quantity inside is just enough to make you wonder whether you have accidentally bought stationery. The same thing has happened to the rest of the kitchen and, increasingly, the bathroom. Detergent has become laundry care. Pickle is a small-batch preserve. Tea is a single-estate experience. Ice cream is slow-churned. Salt has become strangely ambitious. It can be Himalayan, smoked, pink, infused, mineral-rich, and important enough to deserve a place at the table. Water has perhaps had the greatest career transformation. For most of our lives, water was the one thing nobody had to explain. You drank it. You asked for more. Now there is water that sounds like a minor European aristocrat, comes in a glass bottle, is alkaline, and costs enough to make you briefly consider dehydration.

Improving ‘decision velocity’ in bureaucracy

Shashi Tharoor has a review of a new book, Decision Velocity: Rational Abdication, the Fear Tax, and the Road to Viksit Bharat, by retired IPS officer OP Singh. Singh describes a phenomenon of not taking decisions, or ‘rational abdication’, which imposes a “fear tax” that is extracted in the form of unconscionable delay. Tharoor writes

Singh’s “Fear Tax” is the price society pays when decision-makers conclude that the safest course is not to decide at all. It is the outcome of a sclerotic system where, for the desk-bound officer, the risk of a decisive signature far outweighs the perceived safety of a prolonged silence. As Singh so eloquently demonstrates, our current accountability frameworks audit the traceable action with inquisitorial rigour, yet remain blissfully indifferent to the absent decision. Hence, it’s safer not to act at all…

We have inherited, and subsequently ossified, a colonial-era machinery designed for control, not for velocity. In such a framework, a file that moves is a file that invites scrutiny; a file that rests in a dark corner of a cabinet is a file that poses no threat to its custodian. The result is a Kafkaesque reality where the paralysis of the pen is too often mistaken for the prudence of the professional. This is not merely an inconvenience; it is a profound failure of the social contract. When the state abdicates its responsibility to act, it does not merely delay a project or a permit; it stifles the latent potential of a citizenry waiting for the basic friction of governance to be removed.

Tharoor offers some suggestions.

We must move from an obsession with compliance to an obsession with outcomes… We must evolve our oversight mechanisms to distinguish between honest mistakes (the byproduct of necessary risk-taking in a complex environment) and true negligence. A system that does not tolerate the possibility of a wrong decision will rarely produce a right one.

Furthermore, we must incentivise administrative courage. This involves creating "safe spaces" for decision-making, where the rationale for action is documented and respected, and where the institutional culture shifts from "How can I avoid this?" to "How can this be made to happen?"

There cannot be any argument about these. The challenge is to operationalise tolerance of honest mistakes and creation of “safe spaces”. 

The operationalisation of these cannot happen by diktat or homilies. Instead, it requires building them into bureaucratic routines and institutional processes. The former involves changing norms within the government on taking risks, accepting failures, and even celebrating risk-takers. The latter involves wiring the institutional processes to accommodate these. 

I can think of three process safeguard reforms.

An important nudge in this direction would be to amend the Prevention of Corruption (PC) Act to clarify that no public servant shall be subjected to any enquiry, inquiry or investigation under this Act, nor prosecuted, solely based on an opinion, advice, recommendation or decision recorded or taken by him in the discharge of his official functions, unless there is independent material to show that the act was actuated by a corrupt motive or by the expectation or acceptance of an undue advantage. It should also be clarified that an error of judgment, or an unintended, unsuccessful, or sub-optimal outcome of such a decision, shall not constitute such material. This would go beyond the procedural safeguard provided by Section 17A of the PC Act (whose scope is currently not yet settled and awaiting a Supreme Court bench verdict) and provide a substantive threshold. 

Another would be to amend the Right to Information Act and insulate the contents of deliberative processes (opinions, advice, recommendations, notings, drafts or inter se deliberations recorded by the public servant) from any investigation or be made inadmissible in prosecution proceedings, unless there is evidence that the act itself was done with a malafide or corrupt motive. The US Freedom of Information Act insulates the deliberative process as a safe space from any scrutiny. This would allow people to express opinions freely without hedging for future recriminations. 

A third would be to amend the guidelines on performance audit by the Comptroller and Auditor General (CAG) of India to explicitly make the auditors accountable for examination of honest mistakes and unintended failures before making their performance audit comments. It should be clarified that audit shall assess a decision based on the information, circumstances and options reasonably available to the decision-maker at the time the decision was taken, and not with the benefit of hindsight or of facts that emerged subsequently. Where the decision was within the competence of the decision-maker, followed the prescribed process and consultations, was supported by reasons recorded at the time, and was one that a reasonable and prudent official could have taken on the information then available, the auditor should treat it as an honest exercise of judgment and let it rest. 

These enactments must be complemented with efforts to change the norms on such decisions. 

In this context, an important perspective is to view decision-making through the lens of public and private benefit. Many of the delayed decisions or undecided issues involve the likelihood of private benefit. Consider the following three norms and their implications:

1. There is a tendency to view public interest and private benefit as mutually exclusive and conflicting. Further, since the ethos of bureaucracy is the protection of public interest, any resultant private benefit is considered a matter of concern. 

2. There is a belief that denial of private benefit equates to protection of public interest. However, in many cases, this is not true. On the contrary, it is the opposite - it ends up with a back-ended public loss.

3. There is also the instinctive association of private benefit with corruption. This naturally stigmatises any act that results in private benefit. 

These norms overlook the reality that private benefit can co-exist with public benefit. 

The aforesaid norms are most deeply entrenched among auditors and investigative agencies. They tend to view public actions resulting in private benefit with extreme suspicion. 

For example, it is common to see decisions on procurements and contract management delayed due to these hesitations. Officers are regularly called upon to propose decisions on changes in technical specifications, qualification norms, condonation of delays and time extensions, waiver of liquidated damages, contract terms during renegotiations, and so on. All of them, even when justified, end up invariably benefiting some or other contractors, often also causing an increase in expenditure. This opens the decision to be scrutinised as having caused both private benefit and public loss. 

These entrenched cultural norms are reinforced by the circumstances. Compromises on public interest and promotion of private benefit through corruption have become pervasive. This complicates matters and allows righteousness, biases, and prejudices to enter the audit, investigative, and prosecutorial processes. 

In the circumstances, the challenge is to create a culture that destigmatises actions that confer private benefit, without letting it become permissive enough to encourage corrupt practices. This is hard! Most often, there are no shortcuts to doing the hard work of creating the conditions. 

Resolving the public debt pile facing the developed world

Public debts have soared among advanced economies, with the US leading the way. Their sustainability is now a big risk and casts strong headwinds on economic growth. 

The repercussions are already being felt in the sovereign bond markets, including in the safest of assets, the impregnable US Treasury markets. 

The response, especially the recent actions of the US Treasury, arguably even encroaching into the Fed’s domain, to prevent bond yields from rising are extraordinary. It first announced a temporary swap of the Bank of Japan’s Treasury holdings for dollar cash to pre-empt any liquidation to generate the dollars required to prop up the Yen, and followed up with an announcement to double the purchases of Treasury bonds. More are likely to follow in the months ahead, especially with a President who has announced his intent to wage war against the bond markets. 

Chris Giles has an excellent description of America’s ongoing debt surge.

The most relevant measure of US federal government debt, that held by the public, has risen from $3.4tn in 2000 to $32.3tn now, or a rise from 33.7 per cent to more than 100 per cent of GDP in just over 25 years. More importantly, the burden of servicing that debt has doubled, from 11 per cent of tax revenues in 2000 to 21.5 per cent in the first 10 months of the current fiscal year… With the US on a path to continue running deficits close to 6 per cent a year, even at full employment, the debt-to-GDP ratio is set to rise every year, increasing the call on tax revenues to service that debt and the pressure on the Fed to lower interest rates… 

Since 2000, when the federal government ran a surplus of 2.3 per cent of national income, primary public spending (excluding net debt interest) has risen from 15.5 to 19.9 per cent of GDP. All of this increase can be accounted for by spending on services for an ageing population — social security, Medicare and veterans’ programmes. On the tax side, revenues have fallen from 20 per cent to 17.2 per cent of GDP over the same period, partly a result of a cyclical peak in revenues at the end of the last millennium and partly the result of tax cuts. First came the Bush tax cuts, which were made permanent on a mostly bipartisan basis during the Obama administration, and then came the 2017 Trump tax cuts… It does not need to eliminate the deficit, but does need to put debt back on a downward path, which almost certainly requires a balanced primary deficit — a metric that excludes net interest costs — something the US has not achieved since 2007 and not on a sustained basis since the 1990s.

The combined debt held by the public and federal agencies in the US now touches $36 trillion. The US interest bill has doubled to more than 3% since 2021, and the fiscal deficit is running at nearly 6% of GDP with little prospect of declining anytime soon. 

America does not stand alone. Since the Second World War, as welfare states took hold, public spending as a share of GDP has risen steeply across the developed countries. Further, since the seventies, public debt as a share of GDP has been on a similar upward trend. 

Welfare spending has been rising, and subsidy policies have ratcheted up since the global financial crisis. 

This increase in public spending and borrowings is a political choice made in response to rising expectations from governments among the electorate. 

For measure, in the US, net positive mentions of government support — covering welfare and protectionism — in Democratic and Republican party manifestos have trended higher since the early 1970s, based on calculations from the Manifesto Project’s database. In particular, these mentions have surged since the GFC. This suggests politicians are increasingly pitching policies that extend state assistance to appeal to voters. In the UK, the National Centre of Social Research reported in 2023 that expectations for government to keep prices under control, reduce income differences and provide industry with the help it needs to grow all reached a record, based on the British Social Attitudes survey. Indeed, curbing support hasn’t been easy for governments either. Britain’s Labour party was forced to reverse over £5bn of planned cuts to welfare spending. French politicians are also struggling to agree on how to cut expenditure, given the inevitable pains on the public.

A similar expectation cycle has been set off in the financial markets on monetary policy under the watch of technocratic central bankers like Ben Bernanke and Mario Draghi. These cycles have come to reinforce each other. 

A May 2024 paper by John Cochrane and Amit Seru, senior fellows at the Hoover Institution, concurs. It argues that the expectation of central bank support whenever conditions deteriorate inflates stock prices, fuels leverage and, in turn, raises risks of a self-reinforcing monetary policy intervention and taxpayer-funded bailouts. Central banks have also been cautious about unwinding their QE holdings too quickly, fearing market convulsions. This has left their balance sheets elevated, creating a structurally higher liquidity base in the financial system that props up valuations and market activity today.

The rising indebtedness has close historic parallels. In an FT interview, Thomas Piketty pointed to the experiences of European countries in the 19th and 20th centuries. 

We have a long history of public debt in France and Britain. Britain had more than 200 per cent of GDP of public debt in the 19th century. In the case of France . . . after each of the world wars, it was between 200 per cent and 300 per cent. The good news is that we’ve always found ways to get rid of it, and in each of the three instances I’m referring to, through different mechanisms, it went down to very little — less than 20 or 30 per cent GDP in a few years. It was never repaid, in effect, as opposed to the British solution in the 19th century, where it took basically one century of budget surplus between 1820 and 1914 to reduce the public debt coming from the Napoleonic War period… 

The British approach of the 19th century corresponded with a very aristocratic political system where basically one tier was in power and they wanted taxpayers to reimburse them. Was it the best way to prepare the country for the 20th century? I’m not completely sure, because in effect there was more money put into interest payments than money invested in education. And Britain lagging behind in education with respect to the US or even with respect to Germany or France in the 20th century is the number one explanation for a British decline. I would not recommend doing the same in the future… The most successful experience with large public debt is probably Germany after [the] second world war, where they had this exceptional tax on private wealth which raises a lot of money. This contributes a lot to the reduction of the public debt without any inflation. Of course they were traumatised by inflation in the 1920s, so they didn’t want inflation anymore. The other way, of course, is through inflation, which is a wealth tax on the poor, typically.

A major reason for the surging debt stock since the turn of the millennium has been external wars. It has been estimated that the total cost and future obligations of the post-9/11 wars are about $8 trillion in 2021 dollars, excluding future interest costs on the debt. To put all this in perspective, “the US defence budget in 2025 was over $900bn, equivalent to 35 per cent of total global defence spending and more than three times the defence budget of China, the next most powerful military actor.”

Despite this, the dollar has continued to hold reasonably steady. The exorbitant privilege has ensured the dollar's status as the world's pre-eminent reserve currency and the Treasury market's role as the world's safest haven asset, thereby allowing the US access to unlimited global capital at a low cost. While there are no competitors to the dollar on the horizon, the Treasury's safe-haven status is facing competition. 

For one, the central banks, which accumulated reserves by buying up 63% of the extra debt issued by G-7 governments in 2008-21, are now unwinding their balance sheets by running down the dollar component of their reserves. They are instead pursuing alternatives like gold, commodities and the more liquid currencies of smaller developed countries like Switzerland. As William White has pointed out, this has created vulnerabilities. 

Indeed, a report by the European Central Bank showed this week that gold had now replaced US Treasuries as the world’s top reserve asset. By the end of last year bullion accounted for 27 per cent of all global central bank reserve assets, up from 20 per cent a year before. Treasuries fell from 25 to 22 per cent over the same time. This leaves a gap that has been substantially filled by hedge funds, mainly American owned but often counted as foreign investors because of their bases in tax havens such as the Cayman Islands. Many own Treasuries as part of highly leveraged “relative value trades”, financed by short-term borrowing that has to be constantly rolled over. 

William White, former chief economist of the Bank for International Settlements, points out that this works well — until it does not. White argues that the purchase of government debt by non-bank institutions such as hedge funds depends in turn on their access to short-term financing such as the repo market. He adds: “Should any disturbance interrupt that access, as in March 2020 [during the Covid-19 pandemic] or April 2025 [when Trump announced swingeing tariffs], an intense deleveraging spiral could easily follow.” Recent shocks from hedge fund margin and collateral calls have made the Treasury market more fragile and a potential source of systemic risk.

And there’s more. 

White also worries about fiscal dominance — a phenomenon in which the central bank cannot raise interest rates to meet its inflation target because of the punishing servicing cost of high, short-term public debt. This in turn undermines price stability. Another possible concern is financial repression, where the government forces banks and other financial institutions to buy its IOUs at below-market interest rates.

And with bond yields rising, the debt service costs are increasing. One estimate puts an additional $34 bn in financing costs by the end of the first quarter of next year for G7 nations. And this will only rise. 

Debt reduction can be achieved, as has happened in the post-war era in the UK, France, the US, etc., and more recently in Greece (more on it later). 

It requires that both the stock and flow of debt must be brought down. The former requires that the rate of economic growth (g) must exceed the rate of growth of debt (or the interest rate, r) for a long period, and the latter requires ensuring that the primary balance is positive. 

One can think of some scenarios under which the debt reduction can materialise. The ideal scenario of growth driven by a productivity surprise (say, AI) may be too optimistic, given the need to sustain real GDP growth (g) in the 3-4% range against a real interest rate (r) at 1-2% and the ageing demographics and falling labour force participation rates. Fiscal consolidation resulting in primary surpluses is another possibility, though the very high mandatory spending plus interest (now at 75% of the budget and heading to 80% by 2036) means that austerity alone can only be a marginal contributor. 

There is no historical precedent of grow out of debt at this debt level without either financial repression or fiscal consolidation alongside. The 1990s US surpluses under Clinton were a one-off combining Bush’s 1990 and Clinton’s 1993 revenue measures, the peace dividend, and the dot-com boom’s capital gains windfall. 

Then there are the scenarios of Fed-led financial repression, with moderate or high inflation. However, the inflationary path requires that the debt is long duration, new issuance reprices only slightly, and the social and political consequences are managed. But the US Treasury’s weighted average maturity is only about six years, which sharply limits this pathway to debt reduction, and there may be no political tolerance for sustained inflation beyond, say, 3.5%. In any case, repression and inflation will be important factors eventually in the years and decades ahead. 

They were important contributors to bringing down the debt-to-GDP ratio after the War, and the recent increase in bond buybacks announced by the US Treasury Secretary Scott Bessant are the early steps in a long period of financial repression. 

This brings us to debt restructuring and haircuts, which are unthinkable for the US given that the US dollar is the reserve currency. There is also the possibility of a consolidation forced by a crisis (à la Greece), though it looks unlikely for now and may lie 10-15 years ahead. 

Finally, the option of muddling through, stabilisation without meaningful reduction, is a very strong likelihood for the foreseeable future. Through a combination of mildly negative r-g through soft repression, containing primary deficits, and occasional tailwinds and reforms, debt-to-GDP can stabilise at 110-130%. Japan has run this combination for over two decades with 220-260% of GDP without any crisis, and Italy has done so at 130-140%. 

For debt reduction, the US and others may find Greece an unlikely example. From a peak of 212.6% in 2021, by the end of 2025, Greece reduced its debt-to-GDP ratio to 146.1%, and it is estimated to decline to 125% by the end of the decade. 

This spectacular record-breaking drop of nearly 67 percentage points within a four-year window has been achieved through a combination of strong GDP growth post-pandemic (4-5% real growth), early repayment of its financial rescue packages, negative real rates from the inflation shock, maintaining a primary budget surplus of above 2%, and concessional EU financing. 

This combination of factors is unlikely for advanced countries like the US. In the circumstances, the best hope is a trend of moderate inflation (say, 3%), repression to keep interest rates down, some reversal of the accumulated tax cuts, and some expenditure reduction, all of which will only stabilise the debt at about 120-130%. This would create the conditions for deeper reforms on both the revenues and expenditure sides after a forced crisis sometime in the later part of the next decade.

A scenario which cannot be dismissed is one where the erratic policies of the Trump administration, combined with rising inflation, a supply shock (of the kind in Iran), and an AI-equity market meltdown, spook the bond markets, resulting in a significant spike in bond yields. This could, in turn, force the US Treasury into biting the bullet on revenue and expenditure-side reforms. 

In the meantime, the bare minimum to calm the markets would be to at least ensure that the debt-to-GDP ratio is stabilised by bringing the fiscal deficit under control. But wars and Trump 2.0 policies work in the opposite direction.