Weekend reading links

1. India R&D facts of the week

According to World Bank data, the share of R&D in GDP in India was 0.64 per cent in 1996, rose to 0.86 per cent by 2008, and has steadily fallen since, reaching 0.65 per cent in 2020. The same database shows a steady rise in China —from 0.56 per cent of GDP in 1996 to 2.56 per cent in 2022... In 2020-21, corporate sector accounted for 36.4 per cent of the total R&D expenditure. If one includes the public sector corporations, the share goes up to 40.8 per cent... Corporations in India do not appear to treat R&D as a major factor in their advancement. A detailed 2024 study on R&D by the 1,000 largest listed companies, conducted by the Office of the Principal Scientific Adviser to the Government, offers revealing insights. According to this report, in 2022–23, R&D spending amounted to over 6 per cent of turnover in defence and pharmaceutical companies, around 3–4 per cent in automobiles and healthcare, marginally over 1 per cent in auto-components and heavy electrical equipment, and well below 1 per cent in the remaining 24 sectors covered in the study.

2. Public funded research laid the foundation for Ozempic and Wegovy.

Today, millions of Americans take Ozempic, Wegovy, Mounjaro or one of the other new blockbuster diabetes and weight-loss drugs. Thank Uncle Sam — and a slow, venomous lizard that can survive on just a few meals a year. In 1980, Dr. Jean-Pierre Raufman, a researcher studying insect and reptile venoms at the National Institutes of Health, discovered that venom from the Gila monster had a pronounced effect on the pancreas, prompting it to release a digestive enzyme. This piqued the interest of Dr. John Eng, an endocrinologist at the Veterans Affairs Medical Center in the Bronx, who worked with Dr. Raufman to isolate and identify a novel compound, exendin-4, in the lizard’s venom. A synthetic version of the compound, which stimulates insulin production and slows stomach emptying, was approved for the treatment of diabetes in 2005. It was the first drug in the now booming class of medications known as GLP-1 receptor agonists, which are being studied for their potential to treat a wide range of conditions, including kidney disease, Alzheimer’s and alcohol use disorder.

The article has eight other examples of public funded innovations that have transformed our lives today. 

3. Are Chinese brands now going global?

From Stockholm to Sydney, the electric car gliding silently by is increasingly likely to be Chinese. Mixue, a purveyor of ice-cream and cold drinks, has dethroned McDonald’s as the world’s largest fast-food chain by number of outlets. It is expanding in South America, as is Meituan, a Beijing-based delivery app. Chagee, a chain of tea shops, is on track to have at least 1,300 stores outside China by the end of 2027, mainly in South-East Asia; a few years ago it had barely any. And Pop Mart, a Chinese toymaker, has created a buzz worthy of Disney around its strange grinning (or are they grimacing?) nine-toothed dolls, called Labubus. Fans include Rihanna, a pop star, and Sir David Beckham, a retired footballer.

4. Ruchir Sharma points to three ways in which the US stock market run may be stopped.

There are however three ways this buoyancy might break: the AI narrative shifts again, given that companies are investing hundreds of billions of dollars in AI infrastructure, without quite knowing who will profit, or when. Economists prove uncharacteristically more right than the market about the threat of lower growth and higher inflation from tariffs. Or investors come to realise that the apparent strength of US consumers and corporations is a mirage — the flip side of the massive and rising US government deficit.

5. Martin Sandbu points to the return of financial repression, or the process of steering financial flows to where the government (and not the market) wants them to flow.

Rumours of a “Mar-a-Lago accord”, which would manage the dollar’s value down while forcing global investors to discount and lock in lending to Washington, has produced shocked disbelief by other countries. But it is not just Mar-a-Lago: several policy proposals have surfaced recently that can fairly be grouped together as measures of financial nationalism. These include a tax on remittances, levies on foreign investment stakes by nations with policies Washington disapproves of, and the promotion of dollar-denominated stablecoins and looser bank leverage regulations... China... has retained a non-convertible currency and manages its exchange rate. It uses a network of state-controlled or state-influenced banks, corporations and subnational governments to steer the flow of credit to outlets indicated by various economic development doctrines favoured by Beijing over the years... 

The influential reports of former Italian prime ministers Enrico Letta and Mario Draghi have emphasised that the EU sends several hundred billion euros abroad every year when there are huge domestic funding gaps. This invites policymakers to adopt measures to redirect financial flows. So does the agenda to unify national financial markets. The aim of making the euro a more attractive reserve and investment currency has also been invigorated by Trump’s seeming disdain of the dollar’s role. A big EU-level borrowing programme suddenly looks at least conceivable, and an official digital euro is on the way. In parallel, the UK is trying to coax pension funds to put more savings in the hands of British businesses.

6. Tamal Bandopadhyay charts the impressive 70 years of State Bank of India.

On June 30, 1955, the last day of the Imperial Bank, it had ₹210.94 crore of deposits and ₹116.24 crore credit. The size of the Indian economy at the time was ₹10,977 crore. By March 2025, India’s GDP has risen 3,000 times – ₹330.68 trillion. During this period, SBI’s deposit portfolio has grown 25,000 times, to ₹53.82 trillion, and advances, 35,800 times, to ₹41.63 trillion. In the past 70 years, the bank’s income has risen from ₹8.50 crore to ₹5.24 trillion, and profit, from ₹1.36 crore to ₹70,901 crore. In 1955, it had paid ₹90 lakh as dividend; for FY25, the figure is ₹14,190 crore. The number of employees, too, has risen – from 14,388 to 236,226; profit per employee has grown from ₹90,000 to ₹29,91,000, and branches from 469 to 22,397... 

The SBI has the largest mutual fund under its belt; boasts the second-largest credit card portfolio; and its life insurance arm is among the largest in the private sector. Two of its 18 subsidiaries are listed. On an investment of ₹6,200 crore, the current valuation of the subsidiaries is at least ₹3.5 trillion. Its share in bank deposits is 22.5 per cent and credit, 19.5 per cent. In different business segments, such as retail loans, home loans, et al, its market share varies between 20 and 30 per cent. And, its share in the Pradhan Mantri Jan Dhan Yojana, the world’s largest financial inclusion scheme, is around 30 per cent... with a ₹15.06 trillion retail book, it has at least 25 per cent market share. Ditto for home loans. While it entered the mortgage business late, with a ₹8.3 trillion mortgage book, it is now breathing down the neck of HDFC Bank Ltd, which holds the portfolio of HDFC Ltd following the merger of the home lender with it.

7. Amidst all its industrial prowess, China has lagged behind in high quality industrial products which require precision and engineering excellence, like ball bearings and carbon fibre. 

High-end ball bearings are crucial for reducing friction in everything from high-speed trains and tunnel boring machines to electric vehicles, humanoid robots and drones... for machinery like offshore wind turbines — which are now being built close to 200-metres tall and need to last for around 25 years — bearing manufacturers face “incredible reliability requirements” as their products must withstand “huge” amounts of weight and pressure. While China is by far the biggest single ball bearing market in the world, the $53bn global bearing industry is dominated by Sweden’s SKF along with Germany’s Schaeffler, US group Timken and Japanese companies NSK, NTN and JTEKT... carbon fibre composite cascades, which are used in an engine’s casing to help aircraft land safely. Japanese group Nikkiso has a market share of 90 per cent.

8. Ravaged by the uncertainties induced by Trump policies, the dollar has had its worst first half year since 1973!

9. Another Trump legacy, NSF grants have declined by 51% on a ten-year average, falling below a billion dollars in the first half of the year.

10. Adam Tooze links to this important graphic, about the falling employment in America's biggest companies over the years.

11. Luis Garciano has a great tweet thread (HT: Marginal Revolution) that points to possible limitations of AI in creating very high incremental growth.

The more successful a technology becomes, the less it matters economically. Revolutionary technologies shrink their own importance precisely through their success... Consider history's greatest productivity miracle: artificial light. In 1800, one hour of reading light cost more than a day's wages. By the 1990s, we produced the same light using 1/3,000th the energy. The price fell 40,000x. Modern homes flood with light that would seem miraculous to anyone from 1800. We leave lights burning carelessly, illuminate entire cities all night. Yet lighting is now a trivial fraction of the economy. Total victory made it economically irrelevant. When productivity crushes prices, quantity must rise proportionally to maintain economic weight. But we don't use 40,000x more light than in 1800. Maybe 100x. Human demand has limits. griculture tells the same story. In 1900: 38% of workers, 15% of GDP. Today: 1% of workers, under 1% of GDP. We produce far more food with 98% fewer workers. But we don't eat proportionally more just because food is cheap.

This reveals AI's first constraint: demand inelasticity. When AI makes something essentially free, we don't suddenly want infinite amounts. There's only so much text to generate, images to create, routine tasks worth automating. Second constraint: Baumol's Cost Disease. As AI makes some tasks hyperproductive, wages rise everywhere. But nursing, teaching, therapy, plumbing can't be automated. These sectors must match rising wages without productivity gains. They grow expensive and dominate the economy. Third: O-Ring (named after Challenger). Modern services depend on weakest human link. A restaurant with AI-optimized everything fails if the waiter is terrible. An AI-designed building collapses if contractors mess up. Humans remain the bottleneck.

This explains why technologists and economists can't agree. Epoch sees engineering problems to solve with better AI. Economists see structural forces. You can't engineer away the limits of human demand or the need for human judgment in critical roles. The question isn't "Can AI substitute for humans?" It's "What happens when it does?" History's answer: Automated tasks become economically trivial while the economy reorganizes around what remains human. Growth is constrained by what's hard to improve, not what we do well.

Like electric light, AI will generate massive consumer surplus - the gap between what we'd pay and what we actually pay. But consumer surplus doesn't show up in GDP. The lighting revolution transformed civilization yet its economic footprint nearly vanished. Steam, electricity, computers delivered enormous benefits while their economic importance shrank through success. AI will transform society profoundly. But 20% GDP growth? History says no.

Garciano has a full post here.

12. The impact of the Big Beautiful Bill in the US.

It imposes steep cuts on Medicaid for America’s poorest to partly fund tax cuts for its wealthiest. Between 11mn and 16mn would lose health insurance. Millions more would lose food assistance. The bottom 10th of Americans would sacrifice $1,600 a year while the top 10th would gain $12,000... Depending on the estimate, Trump will be adding between $3tn and $4tn to the US national debt over the next decade... His budget robs tomorrow to help the rich today.

On the definitive event of the week, the passage of the Big Beautiful Bill (BBB) in the US. Its impact.

 

13. Luring manufacturing back to the US is facing a labour market problem.

The pool of blue-collar workers who are able and willing to perform tasks on a factory floor in the United States is shrinking... About 400,000 manufacturing jobs are currently unfilled, according to the Bureau of Labor Statistics... Difficulty attracting and retaining a quality work force has been consistently cited as a “top primary challenge” by American manufacturers since 2017... Many Americans aren’t interested in factory jobs because they often do not pay enough to lure workers away from service jobs that may have more flexible schedules or more comfortable working environments... Attracting motivated young people to manufacturing careers is also a challenge when high school guidance counselors are still judged by how many students go on to college. College graduates, on the other hand, often do not have the right skills to be successful on a factory floor. The country is flooded with college graduates who can’t find jobs that match their education, Mr. Hetrick said, and there are not enough skilled blue-collar workers to fill the positions that currently exist, let alone the jobs that will be created if more factories are built in the United States... the number of young people going to vocational schools and community colleges... is dropping, not growing.

14. More on China's weaponisation of its manufacturing prowess.

In a setback to Apple’s India expansion plans, Foxconn Technology Group has been sending hundreds of its Chinese engineers and technicians back home from its iPhone factories in India, it is learnt... According to sources in the electronics industry, the Foxconn move may have been prompted by the Chinese government’s focus on strengthening its supply chain. They also point out that some Chinese equipment makers, which had identified land for smartphone plants in the country, have shelved their plans. There’s signalling from the Chinese side that technology for making machines for new products should remain within their country. Also, there are reports of the Chinese Customs indefinitely holding key machines, which are required to be retrofitted on the assembly lines to make iPhone 17 in India.

15. An important sub-plot in the US-Vietnam trade deal which imposes 20% tariffs on Vietnamese imports and double that on those transshipped from there. 

Tran Quang, an executive at a home fragrance company that exports nearly all of its products to the United States said that he supported the steeper duty on transshipment because it could help Vietnamese businesses facing unfair competition from Chinese companies that have invested in Vietnam to escape tariffs. “There are a lot of small Chinese guys who come to Vietnam just to relabel their products before exporting to the U.S.,” he said. Trade and investment from Chinese companies have helped bolster economic growth in Vietnam and the region, but Southeast Asia is struggling to beat back the torrent of goods from China that are putting domestic companies out of business...

The lack of information so far released about the Vietnam deal makes it impossible to fully gauge its impact, experts said. Transshipment could refer to products that originate in China. It could also include things that are made in Vietnam but have a certain percentage of Chinese parts. But if the limits on Chinese components end up being strict, American companies could move their production out of Vietnam, said Matt Priest, chief executive of the Footwear Distributors and Retailers of America, a trade group. “If it’s too onerous or difficult to comply, companies won’t use the opportunity to grow sourcing in Vietnam,” he said. “They may even head back to China if it’s price competitive.” ... The restrictions on the amount of Chinese content in exported products also place a burden on local customs officials who have never been asked to scrutinize exports so closely, raising questions about how effective they will be. Some countries have even discussed setting up entirely different supply chains for the United States.

16. Finally, declining reading levels across countries.

Research published in June by the UK’s National Literacy Trust (NLT) shows that children’s reading enjoyment has sunk to its “lowest point in two decades”. While 62.7 per cent of children aged 5 to eight said they enjoy reading, only 32.7 per cent of children between eight and 18 said they gained “very much” or “quite a lot” of pleasure from it. This is 18.7 percentage points lower than 20 years ago and 1.9 percentage points down from 2024. The decline in the UK is emblematic of a global trend seen across the western world. In the US, according to the National Assessment of Educational Progress, 14 per cent of children reported reading for fun almost every day in 2023, down 3 percentage points from 2020 and 13 from 2012.

Some thoughts on VC and PE funds

Private equity is a type of investment strategy, consisting, among other things, of venture capital and leveraged buyouts, which are considered among the most impactful innovations of financial intermediation. Much of this owes to the spectacular success of firms in the information technology sector. 

Venture capital (VC) works on the principle that there are promising ideas and dynamic entrepreneurs out there who are short of capital. If they can be identified, funded and provided light-touch portfolio support (mainly in the form of forging connections), a few among them will hit the bulls-eye and generate windfall returns that more than make up for the failure of the majority.

The central assumption is that of identification. This, in turn, has two parts. One is the belief that venture capitalists have acquired some form of prescience to spot great ideas in their nascent stages, well before their commercial potential becomes evident. Second, they are also able to identify the great entrepreneurs who are behind those promising ideas.

I’m not sure about the former at all in any credible enough manner that can justify investing tens, even hundreds, of millions of dollars in them. The strategy of making ten such bets in the hope that one or two will hit big appears closer to gambling than a strategy grounded on some sophisticated skills. If confined to a promising technology sector in its emerging phase, then the context itself dictates that some firms must succeed, and if you have large pots of money, you are more likely than not to hit some bulls-eye. This raises questions about the value proposition of the venture capitalist, certainly enough to question their outsized rewards.

The second part about the identification of entrepreneurs raises even more troubling questions. It’s hard to believe even the shrewdest brains can spot great entrepreneurs with a few interactions, with enough confidence to be able to make the kind of large financial bets made by venture capitalists. Except, if they are friends or friends of friends, or part of a connected closed network.

This raises concerns of cronyism and exclusion bias. Wouldn’t there be perverse incentives, especially given that venture capitalists are investing others’ money, and also the moral hazard afoot from the fact that most or many of these bets will fail? What about the inefficiency arising from the exclusion of those several others outside the network?

Given the aforesaid, there are some fundamental questions that one could be asking. How can we say that burning several hundred million dollars to generate one or two unicorns or decacorns is an efficient use of capital? What if, instead, investors should be more discriminating and do rigorous due diligence before investing? What if there is a model whereby the high-risk assuming angel and seed stage investors, including governments, are compensated by the later-stage investors who have a less risky pool to choose from?

It’s not that such questions are not asked elsewhere. In fact, industrial policy is subjected to this scrutiny continuously and has been declared an inefficient and wasteful pursuit by the same set of experts. For example, in the context of the Chinese government’s massive Made in China 2025 campaign to boost strategic industries and achieve technological self-sufficiency, which involved Big Funds picking sectors and firms and pouring hundreds of billions of dollars, experts have been quick to castigate it for its colossal waste. 

But they conveniently overlook the same portfolio aspect of these investments. Who can deny that these investments have resulted in a portfolio which is the Chinese economy that utterly dominates the world across several sectors and technologies? Just in terms of the incremental output, jobs created, surpluses from exports, not to mention the geopolitical power conferred, these investments appear to have generated returns in multiples. In narrower terms of sectors – electric vehicles, EV batteries, critical minerals refining and processing, solar panels, wind turbines, electronics components and products, etc. – the success of those sectoral funds is spectacular. 

Admittedly, in all these cases, the successes have come at a very high cost in terms of the amounts spent. But the logical conclusion from this line of reasoning is that wastage and losses are fine as long as the portfolio generates a high net return. 

If experts can question the collateral wastage associated with the emergence of this portfolio, why are they shying away from scrutinising the VC industry’s capital deployment efficiency on similar lines?

Leveraged buyout (LBO) is different in an important way. LBO funds identify industries and firms that have promise but are now operating far below their potential, either due to poor management or deficient enterprise or some other factor that can be worked upon. If these firms can be bought out and their operational efficiencies improved or business models modified through very active portfolio management, most likely by replacing the entire senior management, then there might be large efficiencies to be realised. LBO funds use significant leverage to supplement investor equity in purchasing firms, and place the debt on the balance sheets of the firms being bought. 

The critical assumption here is that of very active portfolio management. This would include the PE LBO fund changing the management, and in general, necessarily getting into the nuts and bolts of the firm’s business, from the high-level business model to the granular unit economics and small operational details.

There’s no quibbling about the value of this model. In simple terms, it’s about identifying firms that are not managed well (and there are several out there), buying them, and addressing their inefficiencies to unlock the hidden value. Who could dispute this proposition?

Two things follow from this model. One, the PE LBO firm must have the internal domain expertise to be able to do this effectively. There are hard limits to the use of outsourced expertise. But acquiring in-house domain expertise of the quality required to do such portfolio management effectively across several sectors is very difficult. Two, since it demands proficiency and intense engagement, there are binding bandwidth constraints on how many firms, even a large PE fund with several teams can manage.

Taken together, there ought to be a self-limiting (in terms of size) nature inherent to the PE business model. This also means there’s only so much that the fund can generate as returns and pay the General Partners (GPs).

In this backdrop, it’s natural that problems start when PE LBO funds try to scale beyond a certain level in the quest to amplify and expedite returns and payouts. The incentive distortions and inefficiencies surface at multiple levels. Each team is now stretched over far more projects than they can effectively manage, leading them to follow a light-touch portfolio management. Further, as the fund size increases, it becomes increasingly difficult to identify good investment opportunities.

Leverage is attractive, especially when rates are low, to make investors’ equity go further, besides also amplifies the GPs’ returns. The period of the PE industry’s growth coincided with that of ultra-low interest rates in advanced economies. Now that rates are normalising, the PE/VC industry faces serious vulnerabilities. 

The use of leverage also expands the envelope of sectors that become attractive for LBO firms. In fact, LBO firms come to believe that they have a model that can achieve high returns even with low-risk assets. So, they buy out low and stable return assets like those in infrastructure or affordable/public housing, load them up with debt, strip assets, and pay out large dividends and pass the parcel along. 

This creates problems and externalities that are borne by society and taxpayers. The British water and sewage sector, specifically Thames Water, is a classic example. The same logic makes similarly boring, low-return and mass-market assets like kindergartens, salons, gyms, laundromats, vape shops, and so on attractive to LBOs, but with large negative externality risks. Is this practice of amplifying risks by using leverage to increase returns on low and stable return mass market assets desirable?

Finally, the incentive to indulge in financial engineering – excess leveraging, skimping on investments, sale and lease back, raiding pension chests, etc. – and strip assets has become pervasive. The squeeze in exit options has led to PE LBO funds indulge in practices like selling to another fund managed by it at a higher valuation to reset the clock, continuation funds, strip sales of part of a fund’s assets, net asset value borrowing, defer interest payments and add them to debt, transfer the best companies across funds, and so on to raise money to pay LPs and kick the can down the road.

This article is about how PE funds have come to see insurance premiums as an attractive source of credit to finance their activities and have therefore created a financial model where they encourage the securitisation of insurance premiums and then buy those securities. The model gets strained once the insurance companies face a liquidity crunch or when the LBO fund is unable to exit its investments.

All this raises concerns about the negative externalities inflicted by LBO funds when the cost of capital becomes normalised. See this, this, this, this, and this. It is especially important since LBO funds now attract investments from pension funds, insurers, sovereign wealth funds, and public endowment funds, thereby raising questions on how private (and therefore lightly regulated) these funds actually are. 

Subsidies and international trade

The primary international trade challenge facing countries in manufacturing is that they must compete with China. To achieve this, they must bridge a significant competitiveness gap. Econ 101 points to increasing productivity (through the likes of improving worker skills and using the latest technologies), investing in infrastructure, lowering the cost of capital, and easing regulations. 

This overlooks an important aspect. A critical requirement to compete with China is to maximise economies of scale. In most sectors, this, in turn, cannot be achieved without being export-focused. This, in turn, draws attention to export promotion policies, especially important to bridge the competitiveness gap arising in particular from China’s massive economy-wide indirect subsidies. 

This is especially important with component manufacturing, the next stage of value addition in India’s manufacturing strategy. The low domestic market volumes and the need to maximise economies of scale mean that Indian manufacturers must aim to Make in India for the World. 

India does not have anywhere near the fiscal resources to match China in providing economy-wide subsidies. It must rely on direct and targeted export subsidies. But, despite the near complete paralysis of WTO and egregious violation of its provisions by the major economies (read US and China), India has been surprisingly reluctant to support industrial policies that support the promotion of exports.

The WTO’s Subsidies and Countervailing Measures (SCM) Agreement categorises two kinds of “prohibited” subsidies - those “contingent on export performance” (Article 3(1)(a)), and those “contingent on the use of domestic over imported goods” (Article 3(1)(b)). It allows for subsidies that are specific to enterprises, industries, and regions. When the SCM and other WTO Agreements were being negotiated in the nineties, it was thought that only the subsidies contingent on export performance would be trade-distorting in any significant manner. It was also thought that the subsidies specific to enterprises, industries, and regions (or the economy as a whole) could not be sustained at the scale required to distort trade in particular products, much less global trade in general. 

China has proved otherwise. This necessitates a wholesale revisit of the WTO itself. 

In this context, it’s useful to place the issue of trade-related (or export promotion) subsidies in perspective.

For a start, as Lorenzo Rotunno and Michele Ruta show, there has been a sharp rise in the use of domestic subsidies in industrial policy, especially aimed at manufacturing, and by developed and emerging economies. They categorize subsidies into four groups - production subsidies, direct transfers (including state aid and grants), policies resulting in a loss of government revenues (tax breaks), and policies wherein the government assumes risk related to the beneficiaries’ actions (loans). 

They show that trade finance and export subsidies and incentives are the two main types of export promotion policies. While both have declined over time, the latter now dominates and is still employed by several countries. 

They analyse the impact of domestic subsidies on international trade flows. 

Exports of subsidized products from G20 emerging markets increase 8 percent more than exports of other products, with no evidence of selection. The gravity estimates confirm that subsidies promote international relative to domestic trade. These spillover effects are concentrated in some industries, such as electrical machinery, and are stronger when subsidies are given through tax breaks than other policy instruments. 

Reka Juhasz and co-authors have a new paper that uses supervised machine learning tools on data from policy announcements in the Global Trade Alert (GTA) database to analyse policy language (instead of policy instruments) to categorise industrial policies (IPs) across the world over the 2010-22 period. They define industrial policy as deliberate government action aimed at altering the composition of the domestic economy to achieve a public goal. Their findings:

The new data on IP suggest that i) IP is on the rise; ii) modern IP tends to use subsidies and export promotion measures as opposed to tariffs; iii) rich countries heavily dominate IP use; iv) IP tends to target sectors with an established comparative advantage, particularly in high-income countries.

The graphic below points to the eight most commonly used industrial policy instruments, which highlights that export-related measures (mainly trade financing) are the second largest category of IP interventions. In fact, export-related measures run a close second to non-export subsidies among IP interventions. 

This figure reveals that most export-related IP measures are deployed via trade-financing, and, to a lesser extent, financial assistance in foreign markets… export-related IP measures tend to operate by providing financing as opposed to directly incentivising exporting… much industrial policy seems designed to facilitate participation in export markets... Finally, this finding highlights that modern industrial policy requires fiscal resources and high administrative capacity. Specifically, states need sufficient fiscal revenue to subsidize firms and promote exports, as well as the administrative capacity to identify which firms to support.

Michael Pettis and Erica Hogan point out that focusing on direct subsidies conceals the true extent of trade-distorting subsidies. Instead, they show that indirect subsidies have become the main instruments of export promotion. The table below captures commonly used indirect trade subsidies.

For example, surplus economies typically have undervalued currencies as part of their trade strategies. Having an undervalued currency subsidizes manufacturing at the expense of households because households are all net importers—as they do not produce for the purpose of exporting—while net exporters are mostly manufacturers. An undervalued currency makes the manufacturing sector in that country more competitive, while reducing households’ consumption capacity… repression of interest rates below the neutral real interest rate has also been a powerful cause of financial transfers from household savers to manufacturers, as seen in Japan in the 1980s and China in the 2000. Further, overspending on transportation and infrastructure serves as an especially significant transfer from households to manufacturers in China today. Other transfers include centrally directed systems of credit, low penalties for environmental degradation, repressive labor laws, and restrictions on worker mobility.

They argue that such indirect subsidies constitute a transfer of wealth from households to manufacturers. They write 

Economies that heavily subsidize manufacturers at the expense of households will typically have: larger manufacturing shares of GDP than their trade partners, since manufacturers must migrate to jurisdictions where workers are paid the lowest relative to their productivity to remain globally competitive in a hyper-globalized world; lower consumption shares of GDP than their trade partners, reflecting the cost of the subsidies on consumers; and large, persistent trade surpluses, as the repressed household income used to pay for manufacturing subsidies makes it impossible for domestic household consumption to balance trade. When all three conditions hold, it is almost certain that repressed household demand is subsidizing manufacturing. And, indeed, we see this reduction of household consumption and increase in trade reflected in economic data for surplus countries (and the opposite trend for deficit countries). 

Yet another instrument used by China to distort trade is its State Owned Enterprises (SOEs), which systematically cut back on imports during trade wars. Chinese SOEs make up a fifth of Chinese imports from the US, providing the country with a valuable trade policy instrument that others don’t have. 

Felipe Benguria and Felipe Saffie examined the period from early 2018 when the first Trump administration initiated a trade war with China, and found that US exports fell relatively more during the trade war in products with a high Chinese import share by SOEs. It found that while tariffs account for an 8% reduction in trade, SOEs account for another 4% decline. They also find that the period of decline in US exports to China in sectors with a higher SOE share also saw an increase in Chinese imports from the rest of the world, pointing to further evidence in favour of the SOE effect. 

They write also about how China may have used its SOEs as a retaliatory weapon against the US tariffs. 

We have also found evidence that the SOE effect was stronger among industries located in Republican–voting US counties, suggesting a political motivation just like the literature has found for tariffs. Our work is the first to provide evidence on the use of state–owned enterprises as tools of trade policy, in any context… We have shown that while US exports facing Chi- nese tariffs were gradually rerouted toward other markets, this was not the case for exports facing the reduced demand by Chinese SOEs.

A World Bank report finds that subsidies have larger trade-distorting effects than even tariffs.

Subsidies create trade-distorting effects for both agriculture and manufacturing exports... Subsidies can be more distortive to trade flow than existing tariffs barriers. The distortionary effect of subsidies on trade, expressed in ad valorem equivalents, is estimated at 15 percent for agriculture and 8 percent for manufacturing… Trade-distorting subsidies can displace trade and production in other trading partners, with important repercussions for developing countries… A disproportionally large number of programs are implemented by major trading countries that have the economic heft to distort global markets for goods and services.

It also finds that the prevailing global trade rules are ill-equipped to deal with subsidies. As indirect subsidies have become the main distorting factors in international trade, and also given its own current dysfunctional state, it’s time to revisit the WTO’s provisions. 

In any case, since 2019, after the US refused to allow appointment of members to the Dispute Settlement Body that hears appeals from members, the WTO has become a toothless organization. This has also encouraged a revival of bilateral and multilateral associations, and countries have come together to also settle their disputes through negotiations. This trend will only be accentuated as the US under President Trump becomes more and more unilateral in its trade engagements. In fact, the reciprocal tariffs announced by the US signals a clear regime shift back to the pre-WTO era. 

Weekend reading links

1. The Swiss Central Bank has lowered interest rates to zero in its sixth consecutive rate cut, after consumer prices fell 0.1% in May from a year earlier. Switzerland has had negative rates from 2015-22, with the lowest being minus 0.75%.

2. Hybrids are making a comeback.

In the first three months of this year, hybrids — including cars that can and cannot be plugged in — made up about 14 percent of all light vehicles sold in the United States, according to the Department of Energy. That was around twice the market share of fully electric vehicles in that period. Republican legislation working its way through Congress could further lift sales of hybrids. In May, the House passed a policy bill backed by President Trump that would eliminate a $7,500 tax credit available to people who bought or leased electric vehicles. That legislation would also impose an annual tax of $250 on electric cars and $100 on hybrids to finance road projects. The Senate version of the bill introduced this week would do away with the tax credit, too, but does not include the annual tax.

A few large automakers dominate the sale of hybrids. Nearly half the cars and trucks that Toyota and its luxury brand, Lexus, sold in the first five months of the year were hybrids — and sales of those vehicles were up about 40 percent from a year earlier. Ford Motor’s hybrid sales rose 31 percent in the same period. Honda is on track this year for its highest hybrid sales ever, and the hybrid versions of its Accord sedan and CR-V sport utility vehicle now outsell the gasoline-only models. Hybrids are typically powered by a small gasoline engine that is paired with an electric motor driven by a battery that is much smaller and, thus, less expensive than the batteries in fully electric vehicles. These batteries are charged primarily by regenerative brakes and gasoline engines. Plug-in hybrids, which account for a small share of hybrids, have bigger batteries than regular hybrids and can also be charged from power outlets at home or at charging stations. Some plug-ins can go around 50 miles on battery power alone before the gas engine kicks in.

3. The changing nature of business lines and revenues of Reliance Industries.

Rahul Malhotra, director at Bernstein, calculates that Reliance now generates more than half its annual earnings before interest, taxes, depreciation and amortisation from consumer-facing businesses, compared with less than 10 per cent a decade ago.

4. Smartphone assembly has been the PLI's standout success.

5. Trends with India's FDI.

Manufacturing's share of FDI has continued to decline.

6. Why Chinese consumers took to EVs?

XPeng's Mona Max, which has just gone on sale in China for around $20,000. For this price you get self-driving capability, voice activation, lie-flat beds, film and music streaming. Young Chinese graduates, we're told, see all these as standard features for a first car purchase... a huge amount of government spending goes towards making EVs financially attractive, according to the CSIS study. Members of the public receive subsidies for trading in their non-electric car for an EV as well as tax exemptions and subsidised rates at public charging stations. These perks drove Mr Lu to go electric two years ago. He used to pay 200 yuan ($27.84; £20.72) to fill up his car for 400km (248 miles) of driving. It now costs him a quarter of that. People in China also normally pay thousands for their vehicle registration plate - sometimes more than the cost of the car itself - as part of government efforts to limit congestion and pollution. Mr Lu now gets his green one for free... Another proud EV owner in Shanghai... says that rather than charge her vehicle at a station, she changes her car's battery at one of the city's many automated swapping stations provided by EV maker Nio. In under three minutes, machines replace her flat battery with a fully charged one. It's state of the art technology for less than the price of a tank of fuel.

7. Some statistics on government spending on public sector units.

The government’s total receipts from disinvestment and dividend from PSUs over this period of 10 years fell from 0.45 per cent to 0.25 per cent of GDP... In contrast, the government has been increasing its capital allocations for PSUs through equity and loans in the last 10 years, from about 0.54 per cent in 2014-15 to about 1.66 per cent of GDP in 2024-25... In 2014-15, total equity and loans to PSUs were estimated at ₹67,512 crore, accounting for just about 34 per cent of the Modi government’s total capex of ₹1.96 trillion. By the end of 2024-25, that share rose to 54 per cent, as PSU equity and loans were estimated at ₹5.48 trillion, out of a total capex of ₹10.2 trillion.

8. State capability, airline industry graphic of the day.

9. Brazilian Supreme Court rules that social media platforms can be held legally responsible for users' posts, forcing them to proactively demove material like hate speech, incitement to violence, etc., even without a prior judicial takedown order. This follows rising concerns in Brazil about harmful digital content, especially on children and youth. 

10. The balance sheet of AI spending and benefits is not looking good.

On the cost side, the effects of AI mania are all too apparent. The four tech companies leading the charge — Alphabet, Amazon, Meta and Microsoft — increased their capital spending by nearly two-thirds, or $95bn, in 2024. As this year got under way, they were planning to boost capex by another $75bn... Bank of America Securities predicts that for the tech industry as a whole, spending on data centres will jump from $333bn last year to about $1tn in 2030. By the end of the period, 83 per cent of the money will go into AI-related investments. 

On the revenue side of the equation, meanwhile, some of the AI leaders are starting to notch up big percentage increases in business — but the extra revenue is counted in the tens of billions rather than the hundreds. Early this year, Microsoft said its annualised revenue rate from AI had climbed 175 per cent to reach $13bn. That is still only about 5 per cent of the total revenue it is expected to produce this year. OpenAI’s revenue run-rate from subscriptions, its main source of income, just topped $10bn, doubling from the end of last year. The rates of increase are notable, but the absolute figures still pale in comparison to the capex.

The distractions of development and public policy

Interestingly, the discourse of development is often marred by distractions. It may not be entirely incorrect to say that a very significant part of the development discourse is avoidable and distracting noise. I blogged here about the distractions caused by the predominance of programs and schemes. 

Bloomberg has a good op-ed that draws attention to the near-universal problem of deficient student learning outcomes and argues that the debate on school choice in the US may be a big distraction.

Cultivating more school options is not the end-all-be-all. As of 2019, 25 states had voucher programs of some type in place, but only 2% of K-12 students are in private school with public vouchers. Only 7% of public school students are in charter schools. The reality is that most kids end up staying in the public schools they are zoned for a variety of reasons: some private schools don’t take vouchers and some charters are oversubscribed or on the other side of town. School choice is particularly ineffective in rural environments... Worse, school choice can become an excuse for policymakers to skirt hard and immediately needed conversations about an ineffective public-school curriculum, classrooms that have morphed into screen zombies, or unaccountable teacher and student performance...

Indeed, the most meteoric change in student achievement this last decade wasn’t from vouchers. It was from a statewide investment in the basics. Since 2013, Mississippi has gone from one of the worst elementary school literacy rates in the country to above average from investing in third grade reading. That included better training for teachers, using a phonics-based curriculum and hiring reading coaches. These investments have been paired with steep accountability: if kids are not literate, they repeat third grade. Instead of falling behind, those kids were further ahead academically by 6th grade for having gotten the basics right. We need more of this, shoring up the foundation.

This echoes across development sectors and public policy areas, and countries, developed and developing. 

The description of school education in the US could be applied to school education systems in India and elsewhere. Smart classrooms and digital technologies, Edtech in general, may have become big distractions from doing the basics right. 

On the same lines, the neglect of public health and primary care is sought to be made up for by focusing on universal health insurance. Consultants and dashboards are the response to basic governance failures in monitoring and supervision of programs, schemes, and policies. The unwillingness (and inability) of basic energy audits and their enforcement is sought to be made up for by installing smart electricity meters to address the persistence of high electricity distribution losses. Instead of focusing on basic governance and improving their property tax base and collection efficiency, cities chase technology solutions like GIS mapping and municipal bonds. 

There’s a false comfort from believing that infrastructure financing gaps can be leapfrogged by replacing public financewith Public Private Partnerships, private capital, and foreign direct investment. Instead of derisking bank finance for infrastructure, public policy prioritises capital market deepening. Instead of managing the governance of processes and ensuring compliance, tax authorities tend to use targets and coercive practices to boost revenues. Instead of improving governance, the reflex reaction to an adverse news item about abuse of regulation is to double down with more layers of regulation.

All this, more generally, is in line with the pervasive belief that the way to fix struggling public systems is through innovation and doing new things instead of doing the basics right and doing them much better. 

The point here is not to reject these innovative approaches and technologies, but to put them in perspective, as (perhaps even distant) supplementary to the primary activities that get marginalised in the hype. This is especially so in systems entrapped in low baseline of achievement.

Internal trade barriers and deregulation

In the context of the debate on deregulation, internal trade barriers are a significant but less discussed cost. 

This blog post by Luis Garciano (HT: Marginal Revolution) points to an IMF report that looked at internal barriers to trade within the European Union (EU) and found them to be equivalent to tariffs at 45% for goods and 110% for services.

The IMF report finds that while trade costs for goods dropped 16% for EU imports, they fell only 11% for internal EU trade. It finds that internal trade among EU countries is less than half that between US states. 

It also highlights the various types of non-tariff and other barriers affecting different goods and services sectors. The graphic below is the total tariff equivalent for various service sector segments.

Garciano points to the lack of harmonisation among EU member countries’ rules across sectors and the failure to adopt mutual recognition as a means to such harmonisation. 

Tej Parikh has a good graphical summary of the importance of internal barriers as an economic friction.

He points to a graphic from QuantGov that highlights the wide variations in regulatory restrictions across US states, arising especially in occupational licensing, tax disparities, and zoning laws. 

Since the volume of internal freight flows in the US is about $20 trillion annually, even small improvements can have significant effects. 

A striking reflection of internal barriers is that Canada trades more with the US than among its provinces!

Such barriers contribute significantly to regional disparities, as in India. 

And they come in the way of integrating markets and realising scale, the major attractor for foreign investors. 

Fortunately, there are several examples from across the world on the impact of reforms to ease internal barriers. Australia’s Mutual Recognition Act in 1992 enabled goods sold in one state or territory to be sold in another without needing to meet further requirements, and also established equivalence in occupations. This contributed to increased domestic freight movement and productivity growth.

India’s GST led to the dismantling of border checkposts and dramatic reductions in border-crossing times across state borders. 

It also translated into increased household expenditures in districts more exposed to the faster border crossings. In fact, faster border crossings increased household expenditures by 2% to 23% across districts. 

A Canadian study found that varying regulations faced by the truck transportation sector in the country added approximately 8.3% to freight rates. 

Finally, in the context of removing internal barriers, Luis Garciano echoes the critical point that I blogged here, that deregulation in general should not be seen as a stroke-of-pen reform but as one that requires sustained engagement. He points to examples of how the implementation of deregulation often seriously detracts from achieving its objectives. 

There are two problems: first, rather than replacing national regulations, EU rules pile on top of them. Second, member states often engage in ‘gold plating’ – adding extra national requirements when implementing EU directives. The result is that even when the EU does create common rules (directives or regulations aiming to harmonize), the outcome is often not a truly single market. New EU rules often don't replace old national ones. Instead, they create additional layers of regulation. A toy manufacturer might need to obey an EU toy safety directive while simultaneously navigating older national rules about specific materials… In a paper on capital market regulations in a top finance journal, Christiansen, Hail and Leuz (2016) find that, in fact, countries' rules diverge more (!) after harmonizing regulations… 

In other words, the new rules often only serve to add new layers of legislation. For instance, the Single Supervisory Mechanism supervises the banks, but so do the national central banks who impose their own requirements of capital or liquidity on all banks operating in their territory. If a French bank operates in Belgium, it has to satisfy French, Belgian and European regulators. This reduces the synergies of operating across borders, and is the main reason our banking systems are, even today, mostly national. Or consider General Data Protection Regulation, which (in spite of being a regulation) still means we have regulators at EU, national and regional level…A publisher that trades across the EU must now keep separate analytics setups for Austria, France, and Italy, while the same tool remains legal elsewhere. The Draghi report notes that there are around 90 tech-focused laws and more than 270 regulators active in digital networks across all EU countries…

Despite EU directives aimed at facilitating the free movement of professionals, a specialized engineering company based in, say, Portugal, that wins a contract for a major infrastructure project in Germany might find that German authorities require their engineers to undergo a lengthy ‘equivalence check’ despite holding qualifications recognized under EU frameworks. For smaller firms or individual professionals, these hurdles can be prohibitive. 

These problems echo across countries.