$942 Billion in Buybacks, $500 Billion in Subsidies: The Real Economy Strikes Back

In 1810, Britain was on the brink of famine. The Continental Blockade had cut grain imports from Europe from 114,000 tons in 1807 to just 14,000 tons in 1808. Wheat prices had risen nearly 60%. Napoleon had Britain by the throat. And then he authorized French and allied grain exports to his enemy — not out of mercy, but because he wanted their gold. Mercantilism, the dominant economic framework of 18th-century Europe, measured national wealth in specie: precious metals. In the logic of the system, a starving Britain paying gold for French wheat was a French victory. Napoleon was not subordinating strategy to accounting — he believed the gold was the strategy. Drain the enemy of specie and you drain them of the means to wage war.

Britain paid in gold, survived the blockade, financed another coalition, and Napoleon's empire collapsed four years later. The framework had mistaken the scoreboard for the game.

The confusion did not die at Waterloo. It mutated, resurfacing two centuries later in a form so familiar that most people working inside it cannot see its edges. Shareholder value maximization — the doctrine that a corporation's sole responsibility is to increase its profits, articulated by Milton Friedman in 1970 and operationalized through stock-based compensation, quarterly earnings guidance, dividends, and the buyback — treats accounting profits and share price as the ultimate score, just as mercantilism treated gold. The financial metric is the goal, not the measure. What happens to the real productive base is, by design, someone else's problem.

The arc: from gold to EPS

Mercantilism dominated European statecraft from roughly the 16th century through the Napoleonic Wars. Its core premise: national wealth equals accumulated specie. Exports brought gold in; imports sent gold out. Trade was zero-sum. Colonies existed to supply the metropole with raw materials and absorb its manufactured goods. The framework was not irrational — in an era when wars were funded by the sovereign's treasury and credit markets were shallow, gold reserves were a binding constraint on state power. But it was blind to what we now call the real economy: productive capacity, technological capability, the skills of a workforce.

David Hume's price-specie-flow mechanism (1752) demonstrated the internal contradiction — a trade surplus brings in gold, which raises domestic prices, which makes exports uncompetitive, which reverses the surplus — but the framework persisted for another half-century. Frameworks outlive their disproof. That is their nature.

The 19th century brought free trade and the gold standard. The 20th brought Keynesian demand management, then monetarism, then the Washington Consensus. Through these transformations, the real economy — what a nation makes, builds, and exports — remained the implicit scoreboard. GDP measured output. Trade balances measured competitiveness. Industrial policy, though unfashionable in Anglo-American economics after the 1980s, was the norm in Japan, Korea, Taiwan, and China.

Then came shareholder value. And the American economy, over five decades, reorganized itself around financial metrics rather than productive ones.

The evidence is clearest when you track three numbers against a single denominator: US GDP. Now measure three shares of it — the slice that goes to corporate profits, the slice that goes to productive investment, and the slice that is returned to shareholders. Compare the post-war norm, before Friedman's essay and before the SEC legalized buybacks in 1982, against the post-1980 outcome, with the 1970s as the transition decade.

The first metric, corporate profits after tax, averaged 5 to 7 percent of GDP during the 1950s and 1960s. The long-term average from 1947 to the present sits at 7.4 percent. But from the early 1980s onward, the profit share began a steady climb. By 2012, after-tax corporate profits reached 11.4 percent of national income. By the second quarter of 2026, they had reached 13.2 percent of GDP — near the highest level since the data begins. Capital's share of the American economy roughly doubled from the post-war norm. Looking at that number alone, the notion that shareholder value was a resounding success is irresistible. Other factors — globalisation's downward pressure on wages, the decline of organised labour, the shift from manufacturing to higher-margin services — surely contributed to the profit surge. But the boardroom decisions about whether to reinvest those profits or return them to owners were the mechanism most directly shaped by the framework.

The second metric — productive investment — requires care because the definition of investment itself changed. The Bureau of Economic Analysis reclassified software as a capital investment in 1999 and research and development in 2013. These were analytically sound moves — software and R&D do build productive capacity — but they broke the historical series. To get an apples-to-apples comparison, we can reconstruct a consistent measure of tangible investment (structures and equipment) plus intangible investment (software, R&D, and the broader categories of intellectual property that the economists Carol Corrado, Charles Hulten, and Daniel Sichel have been estimating since the 1950s). The combined picture shows total investment rising from roughly 15 percent of GDP in the 1950s to roughly 25 percent today. But the composition has shifted dramatically. Tangible investment — factories, industrial buildings, equipment — has been broadly stable or mildly declining as a share of GDP. Nearly all the increase has come from intangibles: software code, patent portfolios, brand equity, and organizational capital. These are real investments. They build genuine productive capacity. But they tend to concentrate value in the firms that own the IP, not in the communities and workforces that once benefited from the factory economy. Manufacturing's share of GDP fell from 22 percent in 1979 to 10 percent in 2024.

The third metric — the share returned to shareholders — is where the arc bends most sharply. Before 1982, share buybacks were effectively prohibited; the SEC considered them market manipulation. Dividends were the only channel for returning cash to shareholders. After the SEC adopted Rule 10b-18, buybacks went from zero to the dominant form of corporate payout. According to an NBER study of nonfinancial firms listed on US exchanges, combined buybacks and dividends rose from 19 percent of operating income in the pre-2000 period to 35 percent in the post-2000 period. A Reuters analysis of 1,900 buyback-active companies found that buybacks plus dividends amounted to 38 percent of their capital spending in 1990, 60 percent in 2000, and 113 percent in the years after 2010 — firms were returning more cash to shareholders than they were investing in the business. In the twelve months to March 2025, S&P 500 companies returned $1.64 trillion to shareholders through buybacks and dividends combined — roughly 5.5 percent of GDP. Between 2000 and 2017 alone, US corporations spent nearly $10 trillion repurchasing their own shares.

The three lines, placed side by side, tell a coherent story. American capital captured a growing share of national income. The investment that followed shifted decisively toward intangibles — valuable, but mobile and concentrated. And an ever-larger share was routed directly back to shareholders. The framework did not prevent investment. It redirected it, toward the forms of capital that maximized the financial score.

The Jack Ma paradox

At Davos in 2017, Alibaba founder Jack Ma was asked about US-China trade tensions. His answer was a dissection of the outsourcing framework from the inside. "Thirty years ago, when I just graduated from university, I heard American wonderful strategy. They outsourced the manufacturing jobs, service jobs. They outsourced the manufacturing to Mexico and China, outsourced the service jobs to India." He called it "a perfect strategy." American companies would keep the IP, the brand, the design, the high-margin layers. The rest — the capital-intensive, low-margin, labor-intensive work — would go to countries eager to industrialize.

Then came the punchline. "It's not that other countries steal jobs from you guys. It's your strategy." The billions in profits that outsourcing generated were not reinvested in American infrastructure or workers. "In the past 30 years, America has had 13 wars at a cost of $14.2 trillion. That's where the money went." The strategy was not a failure, Ma argued. The failure was what America did with the proceeds — and the fact that it was now blaming China for an outcome its own framework had engineered.

The arrangement had worked brilliantly for 25 years. American consumers got cheap goods. American shareholders got soaring returns. Chinese workers got factories and — eventually — skills, supply chains, and their own industrial ecosystem.

The framework's internal contradiction was the same one Hume identified in mercantilism. By outsourcing production, Western firms transferred not just costs but capabilities. Chinese firms moved up the value chain. BYD, which began making batteries for mobile phones, now builds electric vehicles that terrify Detroit and Wolfsburg. Huawei, once a manufacturer of telephone switches under contract, now designs advanced semiconductors that trigger US export controls. The outsourcing framework maximized the financial metric while systematically eroding the real-economy base that the financial metric was supposed to measure.

The shareholder-value framework said outsourcing production was rational because it maximized EPS. The same companies and the same policymakers now say depending on that outsourced production is a national security risk. Both statements cannot be simultaneously true. Either EPS is not the right metric, or national security is not a real concern. The inconsistency is the tell: the framework is breaking.

Frameworks outlive their disproof

The parallel with mercantilism is not that the old framework was wrong and the new one is right. The parallel is structural: both mercantilism and shareholder value maximization elevated a financial metric over the real economy, and both became self-defeating.

The mercantilist pursuit of gold left France with a hollowed-out productive base and an adversary that, denied specie, innovated its way to industrial supremacy. The shareholder-value pursuit of EPS left the US with soaring corporate profits, record payouts to shareholders, and a productive base it is now spending half a trillion dollars to rebuild — and an industrial competitor that used the outsourcing era to close the R&D gap.

Frameworks do not change because they are disproven. They change because they stop working for the people who operate them. The mercantilist framework survived Hume's refutation by four decades. It died when Napoleon lost. The shareholder-value framework survived multiple financial crises and two decades of wage stagnation. It is now under pressure because the consequences — fragile supply chains exposed by COVID, a semiconductor dependency that looks like a geopolitical vulnerability, an adversary that used outsourcing to close the technology gap — have become impossible to ignore.

The emerging framework does not yet have a name. You can see its outlines in the policy instruments: tariffs, industrial subsidies, export controls, carbon border taxes, local content requirements. It measures success not by the financial return on a single firm's capital but by the resilience, diversity, and technological autonomy of the productive base. It asks not "what is the margin?" but "where is the factory?"

Two open questions sit at the center of whether this framework consolidates.

The first is who defines it. The Western version — CHIPS Act, IRA, CBAM — is a series of spending authorizations and regulatory instruments built atop an intellectual tradition that spent 40 years treating industrial policy as heresy. The Chinese version is encoded in five-year plans that have systematically raised R&D intensity from 0.9 percent of GDP in 2000 to 2.8 percent in 2025, surpassing the OECD average, with a stated target of 7 percent annual R&D growth through 2030. The Western approach is reactive, legislated in response to crises. The Chinese approach is institutionalized, embedded in the planning apparatus. Which one proves more durable is not a philosophical question. It is a question about which institutional form can sustain a real-economy framework across political cycles.

The second is whether the deeper confusion is not about shareholder value at all, but about something one level down: our GDP accounting system. GDP is itself a mercantilist relic — it counts the financial value of output, not the resilience of the system that produces it. A semiconductor fab that is profitable but dependent on a foreign supplier for lithography equipment counts the same in GDP as one that is vertically integrated. A buyback and a factory both register as economic activity. The GDP framework was designed in the 1930s to measure the flow of money through a national economy. It was never designed to measure the structural autonomy of that economy. If the new framework is to outlast its predecessors, it will need a scoreboard that measures what it values — not just a different policy mix layered on top of the old accounting.

Napoleon, counting his English gold while his blockade crumbled, was following the best framework of his era. The question is whether the people now building factories and imposing tariffs can see what their own framework is blind to.

---

Sources: NBER Working Paper 11344 (Napoleon's wheat exports); Britannica (Continental System grain data); BLS (manufacturing employment share); BEA/NIST (manufacturing GDP share, corporate profits, nonresidential fixed investment); FRED series A008RE1Q156NBEA (nonresidential fixed investment share of GDP); Corrado, Hulten, and Sichel (2005, 2006, 2009 — intangible investment reconstruction going back to 1950s); US Census Bureau via FRED (US-China trade balance); NBER Working Paper 26958 (buybacks/dividends as share of operating income); Reuters (buybacks/dividends as share of capex, 1990–2015); S&P Dow Jones Indices (buyback and dividend totals through Q1 2025); NSF/NCSES Science & Engineering Indicators (global R&D shares); OECD (China R&D intensity); Yicai Global (China 15th Five-Year Plan R&D targets); CNBC / World Economic Forum (Jack Ma Davos 2017 transcript).