IT WAS as though the world had a new appetite. A Kentucky Fried Chicken (KFC) outlet opened near Tiananmen Square in 1987. In 1990 a McDonald’s sprang up in Pushkin Square, flipping burgers for 30,000 Muscovites on its first day. Later that year Ronald McDonald rolled into Shenzhen, China, too. Between 1990 and 2005 the two companies’ combined foreign sales soared by 400%.

McDonald’s and KFC embodied an idea that would become incredibly powerful: global firms, run by global managers and owned by global shareholders, should sell global products to global customers. For a long time their planet-straddling model was as hot, crisp and moreish as their fries.

Today both companies have gone soggy. Their shares have lagged behind America’s stockmarket over the past half-decade. Yum, which owns KFC, saw its foreign profits peak in 2012; they have fallen by 20% since. Those of McDonald’s are down by 29% since 2013 (see article). Last year Yum threw in the towel in China and spun off its business there. On January 8th McDonald’s sold a majority stake in its Chinese operation to a state-owned firm. There are specific reasons for some of this; but there is also a broader trend. The world is losing its taste for global businesses.

Their detractors and their champions both think of (unless otherwise specified)—as the apex predators of the global economy. They shape the ecosystems in which others seek their living. They direct the flows of goods, services and capital that brought globalisation to life. Though multinationals account for only 2% of the world’s jobs, they own or orchestrate the supply chains that account for over 50% of world trade; they make up 40% of the value of the West’s stockmarkets; and they own most of the world’s intellectual property.

Although the idea of being at the top of the food chain makes these companies sound ruthless and all-conquering, rickety and overextended are often more fitting adjectives. And like jackals circling an elderly pride, politicians want to grab more of the spoils that multinational firms have come to control, including 80m jobs on their payrolls and their profits of about $1trn. As multinational firms come to make ever more of their money from technology services they become yet more vulnerable to a backlash. The predators are increasingly coming to look like prey.

It all looked very different 25 years ago. With the Soviet Union collapsing and China opening up, a sense of destiny gripped Western firms; the “end of history” announced by Francis Fukuyama, a scholar, in which all countries would converge towards democracy and capitalism seemed both a historical turning-point and a huge opportunity. There were already many multinationals, some long established. Shell, Coca-Cola and Unilever had histories spanning the 20th century. But they had been run, for the most part, as loose federations of national businesses. The new multinationals sought to be truly global.

Companies became obsessed with internationalising their customers, production, capital and management. Academics draw distinctions between going global “vertically”—relocating production and the sourcing of raw materials—and “horizontally”—selling into new markets. But in practice many firms went global every which way at once, enthusiastically buying rivals, courting customers and opening factories wherever the opportunity arose. Though the trend started in the rich world, it soon caught on among large companies in developing economies, too. And it was huge: 85% of the global stock of multinational investment was created after 1990, after adjusting for inflation (see chart 1).

By 2006 Sam Palmisano, the boss of IBM, was arguing that the “globally integrated enterprise” run as a unitary organisation, rather than as a federation, would transcend all borders as it sought “the integration of production and value delivery worldwide”. From the Seattle demonstrations of 1999 onwards, anti-globalisation activists had been saying much the same, while drawing less solace from the prospect. The only business star to resist the orthodoxy was Warren Buffett; he sought out monopolies at home instead.

Such a spree could not last forever; an increasing body of evidence suggests that it has now ended. In 2016 multinationals’ cross-border investment probably fell by 10-15%. Impressive as the share of trade accounted for by cross-border supply chains is, it has stagnated since 2007 (see chart 2). The proportion of sales that Western firms make outside their home region has shrunk. Multinationals’ profits are falling and the flow of new multinational investment has been declining relative to GDP. The global firm is in retreat.

The other end of the end of history

To understand why this is, consider the three parties that made the boom possible: investors; the “headquarters countries” in which global firms are domiciled; and the “host countries” that received multinational investment. For their different reasons each thought that multinational firms would provide superior financial or economic performance.

Investors saw a huge potential for economies of scale. As China, India and the Soviet Union opened up, and as Europe liberalised itself into a single market, firms could sell the same product to more people. And as the federation model was replaced by global integration, firms would be able to fine-tune the mix of inputs they got from around the world—a geographic arbitrage that would improve efficiency, as Martin Reeves of BCG, a consultancy, puts it. From the rich world they could get management, capital, brands and technology. From the emerging world they could get cheap workers and raw materials as well as lighter rules on pollution.

These advantages led investors to think global firms would grow faster and make higher profits. That was true for a while. It is not true today. The profits of the top 700-odd multinational firms based in the rich world have dropped by 25% over the past five years, according to FTSE, an index firm. The weakness of many currencies against the dollar is part of the story, but explains only a third of the fall. The profits of domestic firms rose by 2%.

A complementary measure comes from the foreign profits of all firms as recorded in balance-of-payments statistics. Though the data refer to firms of all sizes, big ones dominate the mix. For companies with headquarters in the OECD, a club of mostly rich countries, foreign profits are down by 17% over five years. American firms suffered less, with a 12% drop, partly because of their skew towards the fast-growing technology sector. For non-American firms the drop was 20%.

Profits should be compared with the capital sunk. The return on equity (ROE) of the top 700 multinationals has dropped from a peak of 18% a decade ago to 11%. The returns on the foreign operations of all firms have fallen, too, based on balance-of-payments statistics. For the three countries which have, historically, hosted the most and biggest multinationals, America, Britain and the Netherlands, ROE on foreign investment has shrunk to 4-8%. The trend is similar across the OECD (see chart 3).

Multinationals based in emerging economies, which account for about a seventh of global firms’ overall activity, have fared no better: their worldwide ROE is 8%. Several supposed champions—such as Lenovo, the Chinese company which bought IBM’s PC business and parts of Motorola—have been financial flops. China’s biggest completed cross-border acquisition was of Nexen, a Canadian oil firm, in 2012. Last year the buyer, CNOOC, a state-owned energy firm, wrote a chunk of it off.

About half of the deterioration in multinationals’ ROE over the past 5-10 years is explained by the slump in commodity prices, and thus the profits of oil firms, mining firms and the like. Another 10% of the deterioration is due to banks. Firms that provide the specialist services behind globalisation have also been hammered. Profits have dropped by over 50% from their peak at Maersk, a Danish shipping line, Mitsui, a Japanese trading house, and Li & Fung, a supply-chain agent for retailers.

The pain extends beyond these core industries, however. Half of all big multinationals have seen their ROE fall in the past three years; 40% fail to make an ROE of over 10%, widely seen as a benchmark of whether a firm is creating any value worth speaking of. Even at powerhouses such as Unilever, General Electric (GE), PepsiCo and Procter & Gamble, foreign profits are down by a quarter or more from their peak. The only bright spot is the technology giants. Their foreign profits comprise 46% of the total foreign earnings of the top 50 American multinationals, up from 17% a decade ago. Apple made $46bn abroad last year, more than any other firm and five times more than GE, often seen as America’s bellwether.

These figures mean multinationals are no longer achieving superior performance. The Economist has examined the record of the 500 largest firms worldwide. In eight out of ten sectors, multinational firms have expanded their aggregate sales more slowly than their domestic peers. In six out of the ten sectors they have lower ROEs (see chart 4). For American firms, returns are now 30% higher in their home market, where cosy oligopoly has become more enticing than the hurly-burly of an unruly world.

Individual bosses will often blame one-off factors: currency moves, the collapse of Venezuela, a depression in Europe, a crackdown on graft in China, and so on. But the deeper explanation is that both the advantages of scale and those of arbitrage have worn away. Global firms have big overheads; complex supply chains tie up inventory; sprawling organisations are hard to run. Some arbitrage opportunities have been exhausted; wages have risen in China; and most firms have massaged their tax bills as low as they can go. The free flow of information means that competitors can catch up with leads in technology and know-how more easily than they used to.

As a result firms with a domestic focus are winning market share. In Brazil two local banks, Itaú and Bradesco, have trounced global lenders. In India Vodafone, a Western mobile-phone operator and Bharti Airtel, an Indian multinational active in 20 countries, are losing customers to Reliance, a domestic firm. In America shale firms stole a march on the global oil majors. In China local dumpling brands are eating into KFC’s sales. A blend of measures for listed firms shows that multinationals’ share of global profits, 35% a decade ago, is now only 30%.

So much for the investors. What about the second constituency for multinationals, the “headquarters countries”? In the 1990s and 2000s they wanted their national champions to go global in order to become bigger and brainier. A study by McKinsey, a consultancy, based on 2007 data, outlined the sort of benefits they were after. Multinationals operating in America, which accounted for 19% of private-sector jobs, were responsible for 25% of private wages, 25% of profits, 48% of exports and 74% of research and development. Go them.

Citizens of nowhere

The mood changed after the financial crisis. Multinational firms started to be seen as agents of inequality. They created jobs abroad, but not at home. Between 2009 and 2013, only 5%, or 400,000, of the net jobs created in America were created by multinational firms domiciled there (although preliminary figures suggest that job creation picked up sharply in 2014). The profits from their hoards of intellectual property were pocketed by a wealthy shareholder elite. Political willingness to help multinationals duly lapsed.

As a result, the tapestry of rules designed to help businesses globally is fraying. Global accounting, antitrust, money-laundering and bank-capital rules have splintered into American and European camps. Takeovers of Western firms now often come with strings attached by governments to safeguard local jobs and plants. Two American-led trade deals, known as TPP and TTIP, that gave protection to intellectual property, have flopped. The global tribunals that multinationals use to bypass national courts have come under attack.

The deep roots of globalisation mean that trying to favour domestic companies by erecting tariffs no longer works as once it did. Over half of all exports, measured by value, cross a border at least twice before reaching the end-customer, so such tariffs hurt all alike. This does not mean that the inept or ignorant will not try them. But it does encourage the use of other avenues to try and right perceived wrongs, such as the tax system and good old political muscle.

A typical multinational has over 500 legal entities, some based in tax havens. Using American figures, it pays a tax rate of about 10% on its foreign profits. The European Union (EU) is trying to raise that figure. It has cracked down on Luxembourg, which offered generous deals to multinationals that parked profits there; it also hit Apple with a $15bn penalty for breaching state-aid rules by booking profits in Ireland, with which it had a bespoke tax deal. America, for its part, has barred big firms from using legal “inversions” to shift their tax base abroad, most notably in the case of Pfizer, a pharmaceutical company that is America’s third-largest foreign earner.

Republicans in Congress are debating changes to the tax code which would see exporters and firms bringing profits home pay less than before, while firms shifting production abroad would face levies. Meanwhile, some firms have apparently been browbeaten into outsourcing decisions about where to base factories by Donald Trump, the new American president. On January 3rd Ford, a carmaker, agreed to cancel a new plant in Mexico and invest more at home. Mr Trump also wants Apple to shift more of its supply chain home.

If these trends continue global firms’ tax and wage bills will rise, squeezing profits further. If American multinationals shifted a quarter of their foreign jobs home, at American wage rates, and paid the same tax rate abroad as they did at home, their profits would fall by another 12%. This excludes the cost of building the new plants in America.

Of all those involved in the spread of global businesses, the “host countries” that receive investment by multinationals remain the most enthusiastic. The example of China, where by 2010 30% of industrial output and 50% of exports were produced by the subsidiaries or joint-ventures of multinational firms, is still attractive.

Argentina’s government wants to draw in foreign firms. Mexico has just sold stakes in its oilfields to foreign firms, including ExxonMobil and Total. India has a campaign called “make in India” to attract multinational supply chains. An index through which the OECD seeks to gauge the openness of host countries shows no overall deterioration since the financial crisis.

But there are gathering clouds. China has been turning the screws on foreign firms in a push for “indigenous innovation”. Bosses say that more products have to be sourced locally and intellectual property often ends up handed over to local partners. Strategic industries, including the internet, are out of bounds to foreign investment. Many fear that China’s approach will be mimicked around the developing world, forcing multinational firms to invest more locally and create more jobs—a mirror image of the pressures placed on them at home.

The price of hospitality

Host countries may also become less welcoming as activity shifts towards intangible services. For the top 50 American multinationals, 65% of foreign profits now come from industries reliant on intellectual property, such as technology, drug patents and finance. A decade ago it was 35%, and the share is still rising. (It is much lower in Europe and Japan, which do not have big technology firms.) There is no serious appetite among multinationals to recreate in Africa or India the manufacturing centres they spurred on in China, which removes a reason for those host countries to welcome them. The jobs and exports that can be attributed to multinationals are already a diminishing part of the story. In 2000 every billion dollars of the stock of worldwide foreign investment represented 7,000 jobs and $600m of annual exports. Today $1bn supports 3,000 jobs and $300m of exports.

Silicon Valley’s latest stars are already controversial abroad. In 2016 Uber sold its Chinese operations to a local rival after a brutal battle. In December India’s two digital champions, Ola, a ride-hailing firm, and Flipkart, an e-commerce site, said the government should protect them against Uber and Amazon. They argued that their rivals would build monopolies, create few good jobs and ship the profits to America.

The last time the multinational company was in trouble was in the aftermath of the Depression. Between 1930 and 1970 their stock of investment abroad fell by about a third relative to global GDP; it did not recover until 1991. Some firms “hopped” across tariffs by building new factories within protectionist countries. Many restructured, ceding autonomy to their foreign subsidiaries to try to give them a local character. Others decided to break themselves up.

Today multinationals need to rethink their competitive advantage again. Some of the old arguments for going global are obsolete—in part because of the more general successes of globalisation. Most multinationals do not act as internal markets for trade. Only a third of their output is now bought by affiliates in the same group. External supply chains do the rest. Multinational firms no longer have a lock on the most promising ideas about management or innovation. Where they have enforceable patents over valuable brands they are still at an advantage, as they are in products, such as jet engines, where economies of scale are best created by spreading costs over the entire world. But those benefits are less than they were.

The lack of advantage is revealed in the amount of activity that yields little value. Roughly 50% of the stock of foreign direct investment makes an ROE of less than 10% (40% of the stock if you exclude natural-resources firms). Ford and General Motors make 80% or more of their profits in North America, suggesting their foreign returns are abysmal.

Many industries that tried to globalise seem to work best when national or regional. For some, the penny has dropped. Retailers such as Britain’s Tesco and France’s Casino have abandoned many of their foreign adventures. America’s telecoms giants, AT&T and Verizon, have put away their passports. Financial firms are focusing on their “core” markets. LafargeHolcim, a cement maker, plans to sell, or has sold, businesses in India, South Korea, Saudi Arabia and Vietnam. Even successful global firms have gone on diets. P&G’s foreign sales have dropped by almost a third since 2012 as it has closed or sold weak businesses.

It looks as if, in the future, the global business scene will have three elements. A smaller top tier of multinational firms will burrow deeper into the economies of their hosts, helping to assuage nationalistic concerns. General Electric is localising its production, supply chains and management. Emerson, a conglomerate that has over 100 factories outside America, sources about 80% of its production in the region where it is sold. Some foreign firms will invest more deeply in American-based production in order to avoid tariffs, if Mr Trump imposes them, much as Japanese car firms did in the 1980s. This is doable if you are large. Siemens, a German industrial giant, employs 50,000 in America and has 60 factories there. But midsized industrial firms will struggle to muster the resources to invest more deeply in all their markets.

Politicians will increasingly insist that companies buying foreign firms promise to preserve their national character, including jobs, R&D activity and tax payments. SoftBank, a Japanese firm that bought ARM, a British chip company, in 2016, agreed to such commitments. So has Sinochem, a Chinese chemicals firm that is buying Syngenta, a Swiss rival. The boom in foreign takeovers by Chinese firms, meanwhile, may fizzle out or explode. Many such deals, reliant on subsidised loans from state banks, probably make little financial sense.

The second element will be a brittle layer of global digital and intellectual-property multinationals: technology firms, such as Google and Netflix; drugs companies; and companies that use franchising deals with local firms as a cheap way to maintain a global footprint and the market advantage that brings. The hotel industry, with its large branding firms such as Hilton and Intercontinental, is a prime example of the tactic. McDonald’s is shifting to a franchising model in Asia. These intangible multinationals will grow fast. But because they create few direct jobs, often involve oligopolies and do not benefit from the protection of global trade rules, which for the most part only look after physical goods, they will be vulnerable to nationalist backlashes.

The seeds of something more

The final element will be perhaps the most interesting: a rising cohort of small firms using e-commerce to buy and sell on a global scale. Up to 10% of America’s 30m or so small firms already do this to some extent. PayPal, a digital payments firm, says its cross border transactions, which include activity from such multinationalettes, are running at $80bn a year, and growing fast. Jack Ma, the boss of Alibaba, a Chinese e-commerce firm, predicts that a wave of small Western firms exporting goods to Chinese consumers will go some way to reversing the past two decades of massive American firms importing goods from China.

The new, prudent age of the multinational will have costs. Countries that have grown used to global firms throwing cash around may find that competition abates and prices rise. Investors, who all told have a third or more of their equity portfolios tied up in multinational firms, could face some unpleasant turbulence. Economies that rely on income from foreign investments, or capital inflows from new ones, will suffer. The collapse in profits from British multinationals is the reason why Britain’s balance of payments looks bad. Of the 15 countries with current-account deficits of over 2.5% of GDP in 2015, 11 relied on fresh multinational investment to finance at least a third of the gap.

The result will be a more fragmented and parochial kind of capitalism, and quite possibly a less efficient one—but also, perhaps, one with wider public support. And the infatuation with global companies will come to be seen as a passing episode in business history, rather than its end.