Since 2008, global trade has grown slightly more slowly than global GDP. The Doha Round of World Trade Organization negotiations ended in failure. Transatlantic and transpacific trade negotiations are progressing slowly, held back by the resistance of special interests. But, though many experts fear that protectionism is undermining globalization, threatening to impede global economic growth, slower growth in global trade may be inevitable, and trade liberalization is decreasingly important.

To be sure, for 65 years, rapid trade growth has played a vital role in economic development, with average advanced-economy industrial tariffs plummeting from more than 30% to below 5%. The creation of Europe's single market facilitated increased intra-European trade. Japan, South Korea, and Taiwan achieved rapid economic catch-up on the basis of export-led growth. China has followed the same path over the last 30 years. Trade grew about twice as fast as global output from 1990 to 2008.

But there is no reason why trade should grow faster than GDP forever. Indeed, even if there were no trade barriers at all, trade might grow significantly more slowly than GDP in some periods. Several factors make it possible that we are entering such a period.

For starters, there is the changing pattern of consumption in the advanced economies. Richer people spend an increasing share of their income on services that are either impossible to trade (for example, restaurant meals) or difficult to trade (such as health services). Non-tradable sectors tend to account for a growing share of employment and economic activity.

For several decades, that tendency has been offset by ever more intensive trading of tradable goods, often passing through many countries in complex supply chains. In the future, however, the shift to non-tradable consumption may dominate.

Indeed, trade intensity may decline even for manufactured goods. Trade is partly driven by differences in labor costs. China's dramatic manufacturing growth reflected low wages up to now. But as real wages in China and other emerging economies grow, incentives for trade will decline. The more that global incomes converge, the less trade there may be.

In addition, as the economists Erik Brynjolfsson and Andrew McAfee of MIT have argued in their book The Second Machine Age, rapid advances in information technology may enable increasingly extensive automation. Some manufacturing activities, though few jobs, may well return to developed countries, as the advantages of proximity to customers and lower transport costs outweigh decreasingly important differences in labor costs.

Global trade as a share of GDP may therefore decline, but without adverse consequences for global economic growth. Rising productivity does not require relentlessly increasing trade intensity.

Earth, after all, does not trade with other planets, yet its economy still grows. Optimal trade intensity depends on many factors - such as relative labor costs, transport costs, productivity levels, and economy-of-scale effects. Trends in these factors might make reduced trade intensity not only inevitable but desirable.

Even if that is true, international trade will still play a vital role, and preventing any reversal of past trade liberalization is essential. But further trade liberalization is bound to be of declining importance to economic growth.

With industrial tariffs already dramatically reduced most potential benefits of trade liberalization have already been grasped. Estimates of the benefits of further trade liberalization are often surprisingly low - no more than a few percentage points of global GDP.

That is small compared to the cost of the 2008 financial crisis, which has left output in several advanced economies 10-15% below pre-crisis trend levels. It is small, too, compared to the difference in economic performance between successful catch-up countries - such as China - and other countries that have enjoyed the same access to global markets but have performed less well for other reasons.

The main reason for slow progress in trade negotiations is not increasing protectionism; it is the fact that further liberalization entails complex trade-offs no longer offset by very large potential benefits. The Doha Round's failure has been decried as a setback for developing countries. And some liberalization - say, of advanced economies' cotton imports - would undoubtedly benefit some low-income economies. But full trade liberalization would have a complex impact on the least developed economies, some of which would benefit only if compensated for the loss of the preferential access to advanced-economy markets that they currently enjoy.

This implies that further progress in trade liberalization will be slow. But slow progress is a far less important challenge to growth prospects than the debt overhang in developed economies, or infrastructure and educational deficiencies in many developing economies. That reality often goes unacknowledged. The importance of past trade liberalization has left the global policy establishment with a bias toward assuming that further liberalization would bring similar benefits.

But while the potential global benefits of trade liberalization have declined, reduced trade intensity might still impede economic development in some countries. Only a handful of economies over the last 60 years have fully caught up to advanced-economy living standards, and all relied on export-led growth to drive productivity and job creation in manufacturing. Relying solely on that model will be more difficult in the future. China is so big that it must develop domestic drivers of growth at an earlier stage of development than did Japan, Taiwan, or South Korea; as a result, its exports will inevitably decline (relative to GDP).

Meanwhile, for some low-income countries, increased manufacturing and service-sector automation of the sort described by Brynjolfsson and McAfee, whether within advanced economies or within China's established industrial clusters, will make the path to middle- and high-income status more difficult to achieve. That poses important challenges for development policy, which further trade liberalization can alleviate only marginally.