Why Trade Grew Faster Than Output for Decades

The options around the growth of trade relative to output are set out side by side below, with the conditions that genuinely favour one over the other.

The observation

Over an extended period, measured world trade grew considerably faster than measured world output, which attracted a great deal of comment. The ratio of trade to output rose steadily, and that ratio became a standard indicator of how integrated the world economy had become. More recently the two growth rates have been closer, which prompted equally extensive commentary about integration stalling or reversing.

Both episodes are better understood by asking what the numerator actually counts rather than by treating the ratio as a single meaningful quantity. The measurement issue does not explain everything, and it explains enough to change the interpretation substantially.

Fragmentation inflates the numerator

When production splits across countries, an intermediate good crosses borders several times and its full value is counted at each crossing. Output, by contrast, counts value added once, so a product made in three countries adds three border crossings and one unit of output.

Deepening fragmentation therefore raises measured trade relative to output even when the underlying activity grows at the same rate. As fragmentation deepened rapidly, this mechanical effect contributed meaningfully to the observed divergence in growth rates. When fragmentation stopped deepening at the same pace, that contribution faded and the growth rates converged.

The genuine expansion underneath

Falling transport costs made goods tradable that previously were not, which is a real expansion rather than a counting effect. Reduced tariffs and improved trade administration lowered the cost of crossing borders, adding to the range of viable trade.

Communication technology allowed coordination of dispersed production, which enabled the fragmentation itself. Large economies integrating into world markets added substantial new trade that had genuinely not existed before. All of these are real effects, and the measurement artefact sits on top of them rather than replacing them.

Composition matters

Trade in goods is measured comprehensively at borders while trade in services is measured less completely and with more difficulty. As economies shift towards services, a growing share of economic activity is in a category where trade is harder to observe. Digitally delivered services in particular are poorly captured by frameworks designed for physical shipments.

The apparent slowdown in trade growth may therefore partly reflect activity moving into categories that statistics measure less well.

This is an active area of statistical work, and estimates of the scale of the gap vary considerably.

Why the ratio is a poor summary

Comparing gross trade with value-added output puts two different kinds of number into a ratio, which limits how much it can mean. The ratio also varies enormously by country size, since small economies trade more relative to output for reasons of scale alone. Changes in commodity prices move the trade figure without any change in volumes, adding further noise to the series.

None of that makes the indicator useless, and it does mean single-year movements deserve less attention than they usually receive. Value-added trade measures give a cleaner picture of integration and are available with more delay and less granularity.

Company disclosures describe a supply chain one tier deep, and the fragile part is usually three tiers down.

What the long view suggests

Integration deepened substantially over several decades through a combination of real cost reductions and structural fragmentation. The pace of deepening has changed, which is a different statement from integration reversing, and the evidence supports the first more than the second. Trade patterns have also become more regional in some sectors, which changes the geography without necessarily reducing the volume.

Over a shipping cycle, predicting the next phase requires assumptions about technology, policy and costs that nobody can make with confidence. Describing the mechanisms that would drive it in either direction is more useful than forecasting which will dominate.

Side by side

The takeaway

The trade-to-output ratio mixes two different measures. Read the mechanisms rather than the ratio.

Capacity takes a decade to build and one quarter to look like a mistake.