Asia-Pacific Investment Report 2011 — The Map Nobody Had Drawn

In 2011 two separate natural disasters in Asia interrupted manufacturing worldwide, and the pattern of the interruption revealed something structural that had been accumulating for decades.

The revealed facts were consistent across both events:

• Firms with several qualified suppliers discovered those suppliers depended on a common source, frequently several tiers below the relationship they managed.
• The common source was often the only meaningful global producer of a specialised input — a particular grade of material, a specific component, a coating or pigment.
• The affected firms did not know this before the interruption. Procurement visibility extended to direct suppliers and stopped.
• And the concentration was geographic as well as corporate. Even where several producers existed, they were frequently clustered in the same region — sometimes the same industrial park.

Why this happens without anyone choosing it is the report's central mechanism. Specialised manufacturing exhibits strong economies of scale and learning. The producer with the largest volume has the lowest cost and the most accumulated process knowledge, wins more business, and extends the advantage. Over decades this converges toward one or two global producers for any sufficiently specialised input.

No participant makes a bad decision. Each buyer selects the best supplier on price, quality and reliability. The aggregate outcome is a single point of failure that emerges from a series of individually correct choices — the same structure as the financial concentration the Global Investment Outlook 2011 describes, arriving through manufacturing.

The second finding concerns insurance. Most affected firms found their business interruption cover did not respond, because the standard policy triggers on physical damage to the insured's own property. A firm whose factory was undamaged but whose supplier's factory was destroyed had a large loss and no claim.

How concentration happens without being chosen

The mechanism deserves careful development, because it explains why the problem was invisible rather than ignored.

Start with a specialised input — say, a particular chemical used in a manufacturing process, required in small quantities, with demanding purity specifications.

The economics of producing it:

• High fixed costs — the plant, the process development, the qualification testing.
• Low marginal costs once running.
• Strong learning effects. Yield improves with cumulative production, and yield is most of the cost.
• And high switching costs for customers, since qualifying a new supplier requires testing that can take a year or more.

The competitive dynamic that follows:

1. Several producers exist initially.
2. One achieves higher volume, so its unit costs fall and its yields improve.
3. It wins more customers on price and reliability.
4. Its advantage compounds. Return to step 2.
5. Sub-scale competitors exit, because the fixed costs cannot be covered at low volume.

The end state is one or two global producers, which is the efficient outcome by every standard measure — lowest cost, highest quality, least capital tied up in duplicate capacity.

Why no buyer prevents it:

• Each buyer's decision is individually correct. Choosing the cheapest, best qualified supplier is what a procurement function is for.
• The buyer cannot see the aggregate. It knows its own suppliers, not the industry's structure.
• And maintaining a second source is costly — qualification, minimum volumes, higher prices — with a benefit that is invisible until it is enormous.

The concentration was nobody's decision. It was the sum of thirty years of everyone correctly choosing the best supplier available, and it produced a structure no one would have designed.

The tier problem compounds it. A firm may deliberately maintain three suppliers of a component. If all three buy the same sub-component from the same source, the diversification is nominal. Detecting this requires visibility three or four levels deep, which requires cooperation from suppliers who consider their own supply chains confidential.

Why the insurance did not respond

The insurance finding is practically important and was a genuine surprise to most affected firms.

How business interruption cover normally works: it pays for lost profit when the insured cannot operate because of physical damage to the insured's own property — a fire, a flood, a collapse. The physical damage trigger is fundamental to the product, because it makes the loss verifiable and bounds the insurer's exposure.

What happened in 2011: a firm's own factory was undamaged and fully operational. It could not produce because a component did not arrive. There was no damage to its property, so the standard trigger was not met.

The available extensions and their limits:

• Contingent business interruption cover exists and does respond to a supplier's damage. But it typically requires the supplier to be named in the policy — and a firm that does not know its tier-three suppliers cannot name them.
• Limits were generally low relative to the losses, because the exposure had been underwritten as a minor extension rather than a principal risk.
• And the causal chain must be documented. Proving that a specific loss resulted from a specific supplier's damage, through several intermediate firms, is evidentially demanding.

The structural point generalises well beyond insurance:

A risk that has not been named cannot be transferred. The insurance gap was a direct consequence of the visibility gap — you cannot insure a dependency you have not identified.

The market response over subsequent years was the development of broader supply chain cover, non-damage triggers, and parametric products paying on a defined event rather than a proven loss. These remain a small share of the market, and the Global Investment Outlook 2021 shows the gap was still substantial a decade later.

The trade-off that had only ever been made one way

The events forced attention onto a choice that firms had been making implicitly, in one direction, for thirty years.

Efficiency and resilience are genuinely in tension, and the tension is not resolvable by cleverness:

Efficiency	Resilience
Minimal inventory	Buffer stock
Single best supplier	Multiple qualified suppliers
Geographic concentration for logistics	Geographic distribution
Highest-yield process	Redundant capacity
Capital returned to shareholders	Capital held against disruption

Every resilience measure costs money continuously. Every efficiency measure returns money continuously.

Why the choice had gone one way for so long:

• The efficiency gains are certain, immediate and measurable. They appear in this quarter's margin.
• The resilience benefit is uncertain, deferred and counterfactual. It appears as a loss that did not happen, which no accounting system records.
• Capital markets reward the first and cannot observe the second. A firm holding buffer inventory has worse working capital metrics and no way to demonstrate why.
• And the period had been unusually stable, so the base rate of disruption observed by any individual manager was very low.

This is the same asymmetry the Asia-Pacific Investment Report 2008 identifies in reserve accumulation — continuous visible cost against rare invisible benefit — and it produces the same systematic under-provision.

What actually changed after 2011 was less than the commentary suggested. Firms mapped their chains more deeply, and some added buffers. But the economic incentives were unchanged, and buffers erode when they are not tested. The Global Investment Outlook 2021 documents the same absent buffers a decade later, which is the strongest evidence that the 2011 lesson did not durably alter behaviour.

The honest conclusion is that this is not a knowledge problem. Firms understood the trade-off. It is an incentive problem, and it persists because the party bearing the disruption cost is often not the party that captured the efficiency gain.

Integration and correlation

The region's economic integration is usually discussed as an unambiguous good, and 2011 showed its risk dimension.

What integration achieved: a manufacturing system where each economy specialises in the stages it does best, components move across borders multiple times, and the finished product is far cheaper than any single-country production could achieve. The efficiency gains are large and real.

What it also created:

• Geographic clustering of specialised stages. Integration means each stage happens where it is done best, which means each stage happens in one place.
• Correlated exposure to regional events. A natural disaster, a port closure or a political disruption affects many stages of many chains simultaneously.
• And thin buffers between stages, since integration's efficiency comes precisely from eliminating the inventory that borders used to require.

The investment implication is about apparent versus actual diversification:

A portfolio holding manufacturers in five countries across the region may hold five positions on the same industrial cluster. The country labels suggest diversification the production map does not support.

This is the archive's most persistent finding arriving through a new route — the same structure as euro-area sovereign and bank exposure in the Global Investment Outlook 2010, and as energy credit inside diversified high yield in the US Venture Capital Report 2014. In each case the line items are diverse and the underlying exposure is one thing.

The practical test is the same in all three: ask what single event would damage every position at once. If the answer is short and specific, the diversification is nominal.

Japan's position in the chain

Japan's specific role illustrates why a mature economy with modest headline growth can matter enormously to global production.

The position: Japanese firms held dominant global shares in a range of specialised upstream inputs — advanced materials, precision components, specialised chemicals and manufacturing equipment. These are typically low-volume, high-value, technically demanding products with long qualification cycles.

Why the position was durable:

• Accumulated process knowledge that is genuinely hard to replicate, much of it not written down.
• Long qualification cycles that lock in incumbents once approved.
• Small markets by revenue, which makes entry unattractive relative to the technical difficulty — the market is too small to be worth attacking and too critical to do without.

The consequence for the global system: an economy frequently described as stagnant, on the basis of its headline growth, occupied positions in the production chain whose interruption stopped assembly lines worldwide.

The investment lesson is about where value and criticality sit. Headline growth measures the size of an economy's output, not its systemic importance. A supplier of an irreplaceable input has pricing power and criticality entirely disproportionate to its revenue — which is a good business and a systemic vulnerability at the same time.

The Japan Investment Report 2019 develops this position in detail, and the Asia-Pacific Investment Report 2018 covers what happened when these dependencies became a policy concern rather than a commercial one.

One flood, one industry, one price

The clearest single demonstration came from the second event, and it is worth following in detail because the whole chain is documented.

The setup: a substantial share of global production of a particular computing component was concentrated in one country, and much of that within a small number of industrial estates on the same floodplain. The concentration had emerged through the process described above — scale economics, supplier clustering around assemblers, and thirty years of individually sensible location decisions.

What happened when the estates flooded:

• Global output of the component fell sharply, because the affected capacity was a large share of the world total.
• Prices roughly doubled and stayed elevated for more than a year, since replacement capacity takes that long to build and qualify.
• Downstream manufacturers on other continents cut production of finished goods they could not complete.
• And the effects reached firms with no supplier in the affected country at all, because they competed for the same constrained global supply.

Four features make this the most instructive case in the period:

The affected firms were mostly not in the flood zone. The damage was physical and local; the economic loss was global and mostly suffered by parties with undamaged property — which is exactly the insurance gap described above.

The price effect lasted far longer than the disruption. Production resumed within months; prices took over a year to normalise, because rebuilding qualified capacity is slow and because buyers rebuilt inventory, adding demand on top of recovering supply.

Competitors of the affected producers benefited enormously, capturing volume and pricing they could not have won on merit. A disruption is a transfer within an industry as well as a loss to it — which means the sector-level impact is far smaller than the impact on the affected firms, and a sector exposure measures the wrong thing.

And the concentration was entirely knowable in advance. Production shares by country were published. Nobody had connected the industry map to the flood map, because those are different documents held by different professions.

The information required to see this coming existed in full. It was distributed across a production atlas, a hydrology map and a corporate ownership register, and no one had reason to hold all three.

The Global Investment Outlook 2021 shows the same pattern in a different component, a decade later, with the same surprise.

What an allocator could act on

Ask what single event would damage every position simultaneously. Country labels and sector labels both conceal shared dependence on a single industrial cluster or a single upstream producer.

Treat tier-one supplier diversification as unverified. Three qualified suppliers buying from one sub-supplier is not diversification, and the information required to check sits below where procurement normally looks.

Ask manufacturers directly what they know about tiers two and three. The quality of the answer is itself the finding — after 2011, an inability to answer is a management signal rather than an information gap.

Note that a named risk is a transferable one and an unnamed risk is not. Contingent business interruption cover generally requires the supplier to be identified, so visibility gaps become insurance gaps mechanically.

Read persistent under-provision of resilience as an incentive problem, not an ignorance one. The efficiency gain is measured and rewarded; the avoided loss is counterfactual and invisible.

Look for criticality rather than size in upstream suppliers. A small, slow-growing producer of an irreplaceable input has pricing power and systemic weight that headline metrics miss entirely.

What 2011 established

• Geographic and corporate concentration emerges from ordinary competition, with scale and learning effects converging toward single global producers.
• Visibility stops at tier one, so the concentration was undetected rather than accepted.
• Standard business interruption cover did not respond, because it triggers on the insured's own physical damage.
• The efficiency-resilience trade-off is an incentive problem, which is why the 2011 lesson did not durably change behaviour.
• Regional integration concentrated exposure geographically even as it diversified it nationally.