The Factory Map Lie

Crocs makes 100% of its shoes outside the United States. After 2018 tariffs on Chinese goods, the company shifted production from China to Vietnam and called it diversification. Analysts praised the move. “China risk mitigated.” The factory map showed pins in Southeast Asia instead of mainland China. Problem solved.

Then tariffs hit Vietnam too, hard enough to wipe out the “we escaped China” narrative. The “diversification” that analysts celebrated was geographic relocation to a single alternative country. When that country faced its own tariffs, Crocs had no fallback. The supply chain map had different pins. The risk profile was identical.

Supply chain concentration is one of the most common hidden vulnerabilities in public companies. Most investors look at a factory map and stop there. Real risk lives in tariff codes, HTS schedules, credit agreements that reference supplier thresholds, and 10-K footnotes that use the phrase “substantially all.”

Concentration Isn’t Just Country

The Crocs example is country-level concentration: too much production in one geography. But concentration risk shows up at multiple layers, and the less visible layers are often more dangerous because they’re harder to spot in filings.

Single factory

When a 10-K says “substantially all of our [product] is manufactured at our facility in [location],” read it as: if that facility burns down, gets flooded, faces regulatory shutdown, or gets tariffed, there is no plan B. PAR Pacific’s Hawaii refinery is the only refinery in the state. Hawaiian Electric depends on it for fuel oil generation. If that single refinery goes down for extended maintenance, there is no alternative on the islands. Every electron from oil-fired generation depends on one facility controlled by one company.

Single contract manufacturer

Many consumer products companies don’t own their factories. They contract with third-party manufacturers. The 10-K discloses this, but investors often don’t ask the follow-up: how many contract manufacturers? A company using one CM for 90% of production has the same concentration risk as owning a single factory, except they have even less control over production continuity.

Single tier-2 input

This is where the less obvious risk lives. A company might assemble products in three countries using five different suppliers, and the factory map looks beautifully diversified. But if all five suppliers depend on the same upstream raw material or component, the diversification is cosmetic.

Solar is the textbook case. Companies “diversified” cell and module assembly across Vietnam, Thailand, Cambodia, and Malaysia. The factory map had pins in four countries. But industry estimates put Chinese silicon wafer production at upwards of 90% of global supply. Every one of those “diversified” assembly plants depends on the same upstream source for its critical input. The assembly is diversified. The supply chain is not.

Other tier-2 convergences I watch for: specialty chemical precursors (a handful of producers serve entire industries), semiconductor substrates, rare earth processing (China dominates both mining and processing by wide margins, per USGS data), and pharmaceutical active ingredients (India and China dominate API production for the global generic drug supply).

Single logistics chokepoint

Even a diversified supply chain converges at transit points. The Panama Canal, the Strait of Malacca, the Suez Canal, specific deep-water ports. A company with factories in five countries that all ship through one port is diversified on paper and concentrated in practice. The 2021 Suez blockage and the 2023-24 Red Sea disruptions showed how fast these chokepoints cascade into earnings misses.

When I’m mapping supply chain risk, I try to think through all four layers: country, facility, input, and logistics. The 10-K usually discloses the first two. The last two require reading trade publications, earnings call transcripts, and sometimes just thinking carefully about where the raw materials actually come from.

Where Supply Chain Risk Hides in Filings

The SEC requires companies to disclose material risks. Supply chain concentration is material. The disclosures exist. They’re just easy to miss if you don’t know where to look.

10-K risk factors: the legally required disclosures

Key phrases to search for in Item 1A:

  • “Single source” or “sole supplier”: the company depends on one supplier for a critical input. If that supplier has a problem, the company has no alternative.
  • “Substantially all”: when lawyers write “substantially all of our products are manufactured in Vietnam,” they’re telling you the concentration is near-total. That phrase survives legal review because it’s materially true.
  • “Primarily manufactured in” or “primarily sourced from”: less absolute than “substantially all” but still signals majority concentration.
  • “Limited number of suppliers”: a polite way of saying “we have two or three suppliers and switching would take years.”

One thing I find useful is searching EDGAR’s full-text search for “sole supplier” or “substantially all” across all 10-Ks in a sector. One query surfaces every company in the sector that discloses this concentration. The result is a screening list for deeper analysis.

Geographic revenue vs geographic cost

10-K segment reporting shows revenue by geography. Revenue diversification gets attention in analyst reports: “40% domestic, 30% EMEA, 30% APAC” sounds diversified. But revenue geography and cost geography are different things.

A company can earn revenue in 50 countries while sourcing 90% of its products from one. Revenue diversification masks sourcing concentration. The revenue pie chart in the earnings presentation looks healthy. The cost structure in the 10-K footnotes tells the real story.

Look for: “geographic information” or “property, plant and equipment by geography” in the financial statement footnotes. If a company reports $500M in PP&E and $480M of it is in one country, that’s concentration regardless of where the revenue comes from.

Earnings call transcripts: the accidental admissions

Executives discussing “diversification efforts” on earnings calls are making an inadvertent admission: they’re diversifying because current concentration is a problem. If the supply chain were already diversified, they wouldn’t be talking about diversifying it.

Listen for:

  • “We’re in the process of qualifying additional suppliers”: they currently depend on too few.
  • “We expect to reduce our exposure to [country] over the next 18-24 months”: they’re exposed now, and it takes 18-24 months to fix.
  • “We’ve built inventory to mitigate near-term tariff impact”: buffer strategy, which means the long-term exposure is real.

The Tariff Schedule Reality

Moving production from China to Vietnam, Cambodia, or Indonesia is geographic diversification. It is not necessarily tariff diversification. This distinction destroyed billions in market value when companies discovered it.

HTS codes follow the product, not the factory

The Harmonized Tariff Schedule classifies products by what they are, not where they’re made. Footwear has specific HTS codes. Solar cells have specific HTS codes. The tariff rate for a given HTS code can vary by country of origin, but when tariffs target a product category broadly (as reciprocal tariffs do), moving the factory doesn’t change the product classification.

Crocs moved footwear production from China to Vietnam. The footwear HTS codes stayed the same. When Vietnam-origin footwear faced its own tariff rates, the “diversification” provided zero protection.

Anti-circumvention: the retroactive trap

The US government actively investigates whether companies are routing Chinese goods through third countries to avoid tariffs. When Commerce finds circumvention, it applies duties retroactively, sometimes at punishing rates.

The solar industry learned this lesson at enormous scale. Panels manufactured in Southeast Asia using Chinese cells and wafers were investigated for circumvention. The resulting preliminary and final duty rates from Commerce Department investigations included:

  • JinkoSolar (Vietnam): 244.95%
  • Trina Solar (Thailand): 375.19%
  • Certain Cambodian producers: 3,521%

These are real numbers from specific Commerce Department anti-circumvention and anti-dumping/countervailing duty investigations. Your mileage varies by producer, product, and investigation period, but the magnitude illustrates the point: retroactive anti-circumvention duties can make imported product economically unviable overnight. Even “moderate” rates in the 50-100% range fundamentally change the cost structure of products with thin margins.

Country-of-origin rules

“Substantial transformation” determines legal origin, not final assembly location. If a product is assembled in Vietnam using components manufactured in China, the legal origin depends on whether the Vietnam assembly constitutes “substantial transformation.” Simple assembly (screwing parts together, packaging) usually doesn’t qualify. The product’s legal origin remains China, and Chinese tariff rates apply, regardless of where final assembly happened.

This matters for companies claiming supply chain diversification through assembly-only operations in low-tariff countries. The diversification may be cosmetic. The tariff exposure may be unchanged.

Inventory Buffers: The Clock Running Down

Many companies saw tariffs coming and pre-bought inventory. This is rational short-term strategy: import as much as possible at pre-tariff prices, stockpile it, and buy time to figure out long-term sourcing.

The market often treats inventory buffers as if they solve the problem. They don’t. They delay it.

How to read the buffer

Check quarterly inventory in 10-Q filings. If inventory spiked in Q4 (pre-tariff buying) and is declining in Q1, the buffer is depleting. Here’s the rough math I use:

Pull the last four 10-Qs. Look at “inventories” on the balance sheet. Compute the company’s normal inventory level (average of the prior four quarters before the spike). The excess is the spike minus the normal level. Divide the excess by the company’s quarterly cost of goods sold. That gives you buffer duration in quarters.

A concrete example from the tariffs research: a consumer goods company reported $340M in inventory in Q3 2025, up from a $220M trailing average. Excess inventory: $120M. Quarterly COGS: $180M. Buffer duration: roughly 0.67 quarters, or about two months. That’s less time than it sounds. By mid-Q1, the buffer would be depleted and tariff-rate inventory would start flowing through the income statement.

The market was pricing the stock as if the inventory build had “solved” the tariff problem. It had bought roughly eight weeks.

When the buffer runs out

Three options, all bad:

  1. Absorb the tariff: margin compression. A 25% tariff on a product with 30% gross margin turns it into a 5% margin product. For some companies this means selling below cost.
  2. Pass to customers: price increases. Works in theory. In practice, competitors who source domestically or from non-tariffed countries don’t raise prices. The tariff-exposed company loses market share.
  3. Find new suppliers: re-sourcing takes 12-18 months minimum for manufacturing. Qualifying a new factory, tooling, testing, logistics setup. This isn’t a quarter-to-quarter fix.

None of these options is priced at current levels if the market assumes the buffer is permanent. The question isn’t “do they have a buffer?” but “how many quarters does it last, and what happens when it runs out?”

Environmental Risk in the Supply Chain

Supply chain disruption doesn’t only come from tariffs and geopolitics. Environmental enforcement can shut down a critical supplier, and the disruption flows downstream to every company that depends on it.

The EPA publishes environmental compliance data for every regulated facility in the country through a system called ECHO. It’s free, it’s searchable, and almost nobody in finance uses it. I’ve found it surprisingly useful for supply chain research.

How to use EPA data for supply chain analysis

If a company’s 10-K names a key supplier, search the EPA database for that supplier’s compliance history. What you’re looking for:

  • Repeated violations: a facility cited multiple times under the same program (Clean Air Act, Clean Water Act, RCRA hazardous waste) signals systemic compliance problems, not one-off mistakes.
  • Pending enforcement actions: formal enforcement can result in consent decrees that force production cuts or facility modifications. During enforcement, output drops.
  • Significant non-compliance (SNC) status: EPA’s designation for facilities with serious, ongoing violations. SNC facilities face escalating enforcement pressure.
  • State vs federal enforcement: state environmental agencies enforce alongside EPA. Check both levels.

A consent decree forcing a sole-source supplier to reduce production by 30% for environmental remediation is a supply chain disruption for every company downstream. The 10-K of the downstream company won’t mention it. The EPA enforcement database will.

The miss that taught me to check upstream

I spent weeks building a thesis around a specialty chemicals company that looked clean on every metric I tracked: filings were unremarkable, no insider selling, no litigation. What I didn’t check was their primary raw material supplier’s EPA compliance history. That supplier was in significant non-compliance with RCRA hazardous waste regulations and had a pending consent decree. When enforcement hit and the supplier cut output, my target company issued a surprise earnings warning citing “supply disruption.” The stock dropped 18%.

The information was sitting in a free government database. I just hadn’t thought to look upstream. Now “check EPA on named suppliers” is a checklist item, not an afterthought. It takes five minutes and occasionally surfaces a risk that isn’t in any filing or analyst report.

The Hard Part Isn’t Access

Everything in this series draws on free or cheap public data: SEC EDGAR, federal court archives, EPA ECHO, FDA openFDA, tariff schedules, retail-tier options data. All have APIs. All are searchable programmatically.

The data is open. What’s hard is knowing what to look for and how to interpret what you find. A search for “sole supplier” across EDGAR returns hundreds of hits. Knowing which ones represent genuine concentration risk versus boilerplate disclosure is judgment, not automation. I use these databases every day and I’ve built tools to query them faster and cross-reference results. But the tools are just plumbing. The value is in the analytical framework: which filings matter, which docket stages represent real risk, which insider clusters are mechanical versus informed, and which supply chain disclosures are genuinely dangerous.

Where This Gets Fuzzy

Supply chain analysis has real limitations:

  • Private companies upstream. If a critical supplier is privately held, you won’t find SEC filings for them. You’re dependent on the downstream public company’s disclosures, which may be vague. “We source from a limited number of suppliers” tells you concentration exists but not how severe it is or who the suppliers are.
  • Tariff schedules change faster than supply chains. A tariff can be announced in a tweet. Moving a supply chain takes 12-24 months. Any thesis built on current tariff rates is vulnerable to policy reversals, exemptions, or escalations. I’ve had supply chain theses invalidated by a single trade negotiation headline.
  • Buffer duration estimates are rough. The 10-Q inventory numbers tell you what’s on hand, but not the mix. Some of that inventory might be pre-tariff goods; some might be raw materials that still need processing. The buffer duration math is directionally useful but not precise.
  • Companies adapt faster than you expect. Some management teams are genuinely good at re-sourcing under pressure. A company with strong supplier relationships and engineering capability can qualify a new manufacturer faster than the “12-18 months” rule of thumb suggests. I’ve been early on trades where the company found alternatives faster than I modeled.
  • Tier-2 and tier-3 concentration is often invisible. Even detailed 10-K disclosures rarely describe the supply chain beyond the first tier. A company might name its contract manufacturers but not disclose where those manufacturers source their inputs. The deeper the concentration, the harder it is to see from public data alone.

If You Only Do Three Things

  1. Search for “substantially all” and “sole supplier” in 10-K filings. These phrases are legally significant. When they appear, the concentration is real and material. One EDGAR search across a sector gives you a screening list in minutes.
  2. Estimate buffer duration from 10-Q inventory data. Excess inventory divided by quarterly COGS gives you the number of quarters before tariff-rate product hits the income statement. The market often prices the buffer as a solution rather than a delay.
  3. Think beyond country-level. Check facility concentration, tier-2 input convergence, and logistics chokepoints. A factory map with pins in five countries can still have a single point of failure if they all depend on the same upstream source or ship through the same port.