Small, Numerous, and Under-Scaled: What US Export Data Reveals About MSME Global Reach

What the export ledger actually shows

The headline is genuinely striking, and it is government data, not estimate. In its 2024 Small Business Profile, the U.S. SBA Office of Advocacy reports that of all identified US exporting firms in 2023, 270,014 were small, which is 97.2 percent of exporters by count. On the value side those same firms accounted for 588.4 billion dollars in exports, which is 33.0 percent of the identified-firm total.

Set those two numbers next to each other and the tension is obvious. A population that is 97 percent of the exporter count produces one third of the exported value. The average large exporter therefore moves a far larger volume of goods across borders than the average small one, by roughly an order of magnitude once you divide value by firm count in each group. The small-firm export story is a story about how many participate, not about how much they carry.

This is not a complaint about unfairness and it is not evidence that small firms fail at exporting. Both numbers are large. Hundreds of thousands of small US firms do reach foreign customers, and more than half a trillion dollars is a substantial share of national trade. The interesting question is structural: why does participation cluster at the small end while value concentrates at the top, and is that a durable feature of trade or a temporary artifact.

A theory that predicted the shape: Melitz and the productivity threshold

The pattern in the SBA ledger is not a surprise to trade economists. In 2003, Marc Melitz published a model in Econometrica that has organized the field ever since. Its central move was to stop treating an industry as a set of identical firms and to treat firms as heterogeneous, differing in productivity, and then ask which of them would choose to export.

His answer was a threshold. Selling abroad carries fixed costs that selling at home does not: foreign market research, distribution and compliance, adapting a product or service to another market, and the sunk investment of establishing a presence a buyer will trust. A firm only clears those fixed costs if it is productive enough to earn back the outlay on the export sales it can win. Below a critical productivity level, exporting loses money and the rational firm stays domestic. Above it, exporting pays. The model therefore predicts self-selection: exporters are not a random sample of firms, they are the systematically more productive ones.

Why this is a scale story, not a count story

It is tempting to read "97 percent of exporters are small" as a triumph of small-business globalization, and to read "only 33 percent of value" as a failure. Both readings mistake the unit. Count and value answer different questions. Count tells you how widely the ability to export is distributed. Value tells you how much economic weight sits behind each participant.

Access to cross-border trade is broad while capacity within it is concentrated. A small firm can be a real exporter and still be under-scaled relative to the firms that dominate the value line, not because it is doing anything wrong but because per-firm export scale is exactly what the productivity threshold rations. Melitz's model builds this in: firms sort continuously by productivity, so even among successful exporters, output and export intensity rise steeply with the firm's position above the threshold.

For a small firm weighing whether to chase foreign demand, that reframing is useful rather than discouraging. The relevant question is not "can firms my size export," because the data says hundreds of thousands do. It is "where do I sit relative to the fixed costs of the specific border I want to cross," because that position, not size alone, is what the theory says determines whether the attempt pays.

The threshold re-forms at every new border

Melitz wrote about physical goods trade, but the logic is about fixed costs of entry, and fixed costs appear wherever a firm tries to reach a market it does not already serve. That is why the same selection dynamic keeps reappearing as the definition of a border changes.

The policy layer has been working to lower those costs. The WTO frames its e-commerce moratorium, which keeps digital transmissions tariff-free, and its Trade Facilitation Agreement provisions on e-signatures, paperless trading, and e-invoicing, as explicit small-business enablers. Yet the WTO's own handbook for policymakers names the constraints that remain: small firms "shoulder disproportionate burdens" from non-digital customs procedures, lack familiarity with e-commerce mechanics, and face persistent transport and logistics gaps. The legal door is opening faster than the operational one. In Melitz's terms, the fixed cost of the border is falling on paper while much of it persists in practice, so the threshold moves but does not disappear.

The newest border is machine-mediated distribution. As buyers increasingly discover and compare suppliers through search engines and generative answer engines rather than through a salesperson or a trade directory, being legible to those systems becomes a new fixed cost of reach, a structured, crawlable, corroborated presence that a firm either invests in or does not. That is a threshold like any other. It rations which firms are found by an out-of-market buyer, and it is one a small firm can actually choose to clear.

The domestic mirror raises the stakes

Cross-border reach does not happen in a vacuum; it is a decision made by firms that are already operating on thin margins at home. The same SBA data that reports the export split also documents how fragile the small-firm base is. There are 34.8 million small businesses in the US, accounting for 45.9 percent of private-sector employment and 43.5 percent of GDP, so the population is economically central. It is also volatile: between March 2023 and March 2024, small businesses accounted for 982,940 closings against roughly 1.1 million openings, a near one-to-one churn ratio that has no equivalent among large firms.

That fragility changes how the productivity threshold should be read. For a firm with little margin for error, a fixed-cost investment that does not clear the threshold is not a missed upside, it is capital and attention it could not afford to lose. The selection logic is therefore prescriptive as well as descriptive: it argues for honestly assessing where a firm sits relative to the fixed costs of a given expansion before committing, rather than treating "go global" or "go digital" as a costless default.

What this predicts, and where it stops

A good theory is worth more than a single statistic because it generalizes, and this closing section states plainly what Melitz's model does and does not license. What it predicts reliably: exporters will be a productivity-selected minority of value even when they are a majority of the count; lowering the fixed cost of a border widens participation without necessarily equalizing per-firm scale; and each new distribution layer that carries its own entry cost will re-sort firms against a fresh threshold. The 2023 export split is one clean confirmation of the first prediction.

What the model does not do is promise that any individual firm will clear a given threshold, or that clearing it guarantees a return. It is a model of averages and incentives, not a forecast for one business. The emerging question, whether digital and agentic distribution genuinely lowers the threshold for the smallest firms or simply relocates it, is not yet settled by the data, and it should be held as an open question rather than an assumed win. The value of the framework is that it tells a firm what to measure before it spends: its own position relative to the fixed costs of the reach it wants.