There’s a certain difficulty, almost a stubbornness, in tracing the early days of wind turbine manufacturing—especially when relying on legacy classification systems. ISIC 2825, the “Manufacture of machinery for mining, quarrying and construction,” is a category that seems, at first, curiously removed from the image of sleek wind farms across Northern Europe. Yet in 2005, this code is where most of the pioneering wind turbine component manufacturers could be found, at least as far as official records were concerned.
For an economist, this mismatch is both frustrating and revealing. The ISIC system, by necessity, groups together a wide spectrum of machinery production, from cement mixers to drilling rigs. Wind technology, still something of a newcomer at that moment, simply did not have a clear niche. So any attempt to map the sector begins with an acceptance of this ambiguity. That’s the first lesson: the data will not cooperate entirely.
Still, there are ways to tease out the story. The first step is to extract all firm registrations and output data tagged under ISIC 2825 for the year 2005, in the country or region of interest. In the context of early European Union green directives—such as the 2001 and 2003 renewables targets—most national statistical agencies maintain business registries or output reports with this coding. The resulting list is, inevitably, a mix. Some firms are legacy heavy-equipment makers, others are newcomers, and a handful are wind-specific manufacturers, sometimes operating under more general company names.
Filtering this list is both art and science. Company descriptions, archived websites, and trade show participation records can reveal which entities actually produced wind turbine components: blades, gearboxes, towers, or full nacelles. In many cases, it’s necessary to look for indirect evidence—a mention of “wind energy,” “renewables,” or a contract with one of the known wind farm developers. Patents can also provide valuable signals. A surge in filings for rotor technology or grid integration components, mapped against ISIC 2825 registrants, often helps isolate the most relevant producers.
The next step involves output data. Annual reports, export records, and—where they exist—sectoral studies published by trade associations give a sense of scale. Not every firm discloses the breakdown between wind and other machinery lines, but total machinery output, combined with qualitative data (such as public announcements of new wind-related production lines or investments), can be triangulated into a plausible estimate.
Correlating this manufacturing output with the timeline of EU green directives is where the analysis gets interesting, and slightly messy. The directives themselves—the push for renewable energy quotas, grid access reforms, subsidy schemes—rolled out over several years, not in one clean sweep. The policy announcements may predate actual production increases by a year or more. Some firms responded quickly, scaling up wind component lines almost in anticipation. Others lagged, either waiting for regulatory certainty or for downstream demand to materialize.
A practical way to approach the correlation is to align firm-level output data with key policy milestones. For example, a sudden uptick in wind gearbox exports in Denmark or Germany might follow closely after new feed-in tariffs or EU-level targets were formalized. But these relationships are rarely perfect. Factors such as raw material prices, local planning approval bottlenecks, or even technology transfer agreements with non-EU partners can muddy the picture.
A more nuanced layer comes from interviews and trade press archives. Company managers, at least those willing to talk for the record, sometimes describe their own timelines in relation to policy—admitting that a specific directive triggered a retooling of their production lines or justified new investment in blade manufacturing. While such anecdotes can’t substitute for systematic data, they do help explain sudden shifts that may otherwise appear arbitrary.
Documentation is key throughout the process. Recording how firms were filtered, which data sources were prioritized, and any decisions about ambiguous cases ensures that later researchers—or policymakers—can revisit and test the underlying logic. The boundary between “true” wind turbine manufacturers and adjacent machinery producers will always be somewhat porous, especially during these formative years. A degree of uncertainty is inherent, and, perhaps, even instructive.
ISIC 2825 may not have been designed with renewable energy in mind, but by drawing on a combination of registry data, output metrics, and contextual cues, it’s possible to build a convincing map of the sector’s early growth. The link between policy and production, while never perfectly linear, is visible enough to support thoughtful analysis and, with any luck, more targeted policymaking as wind energy continues its uneven ascent.