A model puts the possible 2024 omission between 0.60% and 1.60% of GDP, depending on which digital assets are counted.
The researchers adapted a model of data-driven production to China and used asset disclosures from 5,120 listed companies to estimate digital investment. They calculated a 2024 GDP adjustment of 2.16 trillion RMB under a broad definition that includes telecommunications networks, compared with 0.82 trillion RMB under a narrower definition that excludes them.
The estimates are not statistical bounds. They depend on assumptions about how digital investment maps to the value of data, the assets included and the extrapolation from listed to unlisted firms. For listed firms alone, the broad-definition adjustment was 0.91% of GDP before extrapolation.
What the estimates include
Within the study’s model, data generated as a by-product of transactions is output that national accounts do not price, so measured GDP understates output from this source. For 2024, the estimated adjustment was 1.60% of GDP, or 2.16 trillion RMB, under the broad asset boundary, and 0.60%, or 0.82 trillion RMB, under the narrow boundary.
The broad estimate for listed firms alone was 0.91% of GDP. Its calibrated scenario range was 0.77% to 4.07%, reflecting different model calibrations rather than sampling uncertainty. The study also found that a keyword classifier initially missed much of the value of digital fixed assets because it failed to recognize one written variant of the word for communication; the researchers corrected the series to include telecommunications and network assets in the broad measure.
Evidence and caveats
The study combines a data-economy model with a bottom-up measure of digital investment drawn from itemized asset disclosures by 5,120 listed firms. Its estimates are calibrated scenarios, not statistically estimated intervals. The model does not determine the correct asset boundary, and the value of data depends on a normalization that Chinese data do not identify.
The broad whole-business estimate also extrapolates beyond listed firms using an assumption about unlisted companies that the researchers cannot verify. The study does not measure the accuracy of the GDP figures already recorded and is not an estimate of the net bias in China’s GDP. Asset disclosures report closing balances rather than investment, and the classifier audit found important errors, including missed telecommunications assets and false positives involving some electrical infrastructure.
// Source
Digital Economy and Sustainable Development · 2026 · DOI: 10.1007/s44265-026-00087-8
Authors: Wang Weijia
Institutions: Xi'an Jiaotong University
