BIS: AI and productivity in European firms
On 23 January 2026 the Bank for International Settlements published Working Paper 1325: in data on more than 12,000 EU and US firms, AI adoption raises the level of labour productivity by 4%, with no adverse effect on employment.
Why it matters
There is now an estimate of AI's causal effect on firms outside the US, not a forecast or a poll: the authors instrument EU adoption with the adoption rates of US peers. An editorial assessment; as the authors say, the finding is about the short run.
What the page says. Authors: Iñaki Aldasoro, Leonardo Gambacorta, Rozalia Pal, Debora Revoltella, Christoph Weiss, Marcin Wolski. Data: matched EIBIS-ORBIS, more than 12,000 non-financial firms in the EU and the US; EU firms' AI adoption is instrumented by the adoption rates of US peers. Results: the level of labour productivity is 4% higher; the gain comes through capital deepening, not through job losses; adopting firms pay higher wages and are more innovative; the benefits concentrate in medium and large firms; complementary investment in software, data and workforce training matters. The authors themselves say longer-term effects remain uncertain. What the record does not claim. The page was read, the 36-page PDF was not, so the record does not describe the method, the sample construction or the robustness checks. The outside report that named this paper gave a finding about an 'uneven distribution' with no figure; the 4% is taken from the page. The views are the authors', not the BIS's, as the page itself says.