Industrial production rose by 9.4% yoy, and this result cannot be attributed to any particular sector – production in manufacturing alone accelerated from 0.0% to 9.1% yoy, with as many as eight industries recording double-digit growth. March is a unique month – each year the semi-annual order cycle marks (alongside September)  the seasonal peak of production. In our opinion, the monthly increase in production in March is also a very cyclical variable – during periods of recovery/expansion, production growth is greater than would be indicated, for example, by the number of working days. Nonetheless, we had assumed such an effect in our forecast (+5.8% y/y). The actual result (in manufacturing alone: +19.7% month on month) is the fourth highest March increase in history, but the other three benefited from a more favourable arrangement of working days (+3 rather than +2, as in March 2026).

Manufacturing output and working Days (% mom)

Source: GUS, Pekao Analizy

Seasonal adjustment should therefore deliver an impressive result, and in fact, that’s exactly what we got – after removing seasonal and calendar effects, production increased by 7% mom, which is the highest single-month increase since June 2020. In yoy terms, this is the highest growth rate since September 2022. In our view, such a large increase in production in March of this year should be attributed to both cyclical (rebuilding domestic and partly foreign demand + cyclical nature of March production) and one-off factors. There was no shortage of the latter in March: construction was revived after a harsh winter (see the second part of this comment), early Easter probably boosted food and beverage production in March (relative to April), and, most importantly, the outbreak of war in the Persian Gulf may have had an unusual impact on industry. Concerns about the availability of components and raw materials (fertilizers, intermediates inputs for plastics, certain metals) probably prompted companies to stockpile production. A significant part of March’s industrial production surge therefore does not have a lasting character.

Industrial production (February 2020 = 100%)

Source: GUS, Pekao Research

In construction, the main question concerns the return to normality after a very weak start to the year. As a reminder – low temperatures and snow cover greatly restricted activity on building sites in January and February – on average, production in these two months fell by 13.3% yoy. In March, all obstacles were gone and production returned to positive territory, albeit only just. Last month’s result was +0.4% yoy, halfway between our more optimistic forecast and the consensus. The fluctuations in production at the turn of the year were therefore substantial.

What should we expect in the coming months? Fortunately, the data provides a precedent – the start of 2010 was also very frosty and brought comparable declines in production (albeit from a lower base – December 2025 was much better than December 2009). So far, the thaw in construction is also comparable to that from 16 years ago. 2010 teaches us that such a winter affects annual construction results, even if the sector returns in late spring/early summer to the level of activity it would have had without the winter freeze. Consequently, certain delays in investments are very likely and the cut to our forecast made a month ago was justified.

Cumulative change in construction output since December of the previous year (based on s.a. data)

Source: GUS, Pekao Research

In summary, taking into account the number of one-off factors influencing industrial and construction production in March this year, we are not inclined to revise our GDP growth forecasts for 2026 in response to this data. After the retail sales data scheduled for Thursday (where there is considerable potential for a positive surprise), we will present a refined GDP forecast for the first quarter. For now, our working assumption is that GDP slowed from 4.1% to 3.8% yoy. This is a considerably smaller slowdown than we had expected prior to today’s publication, but it is probably linked to lost output in the second quarter due to the aforementioned one-off factors.