No to a Switzerland
of 10 Million
A pre-referendum forecast built from Foresenta's synthetic Swiss population. We combined public demographic data, the media environment, and individual-level modeling to project the result before the June 14, 2026 referendum.
What this referendum was about
Switzerland's "No to a Switzerland with 10 million" initiative, also known as the Sustainability Initiative, asked voters whether the country should cap its permanent resident population below 10 million by 2050. If the population crossed 9.5 million before then, the federal government would have been required to take corrective measures, particularly in areas such as asylum and family reunification. Should the threshold of 10 million be breached, the initiative could have forced Switzerland to renegotiate or terminate international agreements contributing to growth, including the agreement on free movement of persons with the European Union.
The proposal became one of the country's most debated votes because it touched several sensitive issues at once: immigration, pressure on housing and infrastructure, labor-market needs, and Switzerland's relationship with Europe. Critics and supporters both saw it as a high-stakes decision with consequences well beyond demographics.
How we built the forecast
Three layers that work together — a population that mirrors Switzerland, the information environment around the vote, and individual-level reasoning aggregated into a national outcome.
Synthetic population
We assembled a statistically representative synthetic Swiss electorate from multiple publicly available demographic sources — combining age, education, income profile, geography, language region, and household context to mirror the structure of the real population.
Media environment
We monitored and structured referendum coverage across all language regions, then modeled media exposure at the individual level — matching each synthetic voter to the outlets and channels they were realistically likely to consume. Message framing, salience, and visibility were never applied as a uniform signal across the population; instead, their effect on support varied person by person, according to each individual's plausible information diet.
Individual vote simulation
We estimated vote propensity at the individual level inside the synthetic population, then aggregated those decisions upward to produce canton-level and national results — preserving the heterogeneity that polling averages erase.
We do not model poll responses. We model people, their context, and the information that shapes their choices.
This forecast is a snapshot made in advance, based on the information environment available on June 1, 2026. Events, news, or campaign developments after publication can shift the final outcome — which is precisely why the prediction stands on the date it was issued, not in hindsight.
Want a forecast like this for your market?
Foresenta builds synthetic populations and simulates how they will react to products, campaigns, prices, and information environments — before you commit.
Book a Call →