🏛️
WebGL2 is not available in this browser.
Try Chrome or Firefox. Some Safari versions require enabling WebGL in Develop settings.
Readout
Adjust "Cost of survival" to see selection at work.
0% firms
exited
— mean φ
survivors
— mean φ
all firms
φ = firm productivity (tower height). Heights never change.
Survivors were always the tall towers — raising the water hid the short ones, it didn’t improve anyone.
Cost of survival 30%
The survival cutoff: firms below the waterline exit.
Firm heterogeneity (σ) 0.60
How unequal firms are: low = everyone similar, high = a few giants among many small firms.
Dr.Tremendous
Atmosphere
Storm intensity 50%
Visual only — does not change which firms survive. Firms Render quality
Lower this if the scene feels slow.

Each tower is a firm. Its height represents productivity (φ) — drawn once from a lognormal distribution and never changed. The waterline is the zero-profit survival cutoff: firms whose productivity falls below it cannot cover their fixed costs and exit.

As "Cost of survival" rises, the cutoff rises with it and shorter towers go under. Notice that no tower grows — but the average height above water increases. This is selection, not improvement.

This survivorship effect is the engine of Hopenhayn (1992) and Melitz (2003): tighter competition raises aggregate productivity by culling low-φ firms, even though no individual firm improves.

"If you only measure the buildings above water, you'll conclude the flood made firms better. It didn't — it made the sample better."