Like Lomarandil says, society agreed this appropriate. Here society means professional engineering societies, insurance agency's, public officials, financeers, contractors, etc have attempted to balance risk and economy. Using statistical analysis they 'agreed' upon a standard convention for return intervals for the various transient loadings. Also worth noting that design methods set forth by the building code (LRFD, ASCE 7) for example, use adjustments on the design side to scale the wind forces to different levels of return period for different risk category buildings. The levels of wind pressure (for example) that the code supplies are statistically balanced to match those pre-determined return intervals.
Also extending on Lomarandil's point, it is important for owner's to understand the code level design especially for seismic. Most people think if a building is engineered then it is earthquake proof, I have seen some engineers require their client sign an acknowledgement that they understand that the code level of seismic design implies damage at the design level earthquake, maybe so much damage that the building is torn down.
Just digging into a bit more, we can think of the consequences of altering the return interval up or down a bit.
-If you design for a 20 yr return interval level of wind, it would be a lower value than the 50 year. And the risk that the design wind is exceeding during the life of the building is increased.
-On the flip side you design for 100 yr return wind, there is less chance that that level wind event would occur during the lifespan of the building, this means perhaps a lot of the value of structure is being wasted over the life of the building and 'society' has determined that the additional cost is not worth the extra reduction in risk.
the ASCE 7 commentary, and any reading on LRFD design methods and risk factors would shed further light on this topic.