California is actually a cautionary tale IMHO of the perils of a referendum or "votation" system. There's not necessarily an intrinsic problem with the system, but its effectiveness often hinges on how the questions are asked and framed.
For instance, suppose the following: a majority of voters support Project A and a majority of voters support Project B, but Projects A and B are expensive and a majority of voters would oppose spending money on both Projects A and B.
If you do separate, parallel yes-or-no referendums on both A and B (as frequently happens in California), it is likely that both would pass, resulting in a sub-optimal outcome as a sub-majority wanted both.
With a smaller group of legislators, you can do things like non-binding votes to allow proponents to assess support and modify or withdraw their proposals accordingly (if A is more popular than B, then B can withdraw in favor of A) or sequential votes (if A has already passed, then people will vote no on B). The higher number of stakeholders involved with a referendum and the general cost of running an election make this less feasible with the direct democracy system.