Non-Normal Interactions Create Socio-Economic Bubbles

In social networks, bursts of activity often result from the imitative behavior between interacting agents. The Ising model, along with its variants in the social sciences, serves as a foundational framework to explain these phenomena through its critical properties. We propose an alternative generic mechanism for the emergence of collective exuberance within a broad class of agent-based models. We show that our model does not require the fine-tuning to a critical point, as is commonly done to explain bursts of activity using the Ising model and its variants. Instead, our approach hinges on the intrinsic non-symmetric and hierarchical organization of socio-economic networks. These non-normal networks exhibit transient and unsustainable surges in herd behavior across a wide range of control parameters even in the subcritical regime, thereby eliminating the need for the - arguably artificial - fine-tuning proximity to a critical point. To empirically validate our framework, we examine the behavior of meme stocks and establish a direct linkage between the size of financial bubbles and the degree of non-normality in the network, as quantified by the Kreiss constant. Our proposed mechanism presents an alternative that is more general than prevailing conceptions of instabilities in diverse social systems.