This paper asks how heavy state intervention during Japan’s mobilization years changed what stock prices signaled. The authors focus on the large family business groups called zaibatsu. These groups had internal banks and political access, which could give them better access to credit, materials, and government contracts. At the same time a formal stock exchange kept trading through 1930–1943, so prices might still reflect news even as policy rules shifted who got resources.

The researchers build a simple four-portfolio pricing model that treats zaibatsu affiliation as a key characteristic. In the model, being part of a zaibatsu affects both expected payoffs and how market valuations translate into real economic scale because zaibatsu firms faced smaller ‘‘financing wedges’’—that is, lower frictions in getting funds. They then create daily, capitalization-weighted indices and four benchmark portfolios from a two-by-two sorting of firms by zaibatsu membership and military orientation.

To test the model they use a high-frequency event-study design built on a capital asset pricing model with two important relaxations. First, the model allows for autoregressive residuals, which absorb persistent patterns in daily returns. Second, it allows for stochastic volatility, which captures time-varying uncertainty and volatility clustering. This CAPM-AR(p)-SV setup is designed to avoid fragile assumptions commonly used in constant-parameter event studies of wartime daily data.

Their results show several consistent patterns. The model helps explain why market capitalization became concentrated in a few firms, why abnormal returns around events were segmented across the four portfolios, and why cumulative returns could keep adjusting even when the immediate day‑0 return was small. They also find that zaibatsu portfolios were often insulated from broad ‘‘regime’’ shocks but showed concentrated reactions when events revised rents tied to state allocation or group-level continuation values.