中文摘要
我们首次对公共税务披露系统中的实际搜索行为进行了研究，分析了2014年和2015年在挪威发生的约一百万次搜索。我们刻画了这些搜索所构成的社会网络，包括其同质性程度与互惠性，以及目标人物与搜索者的人口统计特征。约四分之一的搜索发生在可识别的家庭与就业网络内。大多数搜索者以与自己相似的人为目标——即网络术语中的同质性——但年轻、低收入的搜索者也会以年长、成功人士和名人为目标。基于搜索相对于报税时序的因果研究设计，并未发现证据表明目标人物在被发现成为搜索目标后会随后增加其申报收入。证据表明社会比较驱动了绝大部分搜索，而非税务合规。然而，即使在均衡状态下出于合规动机的搜索很罕见，公开披露仍可能威慑逃税。
Abstract
We conduct the first-ever study of actual searches done in a public tax disclosure system, analyzing about one million searches done in 2014 and 2015 in Norway. We characterize the social network these searches comprise, including its degree of homophily and reciprocation, and the demographics of targets and searchers. About one-fourth of searches occur within identifiable household and employment networks. Most searchers target people similar to themselves—homophily in network parlance—but young, low-income searchers also target older, successful people and celebrities. A causal research design based on the timing of searches relative to tax filing uncovers no evidence that, upon discovering they were targeted, targets subsequently increase their reported income. The evidence suggests that social comparisons motivate the bulk of searches rather than tax compliance. However, public disclosure may deter evasion even when compliance-motivated searches are rare in equilibrium.