The dictator's information problem

Democracies learn about failures through elections, a free press, courts and organised interests. A dictator who suppresses these also suppresses the signals they carry. Subjects who fear punishment hide their true views, officials report what superiors want to hear, and the ruler cannot tell whether apparent loyalty is real. The more the regime represses, the less it knows.

Ronald Wintrobe (1990, 1998) named this the 'dictator's dilemma': because repression makes people afraid to express disloyalty, the dictator cannot know how much support he has, and so may overspend on both repression and buying loyalty. Milan Svolik extends the problem to the elite, where secrecy makes power-sharing deals hard to monitor. Later research examines the tools regimes use to recover information without losing control:

• Selective censorship: Gary King, Jennifer Pan and Margaret Roberts (2013) found that China's online censors remove posts with collective-action potential while allowing much criticism of officials and policies, which can serve as a source of information.
• Controlled watchdog journalism: Peter Lorentzen (2014) models a regime allowing investigative reporting on lower-level officials, adjusted to the level of social tension.
• Tolerated protest: Lorentzen (2013) argues that tolerating small, economic protests can reveal discontented groups and misbehaving local officials.
• Petitions, inspections and internal reports: Martin Dimitrov (2023) shows how communist regimes in Europe and China built institutions to gather information on public opinion, including citizens' complaints and internal assessments.
• Resource rents: Egorov, Guriev and Sonin (2009) find that resource-rich dictatorships have less free media, because they depend less on bureaucratic performance.

• China's Great Leap Forward (1958–61): local officials inflated harvest reports, procurement was set from false figures, and critics such as Marshal Peng Dehuai were purged in 1959, while a famine killed tens of millions.
• The Soviet economy's reliance on distorted plan-fulfilment reports, a problem that projects such as the OGAS computer network tried to solve technically.
• China's early handling of the outbreak in Wuhan in December 2019 to January 2020, when doctors who warned colleagues were reprimanded by police.

Some scholars argue that the problem is overstated: modern autocracies run opinion surveys, monitor social media and analyse big data, and may know more about their citizens than democratic governments do. Others reply that technical surveillance captures behaviour, not sincere preferences, and does not solve the problem of subordinates misreporting to superiors. Amartya Sen's argument that famines do not occur in functioning democracies with a free press is a related but contested claim.

The information problem is now a central theme in the study of autocracy, and empirical work shows that regimes actively design partial channels of feedback. Whether these channels can substitute for free elections and media in catching large policy errors is unresolved.