Democratic backsliding causes extinction.

Totalitarian states are incompetent. They are bad at forecasting and dealing with disasters (Caplan, 2008).16 This can be seen most clearly in the great famines of Communist China and the USSR, in which millions died (Applebaum, 2017; Becker, 1996; Dikötter, 2010; Snyder, 2010). In comparison, functioning multiparty democracies rarely, if ever, experience famines (Sen, 2010). “Established autocracies” (or “personal”/“sultanist”) are particularly bad, as there are few checks or restraints on arbitrary rule and the whims and ideology of the single individual, even from other elites (Svolik, 2012). From the inside, the “inner circle” around Mao, Stalin, and Hitler seems incredibly chaotic, with elites strongly incentivized to conceal information and encouraged by the autocrat to squabble and feud—so they are divided (Conquest, 1992; Kershaw, 2008; Zhang & Halliday, 2006). If totalitarian states are worse at addressing social, environmental, and technological problems, then a world dominated by them would likely be worse at responding to risks of collapse and extinction.

A world dominated by totalitarian states is more likely to have major wars. States with near-universal adult suffrage rarely (if ever) go to war with one another (Barnhart et al., 2020), so a world dominated by democracies has fewer wars. Miscalculation might be a particular problem for totalitarian states due to personalization and disincentives for accurate information, leading to well-known strategic disasters such as Hitler and Stalin’s blunders in World War II (Bialer, 1970; Noakes & Pridham, 2001), or at a smaller level, Saddam Hussein’s rejection of diplomacy (Atkinson, 1993). War makes collapse and extinction more likely, by raising the chance of weapons of mass destruction being used.

Linked to this, totalitarian states are less cooperative than democratic states. While cooperation is possible (Ginsburg, 2020), their internal norms are characterized by paranoia and treachery, and their lack of transparency limits their ability to credibly commit to agreements. This is bad for all risks that require cooperation such as pandemics or climate change (Tomasik, 2015).

Finally, continued social and scientific progress is likely to reduce risks of collapse and extinction. Social progress could reduce global inequality and other risk factors. Scientific progress could help address natural risks and climate change (Sandberg, 2018), differentially increase defensive rather than offensive power (Garfinkel & Dafoe, 2019), and solve safety challenges in AI or biotechnology (Russell, 2019). However, as we will now discuss totalitarian states would likely inhibit social progress.