A core analytical lesson of this worldview is to distinguish between what people say and what they do — assuming what people do reflects what they want, while what they say may just be what they think you'd like to hear.

Cronies are relatively eager to push for war because they expect to expand their booty and their heads are less likely to roll than the leader's, so they pressure the leader toward war more often than the leader might prefer — and the leader must respond to keep them loyal.

Foreign aid is given in exchange for policy concessions that benefit the donor government's democratic constituents, not the recipient's citizens — because if the recipient would follow those policies anyway, there'd be no need to pay them — making aid harmful to recipient citizens while helping autocrats supplement bribes.

Aid programs have actually been successful, not failures — but what they succeed at is keeping recipient policies aligned with donor desires and keeping autocrats in power, not the publicly stated goals like reducing poverty.

Hitler did not fully mobilize the German steel industry early in WWII (unlike WWI) because he feared losing the loyalty of comfortable middle-class supporters if he stopped producing consumer goods — driving his need for slave labor and the Ukraine — and only cranked up the war effort late when defeat clearly meant being deposed, while the Allies mobilized fully from day one.

Autocrats start wars for two reasons: they have policy beliefs (sometimes that conquering neighbors will expand the extractable resource base via land grabs or tribute), and because their rent-seeking inefficient economies run down over time, creating a shortage of resources to bribe cronies — exemplified by Saddam Hussein's 1991 move on Kuwait's oil fields.

The larger the winning coalition, the more efficient it is for leaders to rule by producing public goods rather than private rewards, because there are too many cronies to bribe individually — coalition size acts as the implicit price of private goods, which de Tocqueville observed.

Foreign aid has the same effect as the natural-resource curse because, like oil in the ground, it is money the leader can tap without raising taxes on labor — a 'free good' that sustains bad leaders in office and oppresses the public.

Democracies cut aid to elected governments whose policies they dislike (e.g., Hamas in the Palestinian Authority) because voters prefer cutting money to inducing better policies over funding abhorrent ones — making it anti-democratic for a re-election-seeking leader to fund foreign agendas their constituents reject.

The World Bank is itself a small-coalition political entity whose loan decisions are made by representatives of guaranteed-risk banks and recipient-country governments who profit from continued flows — so it does what is incentive-compatible for its constituents rather than advancing welfare, and exists fundamentally to make loans.

Democratic leaders are highly selective about wars, choosing fights they believe they will almost certainly win, because losing a war nearly guarantees they are deposed — and when a war goes badly they increase effort (double up) because winning is the public good that matters.

Autocrats can be more adventurous in war because they are judged on delivering goodies to cronies, not policy performance — so they fight in a broader range of circumstances, are less likely to add resources when losing if it would cost their cronies, and survive defeat as long as they keep paying cronies and aren't removed by a victorious foreign rival.

Dictators are never truly able to do whatever they want; even Hitler had to curry favor with a group of relatively comfortable middle-class supporters he viewed as necessary, contradicting the cartoon image of an all-powerful dictator able to wage total war freely.

Once a free press and free assembly are installed, they are hard to reverse without mass slaughter, which requires soldiers willing to fire on mass movements — and they often are not, as seen in the 1991 Soviet coup and the Velvet Revolutions, especially Poland.

The 'selectorate' is the set of people who have a meaningful say in choosing leaders, and it varies greatly in size — the electorate in a democracy, the generals and senior civil servants in a military junta, or roughly 300-400 essential people in North Korea.

The 'winning coalition' is the subset of the selectorate whose support is essential to keep a leader in office, in exchange for which they receive benefits — corruption opportunities, rent-seeking, and bribes.

When the selectorate is very large relative to the winning coalition, the incumbent need not spend as much to keep supporters loyal, because a defector has a low probability of being essential to a rival — leaving more discretion for the leader to keep resources for himself or pursue clever or dumb policies.

The implicit tax rate is higher when the coalition is small: in North Korea, with per-capita income around $600, the average tax rate is about 10%, whereas in the US a person with comparable income pays nothing — so the comparable-income person in North Korea pays a much higher rate to feed the leadership's rent-seeking demands.

Foreign aid not only discourages leaders from reforms but makes it easier for them to purge rivals and tighten their grip — when facing a credible threat, the rational response is to shrink the coalition and cut public goods rather than democratize.

Ahmadinejad's nuclear weapons program is a skillful political strategy: it resonates with young professionals' pride amid a declining economy, restores Iran's primacy as the exporter of Shia fundamentalism against the rising al-Qaeda, creates bargaining leverage, and secures his hold on power — not the act of a madman.