The surge's reduction in deaths is not a measure of success because more troops naturally reduce violence at a moment in time; what matters is long-term sustainable cooperation and binding agreements, which a temporary troop increase does not create.

Between 1898 and the early 1920s the United States militarily occupied Cuba three times (1898-1902, 1906-1909, 1917-1920) each time to overcome instability and create political institutions, yet these efforts ultimately failed, leading to Batista and then Castro and a long period of illiberal rule.

U.S. interventions throughout Latin America and Cuba installed regimes friendly to U.S. political and business interests even when those regimes were illiberal, creating a tension between the rhetoric of spreading liberty and self-determination and the reality of picking winners at the expense of the country's citizens.

People respond to incentives: raising the benefit of a behavior produces more of it, raising the cost produces less; therefore reconstruction is fundamentally about whether occupiers can create rules that give citizens the incentive to pursue ends aligned with liberal democracy, and when those incentives are absent the reconstructed institutions unravel.

In Iraq the historical animosities and centuries-old conflicts among groups and subgroups (Kurds, Sunni, Shia and their subfactions) raise the transaction and bargaining costs of forming a cooperative liberal-democratic settlement, because parties do not trust each other and have violently coerced one another in the past.

Even if Iraqi groups could move toward a larger economic pie under liberal democracy, the uncertainty over each group's future share, combined with the need to set dividing rules in advance, causes fighting that leads parties to forgo the gains they could otherwise achieve—a coordination/bargaining failure.

A credible commitment problem undermines reconstruction: even if rival groups could strike an agreement, they have no reason to believe it is binding once the occupier (the third-party enforcer) eventually exits, so the absence of a credible long-term enforcer discourages cooperative settlement.

Occupiers face an inherent bind: staying long enough to credibly enforce contracts contradicts the rhetoric of liberation and self-determination, because the longer the U.S. stays and intervenes in political, economic and social outcomes, the more indigenous citizens view it as serving U.S. interests rather than as a benevolent liberator.

Insurgents and terrorists in Iraq are motivated significantly by the perception that they will not get a fair share of power or pie in the new Iraq; they want the U.S. gone not because the U.S. harms citizens but so they can have a more dominant role than they would otherwise have, making their incentives destructive to cooperative reconstruction.

There is a systematic disconnect in perceptions: the U.S. views itself as a benevolent actor bringing good ends (private property, wealth, representative government), but Iraqis and neighboring populations often perceive U.S. actions as bad or self-serving, and this perception gap produces negative outcomes regardless of actual U.S. intentions.

When violence is available, the social payoff of cooperating with the occupier flips: cooperation can be rewarded as a badge of honor or punished as a badge of shame through social norms, cultural responses, or actual murder of collaborators, which alters citizens' incentives dramatically.

A liberal democracy of the form the West wants will not emerge in Iraq in the near future; the best achievable outcome is some kind of stability once the relative power relations between groups become clear, and that cooperative equilibrium is unlikely to be one the U.S. would view as good.

The U.S. will likely remain involved in Iraq in some form for centuries—not continuously, but with recurring re-entry—just as it repeatedly returned to Cuba, because each intervention solves a problem that creates new problems requiring subsequent interventions.

Policymakers face a fundamental knowledge problem: what looks like a problem today is not a problem tomorrow, and interventions to fix problems create new sets of problems over years and decades because of limited knowledge and unintended consequences—illustrated by the U.S. once arming Saddam against Iran and funding bin Laden against the Soviets.

American politicians face conflicting public-choice incentives in Iraq: they will have a hard time exiting because of the political cost of what happens right away, and also a hard time staying because of ongoing political costs, producing a seesaw in-and-out pattern of involvement.

Japan and Germany succeeded as reconstructions partly because they were relatively developed countries with a strong national identity—citizens identified as Japanese or German rather than primarily with internal subgroups—removing the inter-group conflict that plagues Iraq.

Japan and Germany were international wars between clear nation-states with a clear winner and an official surrender, so the U.S. entered unambiguously as occupier-in-charge with no insurgency, whereas in Iraq the rhetoric of liberation rather than conquest created ambiguity that fostered resistance.

Reconstruction's success depends on culture and informal institutions (human capital) far more than on physical infrastructure or money: like Microsoft, where rebuilding a burned headquarters is trivial because the people and culture remain, Japan and Germany only needed physical rebuilding, whereas building a Microsoft building in Iraq would fail because the human capital and culture that make it work are absent.

MacArthur succeeded in Japan by utilizing established mechanisms of credibility rather than imposing change wholesale: he photographed himself with the Emperor to acquire legitimacy, sent the Emperor on a public-relations tour to facilitate cooperation, and passed the new constitution through the Diet while accepting ~80-90% of Japanese leaders' suggested revisions—thereby lowering the transaction costs of coordinating people around a cooperative outcome.

Keeping the Emperor alive—despite 70-80% of Americans wanting him tried for war crimes—was a high-risk gamble by MacArthur because the Emperor could have become a rallying point for insurgency rather than a tool of legitimacy; the strategy worked but could have swung the entire reconstruction the other way.

The Iraq constitution superficially resembles Japan's (drafted with U.S. influence but approved by a representative body), but it failed because the U.S. played a key role in selecting the representative body and de-Baathification disenfranchised people, undermining legitimacy and contributing to the insurgency, whereas Japan used genuinely pre-existing, broadly legitimate institutions.

The contrast between the Japanese Emperor and Saddam Hussein is that the Emperor commanded the affection and respect of citizens and could coordinate them legitimately, whereas Saddam achieved internal peace only through coercion (jailing or killing threats), so keeping Saddam in power would not have replicated the Emperor's role—it would have required ongoing gunpoint coercion that large distrustful segments would reject.

Germany's reconstruction succeeded not because of occupation but despite it: the U.S. retained Nazi-era price controls and rationing, and recovery only began when Ludwig Erhard unilaterally abolished the price controls on the radio against the occupiers' rules, because freeing prices allows resources to flow to their most highly valued ends.

Japan and West Germany are not evidence of an American ability to centrally plan liberal democratic institutions; they are evidence of the ability to overthrow a regime (where the U.S. military excels) combined with recognizing and utilizing indigenous institutions (Japan) or, in Germany, succeeding despite imposed drags on the economy.

Bureaucratic infighting in both German and Iraqi reconstruction is explained by the public-choice model of bureaucracy: non-elected, non-profit-driven bureaucrats measure success by budget size and number of subordinates, so each bureau fights to be in charge to maximize its budget and headcount—meaning the dysfunction is structural, not a failure of effort by the Bush administration.

Somalia and Haiti, not Japan and Germany, are the relevant comparisons for Iraq and Afghanistan, because the future threat to the U.S. comes not from a nation-state superpower but from weak, failed states and rogue groups within them, and Somalia/Haiti share the absence of the conditions that made Japan/Germany succeed.

Somalia illustrates mission creep and the perception problem: humanitarian aid intervention expanded toward democratization (bureaucracies expanding their mission to gain power), and delivering aid to one side of a civil war was perceived by rival warring parties as assisting their enemy, causing them to shoot at U.S. forces and forcing the U.S. to either abandon neutrality or exit.

Haiti shows that even a partially successful reconstruction fails over time: the U.S. restored Aristide to power in 1994-96 and tried to build an uncorrupt judiciary and rule of law, but failed, and in 2004 Aristide was again forced from power and into exile, leaving citizens repressed and producing only at subsistence level because the government seizes surplus production.

Democracy is merely a mechanism for electing people (majority rule), which by itself is not good; what makes democracy beneficial to human beings is liberal constraints—protection of property, civil and political rights, and binding constraints on rulers so neither majorities nor elites can exploit citizens. Hamas is a democracy but not a free society.

Holding elections is the relatively easy part of democratization; the difficult part is establishing binding constraints on the winners so they cannot abuse those who voted against them, and failed democracies are precisely those where elections are not genuinely competitive because opponents fear being killed and constitutional protections are not respected.

Westerners have a fetish for nation-states and assume any government is better than none, but when the feasibility set does not contain liberal democracy, propping up weak or failed governments yields very bad outcomes; there is a good argument for allowing weak and failed states to fully collapse, because such governments are a pure cost on citizens benefiting only cronies, and a state weakening or collapsing signals it is not providing goods people want.

A large investment in education and human capital by itself does not produce development or democracy, because human capital is an input that requires complementary institutions—secure property rights and the opportunity to use one's skills productively; without those, human capital is worthless, paralleling the foreign-aid literature's repeated failures (savings gap, then human capital gap) ultimately explained by incentive and information problems.

Iraqi workers under Saddam's state-run economy were conditioned to wait for directives rather than engage in entrepreneurial thought—factory workers asked occupiers how much to produce and what inputs to use—because they were rationally responding to the incentives of a centrally-directed economy, so the latent commercial capacity could not simply be activated.

A free society works best where the need for coercion is least: in developed countries people cooperate voluntarily because they buy into shared belief systems, values and ideals, so the state cannot and need not be everywhere; where those beliefs are absent, continued force is required just to maintain cooperation, which is exactly what is seen in Iraq.

The 1919 post-WWI imposition of democracy on Germany (the Weimar Republic) failed—partly due to reparations and resulting inflation—and led to Hitler, whereas the post-WWII reconstruction succeeded; it is extremely difficult ex ante to isolate the factors that distinguished the two, which means historians cannot confidently identify a replicable recipe.

We know the ingredient list of liberal democracy (property rights, civil rights, political rights, rule of law) but not the recipe to create it where it doesn't exist; the implication is that since we don't know how to cook it, we should not cook in the first place—i.e., we should be highly skeptical of intervention.

Comparing troop numbers across countries (e.g., 'we used X troops in Japan but only Y in Iraq') is a meaningless way to explain success or failure, because 10,000 troops in Japan with no insurgency is fundamentally different from 10,000 troops in Iraq with an insurgency—context cannot be stripped out and transplanted.

The lesson from Japan, Germany, and the post-Civil-War U.S. is not that finding a national leader to rally around guarantees success; a national leader can maintain power for many reasons and can also work against the occupier's goals, and a genuinely legitimate leader opposing reconstruction could generate a worse outcome.

Success in exporting democracy is unpredictable, low-probability, and not creatable ex ante, so the default policy should be non-intervention combined with unilateral free trade; the U.S. should get out of the business of exporting democracy at gunpoint.

Unilateral free trade is a better tool of change than isolation/sanctions toward illiberal regimes: economic sanctions rarely if ever produce regime change, typically hurt the country's citizens by blocking goods, and give bad leaders fodder to blame the U.S.; trade is voluntary and offers both economic gains and cultural exposure to the values and ideals underpinning a free society, though it is not a panacea.

Freedom cannot be imposed because a free society works best where coercion is least; people must self-determine and choose freedom from the bottom up, which means allowing them to decide whether they want it, making 'imposing freedom' an inherent contradiction.

Saddam Hussein is gone, which can be viewed as a good in itself.