Voters may like the idea of direct democracy, but as Garett Jones mentions in 10% Less Democracy, most scholars agree that representative democracy produces better outcomes. Similarly, while voters may thrill more to directly choose their top leader, better outcomes come from having voters pick legislators who then pick, and can remove, the top leader.

Here’s Arend Lijphart with some simple theory:

In parliamentary systems, only the legislature is popularly elected and is the clear and legitimate representative of the people, but in presidential systems both president and legislature are popularly elected and are both legitimate representatives of the people—but it is quite possible and even likely that the president and the majority of legislators have divergent political preferences. … There is no democratic principle to resolve such disagreements. … second problem is “rigidity”: presidents are elected for fixed periods of time. … third serious problem is the “winner take all” nature of presidential elections. … The fourth serious drawback of presidentialism is that presidential election campaigns encourage the politics of personality … instead of … competing parties and … programs.

In his new book Why Not Parliamentarism? Tiago Ribeiro Dos Santos collects much evidence favoring that option:

The main theme is executive subordination to parliament. However, there is no “cutoff” point that makes a country definitely presidential or definitely parliamentary. …

In this book, I argue that parliamentary systems are vastly superior to presidential systems. The word “superior” is deliberately chosen because parliamentary systems outperform presidential ones in just about any aspect worth pursuing by a respectable and universal ethical philosophy. Not only are they more efficient and more protective of individual liberties, but they are also more equal. Parliamentary systems are more stable, are less prone to coups, and are also more adaptable and incorporate changes quicker than presidential systems. They preserve traditions yet also innovate, and they grow faster and better protect the environment.

The evidence he collects is diverse, and for me persuasive. First there are raw correlations:

The number of parliamentary countries is [now] very close to the number of presidential countries. … [In] a ranking of countries by gross domestic product (GDP) per capita, … top 30 countries include oil-rich countries, 2 presidential countries (the US and South Korea), and no less than 22 parliamentary countries. .. [Re the] Sustainable Development Goals, [top 17 are all parliamentary, none of bottom 20 are] … [Re] UN Development Program’s Inequality-Adjusted Human Development Index, … only 2 of the top 30 countries are presidential, and none of the them are in the top 20. …

Then there is history:

We have failed to understand much of [old] European political thought … We read authors such as Edmund Burke, Benjamin Constant, Germaine de Staël, François Guizot, Alexis de Tocqueville, and John Stuart Mill. … [but] have forgotten the concrete, overarching project in which these figures all were involved, the one that made their thought intelligible. That project was parliamentarism. For each of [these] authors, … the defining feature of a free state was that it contained a space for parliamentary politics—an assembly in which political actions were discussed and deliberated and in which executive officials were held responsible. …

The phenomenon of the parliamentarization of Europe easily escapes many analysts because it was a gradual process and is often confused with democratization. … When the proper terms are used, parliamentarization for the expansion of parliamentary powers and democratization for the expansion of suffrage, it becomes clear that the first term is associated with general institutional improvements, while the latter does not have as much evidence in favor of [its value]. So European countries were the first to industrialize and were the first to become parliamentary. … In both a global and regional level, the first countries to adopt parliamentary systems did better earlier.

There is stat analysis of recent national experiences:

The study of presidential systems is teeming with confusion because the most studied presidential system, the US, is very atypical. Presidential powers in the US are severely more limited than in other presidential countries. …

Propensity for lower government spending in presidential systems. … find significant differences in favor of parliamentarism in 10 of the 14 outcome measures, with presidential systems not superior in any of them. Namely, parliamentary systems are associated with higher GDP per capita, better corruption control, bureaucratic quality, rule of law, telephone mainlines, import duties, trade openness, investment rating, infant mortality, and literacy. … In sum, despite not being unanimous, the state of the evidence points to a clear advantage of parliamentary systems.

There is the history of how competitive firms are organized:

Companies organized as presidential systems do not seem to exist [today]; all of them follow a logic similar to the parliamentary system, where shareholders elect a board of directors (the parliament) who hires and fires the CEO (the prime minister). … Unlike national governments, companies were able to experiment with a much wider range of constitutional arrangements, and experiment they did. In FPT’s database of companies before 1850, unincorporated companies had extreme freedom to select a variety of governance arrangements. [Freedom they often used.]

Finally, and to me most persuasive, there is evidence on U.S. cities:

Since the beginning of the 20th century, cities can—and increasingly have—adopted the “council–manager” model of government, whereby citizens elect a council, which in turn is responsible for choosing a city manager in charge of administration. This model stands in contrast with cities that adopt the “strong mayor” model, where citizens elect the mayor and council separately and the mayor is largely responsible for administration. …

For every objective measure that could be investigated, the results were either in favor of the council–manager model or a difference could not be found. … The only people who seem to evaluate mayors better than managers are mayors the selves, which is perhaps not surprising. … managers reduce spending in police forces without losing quality. … less wasteful tax exemptions in manager cities, … treasurers appointed by the council, as opposed to elected, borrow at lower rates. … “council–manager cities have stronger budgetary solvency compared with mayor–council cities,” … council–manager cities have less corruption.

As usual, the studies of variation across nations have a small N problem; if you try to include too many controls, you run out of data. In contrast, for firms N is huge, but one worries that their problems are too different, as boards of directors are rarely elected directly by shareholders. But the problem of city governmence seems close enough to nations, and there N is large. For example, in this study N = 12,238. And even that N is just barely enough to clearly show this:

switches to mayor-council (council-manager) form are associated with a reduction (increase) in spending of just over 9 percent, relative to jurisdictions with no change in government form in that year.

As reviews of this city governance literature seem consistently favor managers over mayors, that cinches it for me.

I’ve spent a lot of time over the years considering far more radical governance proposals. So the very slow trend toward adopting parliamentary systems seems quite discouraging. While I see enormous emotional energy around me devoted to politics, almost none of it tries to push for better overall systems for which the data is quite solid. What hope can there be for more radical but sensible change?

There is an Inter-parliamentary Union (IPU), but the richest country in the world, the US, is not even a member. The approved budget of the IPU for 2020 is around US$18 million. A sign of hope is that the UN adopted a resolution in 2018, establishing a partnership with the IPU and creating the International Day of Parliamentarism, celebrated for the first time on June 30, 2018.

I plan to celebrate that next year.

Added 7p: The author tweets re the growth advantage of parliaments:

That’s a good point. If I had to choose, I would go for McManus and Ozkan’s estimate of a 0.6 to 1.2pp higher yearly growth. I believe this is conservative because of attenuation bias I mention.

— Tiago Santos (@tribsantos) September 28, 2020