When it comes to revolution, geography matters. Capital cities that are densely populated and accessible are vulnerable to revolution, inaccessible and lightly populated capitals are not. That, at least, is the argument of a paper from three economists and political scientists, "Isolated Capital Cities and Misgovernance: Theory and Evidence."
The authors begin from a simple observation: "Countries with isolated capital cities display worse quality of governance." They posit that in countries without constraining institutions (ie democracies), the threat of rebellion is the only constraint on the ruling class. For a rebellion to succeed, whether in overthrowing a government or simply influencing its behavior, it needs a large presence in the capital. Think of the success of massive protests in Cairo compared to the continued inability of Syrian protesters and rebels - always based outside of the capital - to oust President Bashar al-Assad. Isolated capitals should be correlated with and in fact help explain the existence of predatory, autocratic rulers and poorly governed states.
The researchers find that the data backs up their theory. Isolated capitals correlate with reduced power sharing (as measured by a number of indices of democracy or political freedom commonly used in political science), increased transfer of resources from the masses to the elite as shown by the proxy of the income premium of people living in the capital, and worse governance as measured by the World Bank's Worldwide Governance Indicators. Countries with isolated capitals also spend less on military and police forces, as expected due to the decreased threat of rebellion.
The authors also find that rulers do seem to be motivated by this logic. Countries have relocated their capitals about once every six years since World War I, almost always to more isolated environs. This is unsurprising given that capitals are often the most populous city, but the new capitals are often significantly smaller or built from scratch in areas where the geography did not induce the formation of cities. Burma's decision to relocate its capital to remote Naypyidaw in 2005 is a particularly stark example:
"Vast and empty, Burma's new capital will not fall to an urban upheaval easily. It has no city centre, no conned public space where even a crowd of several thousand people could make a visual - let alone political - impression. Naypyitaw (sic), then, is the ultimate insurance against regime change, a masterpiece of urban planning designed to defeat any putative 'colour revolution' - not by tanks and water cannons, but by geometry and cartography. 320 kilometres to the south, Rangoon, with five million people, is home to one-tenth the country's population. But even if that city were brought to a standstill by public protests and demonstrations, Burma's military government - situated happily in the middle of paddy fields in the middle of nowhere - would remain unaffected."
As the passage alludes to, architecture also matters. Although the authors do not address it, nonviolent struggles like the Egyptian Revolution and the Color Revolutions benefitted from large public spaces in the capital that could be occupied, so Naypyidaw chose not to build any. Similarly, the 19th century renovation of Paris that endowed the city with its famous long boulevards was as much an effort to thwart rebellions as beautify and modernize the city. The dense city of Paris had known constant rebellions, and its narrow, chaotic streets facilitated barricades. Introducing wide boulevards allowed the military to more easily monitor the city and respond to rebellions.
Rebellions, revolutions, and the misrule of dictators are motivated by emotions like pride, desperation, and greed and negotiated by politics. But the influence of geography and architecture should not be ignored.