In his recent book, Presidentialism, Parliamentarism, and Democracy, José Antonio Cheibub classes Pakistan as a mixed democracy from 1972-76. His definition of a mixed democracy is where the government is responsible to the assembly, where there is an independently elected president and where the government is also responsible to the president, meaning either that the president can dismiss the cabinet directly or that the president can dissolve the legislature and so dismiss the government indirectly.

So, Cheibub’s definition of a mixed system is slightly different from the definition of semi-presidentialism used here. In the end, though, his classification of countries is very similar to my own. There are some differences. He classes Ireland as parliamentary because the president has no power to dismiss the government either directly or indirectly, but there are no other parliamentary examples that I would class differently. The situation with presidentialism is not quite the same. He includes a number of cases as presidential that I would think of as semi-presidential: Burundi (1993-95), Cuba (1946-51), Guinea-Bissau, Namibia, Peru and Sri Lanka. In most cases, I think this is because Cheibub does not consider that the government in these countries is/was constitutionally responsible to the legislature. (See a previous post on Sri Lanka). However, the category that intrigues me the most is the mixed category. Here, all his examples are those that I would class as semi-presidential with two exceptions: Brazil (1961-62) and Pakistan (1972-76). I have posted about Brazil at that time before. But what about Pakistan?

Pakistan adopted a new constitution in 1973 and it was clearly parliamentary. The president was elected by an electoral college comprising parliamentarians and members of the provincial assemblies. So, I am not sure why post-1973 Pakistan can be classed as mixed. In any case, though, Cheibub’s start date is 1972. Prior to that time there had been only one presidential election, which was in 1965 and it, too, was organised on the basis of what I would consider to be an electoral college. It comprised 80,000 electors from the (then) two parts of Pakistan. My assumption is that Cheibub considers this to be sufficient to constitute an independently elected president. Even so, as far as I understand it, the constitution was formally suspended in 1969 when martial law was decreed. Overall, I am not sure either why 1972 is the start date indicated or why the period extends beyond 1973 when the constitution was parliamentary.

None of these comments should not be taken as a criticism of Cheibub’s book. It is a really interesting book. Please go out and buy it. He makes a convincing case that there is no evidence to suggest that presidentialism is worse then parliamentarism or mixed democracies in terms of the propensity for democracy to collapse. He bases his definitions of regime types on constitutional clauses, which is something I really support. It is just that, given his list of mixed democracies is usually so consistent with my list of semi-presidential countries, I am intrigued by the reasoning behind the decision to class Pakistan as mixed from 1972-76.