Non Democracies

Many studies claim to show that democracies do a better job than non-democracies of improving the welfare of the poor. These claims are consistent with leading political economy models, which suggest that democracies produce more public goods, and more income redistribution, than non-democracies (Acemoglu and Robinson 2005; Boix 2003; and et al. cited in Ross (2006)). Along with taxation from wealthy that helps to finance projects for the poor, democracy empower poor people through elections, and thus force governments to attend to their needs and, better information about the condition of the poor.
According to Michael Ross, whilst it may be true that democracies spend more money on public goods provision, it is not clear that they have any

In the second part came up with his own large sample data analysis (taking in to account the missed variable, wrong model...etc problems in the past studies), and reaches on a conclusion that democracy, has little or no effect on infant and child mortality rates. Finally suggests the democracy can be adjusted for both higher social spending, and unchanged infant mortality.
Ross used infant and child mortality rates as they reflect extreme poverty: the infant mortality rate, which describes number of death of children who fail to reach the age of one; and the child mortality rate, which describes the number death of children who fail to reach the age of five. He used a dataset that includes all states that were sovereign between 1970 and 2000, and had populations over 200,000. The study also used fixed-effects model to control for exogenous global health trends, and used a program called Amelia [Honaker et al. 2001] for the

They also produced faster drops in child mortality than the high-observation states. In other words, poor people were better off initially, and enjoyed more improvements, in the low-observation authoritarian states (e.g., Cuba,Poland, Oman, Libya, and Saudi Arabia) than the high observation authoritarian states (e.g., Zambia, Niger, and Mauritania). Excluding low-observation states in general will hence produce a sample in which authoritarian states have worse records than they do in the