Goldstone et al: A Global Forecasting Model of Political Instability

JA Goldstone and PIT Force, A global forecasting model of political instability (Political Instability Task Force, 2005).

These authors develop a model that is able to predict state failure out 2 years with 80% accuracy. This does not mean that they predict state failure with 80% accuracy, but whether a state will or will not fail with 80% accuracy. This is a simple model, with a certain kind of regime type being understood as the most highly correlated with failure.

Firstly, there was the operationalization of state failure. For these authors, it is defined fourfold: revolutionary wars (over 1,000 battle deaths in one year), ethnic wars (same), adverse regime changes (drop of 6 points in polity score) or genocides/policies (direct targeting of political parties or ethnic groups by governments w/o death numbers specified). This list was compiled with the help of regional experts. Their n values for each: 62 revolutions, 74 ethnic wars, 111 adverse regime changes, 40 genocide/policies from 55-03.

Their list of independent variables was also compiled with the help of area experts and totaled 75.

The broader goals of the political instability task force are to both understand the variables that cause instability, and to create a model that can identify countries that are likely to be unstable.

As failure is an incredibly discrete event, the method used was a case-control method, which is used for identifying rare diseases in large populations. A positive dependent variable is coupled with a negative dependent variable and the populations are compared. Independent variables were also shifted two years prior to the onset year for dependent variables.

The findings of the group were that relatively simple models can be used to identify cases of instability. Even though there have been myriad explanations for the causes of state failure, these are not the case across the board. Goldstone (2001) argues that, “…the origins of a political crisis can best be understood by turning the problem on its head, asking what factors are necessary for a state to sustain stability despite the various problems-economic, political, social- it might encounter” (8).

The authors found that many variables that are traditionally highlighted as being drivers of state instability miss the mark, as they are not able to point to the underlying instability in the political system, but end up exploring effects of political instability. High inflation, high youth bulge and birth rates are indicative of poor governance: “…in countries that are poorly governed, it is more likely that there will arise bouts of high inflation, or sharp economic reversals, or that people will rely more on family support…” (10). State failure is not understood by looking at the list of challenges that a state must face, but rather the resilience of the state to face those challenges.

Two polity variables that lead to very good results in combination was the degree to which a political process was fractionalized, and the degree to which there was open competition for the political office. “The combination of a winner-take-all, parochial approach to politics with opportunities to compete for control of central state authority represents a powder keg for political crisis” (12).

“According to our research, most economic, demographic, geographic, and political variables do not have consistent and staticially significant effects on the risk of instability onset” (14).

“The model essentially has only four independent variables: regime type, infant mortality, a ‘bad neighborhood’ indicator flagging cases with four or more bordering states embroiled in armed civil or ethnic conflict, and the presence or absence of state-led discrimination” (15).