This chart plots the "annual likelihood" of the onset of political instability according to the POLITY score of regimes at the point of onset for all cases of political instability for all countries during the contemporary period, 1955-2006. The chart plots three perspectives on political instability using solid, line-and-marker plots. The dashed line plots use the POLITY2 values and incorporate interpolated information regarding periods of "interregnum" (-77) and "transition" (-88); in this rendition, periods of "interruption" (-66) are treated as missing data. Both of these plots exclude cases of political instability that take place at the time of state independence (8 instability cases). The plots are "smoothed" by taking three-score averages for both POLITY scores and event counts (e.g., the number of POLITY years for POLITY score -7 is calculated by summing the numbers for -6, -7, and -8 and dividing by 3; likewise, the number of instability onsets for POLITY score -7 is calculated by summing the numbers of onsets for -6, -7, and -8 and dividing by three). This "smoothing" technique is consistent with the estimated error in the POLITY scores of +/- 1 point.

The most restrictive perspective is the PITF definition of the "consolidated case" of political instability (plotted by the teal lines), which is a continuous episode of political instability defined by the occurrence of one or more cases of adverse regime change or major political violence, overlapping in time or in succession, during which there is no period of "stability" between the end of an instability case and the beginning of another case that lasts more than five years. This perspective distinguishes between a general condition of "stability" and a phase shift to a general condition of "instability" in a particular state. The "annual likelihood" for this perspective on political instability is calculated by determining the POLITY score of regime at the point of "consolidated case" onset and, then, dividing the number of onsets for a particular POLITY score by the number of "stability years" that particular POLITY score occurs in the dataset (i.e., a subset of the Polity IV data series that does not include POLITY scores for years during which a state is considered to be in an instability condition and for the five years immediately following the end of a period of instability). For more detailed information on the PITF "consolidated case" methodology, see the PITF Web site.

The blue lines plot the "annual likelihood" of the onset of any one event from either of the two categories of political instability events included in the PITF perspective: adverse regime changes and major political violence. The category of "major political violence" includes ethnic wars, revolutionary wars, and genocides and politicides; however, cases of genocides and politicides are included only when they refer to a distinct episode of political violence, rather than an escalation of violence or additional tactical use of violence in an ethnic or revolutionary war. Additionally, six cases of war are "double-counted" in the PITF data as they are designated simultaneously as both ethnic and revolutionary wars; these cases are counted only once in the onset calculations. As onsets in this perspective can occur at any time, regardless of whether a separate episode is already occurring in a country, the "annual likelihood" for this plot is calculated by determining the POLITY score of regime at the point of case onset and, then, dividing the number of onsets for a particular POLITY score by the total number of years that POLITY score occurs in the comtemporary period dataset.

The maroon lines plot the "annual likelihood" of the onset of any one event from either of the three categories of political instability events that include "major democratic transitions" with "adverse regime changes" and "major political violence." "Major democratic transitions" disrupt and destabilize status quo authority regimes in ways very similar to "adverse regime changes" even though we may prefer, or even encourage, changes toward greater democratic (or less autocratic) authority. "Annual likelihoods: are calculated using the same, basic methodology as the two plots already described (above).

In general terms, discounting "major democratic transitions" in political instability analyses tends to bias findings toward over-emphasizing the volatility and risk associated with democratic transitions and under-estimating the volatility and risk associated with autocratic regimes. Indeed, using the PITF "consolidated case" approach gives the impression that "strongly institutionalized" autocracies exhibit (low) levels of risk comparable to that of the most "highly institutionalized" or "consolidated" democracies. Clearly, the relative stability of democratic authority is only realized in the consolidated democracies (+9 and, especially, +10 on the POLITY scale). The various forms of "strongly institutionalized autocracies" are 6 to 10 times more likely to experience an instability event than are the "fully institutionalized (+10) democracies" according to the full instability analysis presented here. "Non-consolidated democracies" have levels of risk for political instability onset similar to that of the autocracies and it is the category of "anocracies" that have the highest risks of instability onset. Unfortunately, it appears to be quite difficult, but not impossible, to avoid the highened risk associated with the condition of "anocracy" as a preparatory "step" to "fully institutionalized democracy." There were 81 cases of "major democratic transitions" in (-10 to -6) autocracies during the contemporary period. Of these, about half (43 cases) resulted in shifts from autocracy to the higher risk category of anocracy, while 32 resulted in shifts to the similarly high risk category of "incomplete democracy" (+6 to +8). Only 6 transitions resulted in shifts from "strongly autocratic" to the lowest risk categories of "highly institutionalized democracies" (+9 or +10).


Back to Polity IV Main Page

Polity IV Home
CSP Logo CSP Text Logo
Copyright 2008