One year after 'Asante Sana', coup sounds still rumbling

One Tuesday, November 14 2017, military elites took control of all security arms of government and moved tanks into Harare's central business district to take control of government. Four hundred and fifty-three consecutive months of successfully holding onto power and thwarting political opponents does not guarantee one a 454th month in office," says One Earth Future (OEF) associate director, Curtis Bell in an article titled, Mugabe: An Unlikely End. True, it didn't guarantee former President Robert Mugabe. Nearly 40 years after Zimbabwe's independence and, Pearl Matibe reflects on the coup d'état - one year later - and coup-related research from OEF.

In the unstable days preceding Mugabe's resignation, many tried to make sense of Zimbabwe's new reality, staying tuned in to every update, and staying in touch with friends and family, under increased military presence to ask what to expect. They weren't looking for concrete predictions — one of the unsettling and disconcerting things about those long hours even well into the night is that no one had any reliable information about what was truly happening — except for the alarming sense and understanding of the life-changing disposition when a military power is in charge.

The country has changed in the past year, and many have grown numb after unrelenting political and socio-economic shocks. What now passes for ordinary, would have once been inconceivable. Then the August 1 killings of unarmed civilians happened; it was staggering when credible evidence emerged of men dressed in military fatigues and balaclavas firing live ammunition from AK-47 assault rifles into unarmed civilian crowds.

The coup-anniversary brings mixed feelings as we recall how the last November footage showed armed soldiers patrolling the streets. Today, Zimbabwe now has a militarised police and intelligence service; a new level of dysfunction for what was once known as the bread basket of Africa and most promising economies.

To begin reflecting on coups, Pearl Matibe (PM) interviewed political psychologist, computational social scientist and researcher, Clayton Besaw (CB). Besaw is a political events forecaster with OEF whose work includes the development of forecasting and machine learning systems for predicting conflict events such as coups, election violence, and intrastate conflict outcomes. OEF is best described exclusively as a content and research expertise organisation.

While you can't ignore the politics of any country, in the interview questions, Besaw provided his expert opinion:

PM: On August 1 2018, Zimbabwe's military deployed soldiers who shot live bullets into civilian crowds. What was your CoupCast machine learning prediction of the likelihood of a coup attempt on that day prior to presidential election results being announced? Do you have a Git-image/file of that to share with our readers?

CB: Unfortunately, our public maps and tables are not backed up by image snapshots, but we do keep historical data in our archives. In lieu of an image, let me provide you with our monthly archives of coup risk for Zimbabwe before the results announcement (July 1 to August 2), after the announcement (August 3 to August 31), and one month after announcement (September 1 to 30). The source of this data comes from our [The] Rulers, Elections, and Irregular Governance Dataset (Reign) monthly updates that are archived over time. I can provide you with the spreadsheets if you wish as well.

Unfortunately, we cannot produce daily risk estimates, because the data we use to predict coup risk can only be updated during a monthly period and we try to update our monthly forecasts after all election results are announced as well.

Brief explanation of each number. Coup risk percentile tells us how Zimbabwe compares to the rest of the world. For example, July's percentile was 97%.

This means that Zimbabwe had an estimated risk of coup attempt that was higher than 97% of all other countries in the globe during the month of July. The coup risk probability is just the raw estimate of how likely a coup was to take place.

Before results announcement in July/Early August: Coup risk percentile = 97%, coup risk probability = 14%

After results announcement in August: Coup risk percentile = 85%, coup risk probability = 4%

One month after announcement in September: Coup risk percentile = 85%, coup risk probability = 4%

Essentially, these forecasts tell us the following story. In the lead up to the election, economic and social conditions in Zimbabwe alongside uncertainty towards the election outcome made Zimbabwe the likeliest country in the world to experience a coup attempt. Zimbabwe is categorised as a type of authoritarian regime (dominant party government), and these types of uncertain elections appear to make such regimes more vulnerable to the risk of a coup taking place.

Once the results were announced, and Mnangagwa declared the winner of the election, CoupCast forecasted a large decrease in the risk of coup probability even though Zimbabwe remains comparatively high in its coup risk percentile. When incumbent governments in these types of regimes do win elections, our data suggests a greater increase in political stability on average.

PM: From your research, what might be a strategy for peacefully reducing coup threats in Zimbabwe?

CB: I want to qualify my answer to this question in two ways. First, I want to be careful in painting a broad picture. Zimbabwe is a unique country with complex political and social circumstances and has defied trends in Africa as a whole. Second, I know that you have been speaking to my colleague Jonathan Powell and I would like you to defer to him regarding strategies for reducing the risk of coups in Zimbabwe in the future. With that said, I can offer a few pieces of insight into how Zimbabwe may avoid further coup events into the future.

The current research on coup attempts suggests that there are multiple factors that may decrease the threat of coups. First, it appears that the structure and independence of a country's military is by far the most important factor. The military seems to be less likely to engage in coup behaviour if they are given adequate resources and materials to keep their decision-makers content and happy. This may be key to keeping them from intervening in civilian political disputes.

Second, it is important that countries like Zimbabwe continue to engage in regional and global economic partnerships and trade. Zimbabwe has a unique political climate and stability likely depends on all parties engaging in compromise and working towards economic development and partnerships. This is likely especially important in the context of regional African trade and development. The more integrated Zimbabwe is into these regional and global partnerships, the less likely it is to turn to military means for solving domestic political disputes and the more likely it is to abide by constitutional and electoral means.

Overally, the political actors in Zimbabwe should try and achieve two common strategies going forward. 1. Provide the military with enough resources and mandate to provide the collective good of national security for Zimbabwe as a whole. 2. Further work together to help Zimbabwe develop greater economic partnerships both regionally and globally.

With those suggestions, I suggest that you integrate them with the wise advice of Powell and consider them with your own wisdom and expertise of Zimbabwean politics. These strategies are likely difficult to implement fully and will require considerable compromise and negotiation going forward.

Other research

Researchers who have studied post-election coups, such as Tore Wig and Espen R,d for example, believe that if the incumbent performed poorly, it could have caused a second coup after July. This may point to the question held in some quarters, both domestically and abroad, on if the Zimbabwe presidential result would be accepted by the Zimbabwe Defence Force (ZDF).

Interesting coup facts

According to the OEF data, out of almost 500 coups since 1950, only two, excluding the Zimbabwe coup, happened to leaders who have been in power for more than 400 months (around 33,3 years). Alfredo Stroessner in Paraguay was ousted by a coup in 1989 after 35 years in power. A coup also took out Haile Selassie in Ethiopia in 1974 following 33 years in power.

More recently, Gnassingbe Eyadema, who had been the President of Togo for 38 years, was replaced by his son by a military order, immediately after his death in 2005. The event is formally considered a coup because the military intervened to preclude a constitutional succession, but Eyadema could hardly be considered to be a sitting incumbent as he died hours before the military acted. This case considered, Mugabe is now the oldest leader to suffer a coup attempt since 1950.

When looking at the data by the age of the leader, rather than leader's tenure, a coup against Mugabe was even less likely. Since January 1950, a leader over 85 years old has never suffered a coup attempt.

The oldest leader to have been ousted by a coup since 1950 has been Habib Bourguiba, the first President of Tunisia, ousted by a coup in 1987. Three leaders were ousted at the age of 82: Alberto Demicheli of Uruguay, Selassie of Ethiopia, and, most recently, Hosni Mubarak in Egypt. With this coup, Mugabe earns the unhappy distinction of being the oldest leader unseated by a coup.

OEF data and methodology

The Reign Dataset (Rulers, Elections, and Irregular Governance) covers political conditions in every country each and every month. They update the data set monthly to reflect the most recent political events, such as coups, world elections, and changes in political leadership. They also provide monthly election coverage and track leadership changes in a series of updates called international elections and leaders.

These conditions include the tenures and personal characteristics of world leaders, the types of political institutions and political regimes in effect, election outcomes and election announcements, and irregular events like coups, coup attempts and other violent conflicts.

The dataset covers more than 200 countries for each month from January 1950 to the present. Reign was created by gathering original data, compiling other datasets on political conditions, reviewing their coding rules, and updating all information to the present [Bell, Cur tis. 2016. The Rulers, Elections, and Irregular Governance Dataset (Reign)].

The CoupCast project, led by Curtis Bell, uses historical data and machine learning to predict the likelihood of a coup attempt occurring in any country on a month-to-month basis. Using historical coup activity, the CoupCast identifies the baseline risk factors and short-term triggers that greatly increase the likelihood of a coup. The forecast is updated monthly.

In brief, here's the CoupCast methodology: Coups, unlike other political crises that unfold over weeks, months, or years, are precisely timed events aimed at ousting a very specific individual from power. This precision means the risk of a coup may vary greatly over the course of a year. It can change instantaneously during transitions between leaders. For this reason, CoupCast estimates a unique risk of a coup attempt for every individual leader for each month he or she is in power.

To generate a predictive model, they collect data on coups and the conditions that decades of coup research link to coup plotting. These data include:

 Historical coup data on approximately 600 coup attempts dating back to 1920.

 Socio-economic conditions including GDP, economic growth, infant mortality, and extreme deviations in precipitation.

 Political conditions including measures of militarisation, democratisation, regime longevity, and the timing and outcomes of elections and referendums.

 Political violence indicators, both within a country and in the region.

 Regional shocks that capture political instability, change, and economic shocks in nearby countries.

 Leader traits, such as age, military experience, time in power, and method of entry.

They use historical data to find the statistical relationships between these variables and coup attempts and then use those relationships to forecast the risk of a coup against every world leader within the next month. They produce monthly estimates in three stages:

Stage 1: Theoretically-Informed Models

Stage 2: Machine Learning

Stage 3: Combined Model

Coup research figures

Below are figures Matibe retrieved from OEF that demonstrate their coup research findings.

Key questions moving forward

The case of Zimbabwe and the trends found in the OEF data provide insight into two important issues. First, elections represent a source of potential instability across both democratic and autocratic regimes. Second, scholars should seek to further understand the relationship between political instability and election cycles, including election announcements, delays, and elections that are ultimately cancelled.

Call-to-action

The military is rapidly losing its appeal with pro-democracy actors. Whatever feud that may or may not be brewing pitting President Emmerson Mnangagwa against Vice-President Constantino Chiwenga (if it exists in earnest!) or either of them against the opposition voices, should be resolved quickly and peacefully. And, any future of Zimbabwe's, without a Nelson Chamisa leadership-governance and international engagement will not produce sustainable peace and stability in Zimbabwe.

As you commemorate or recall the military-linked events of November 2017, it might be that you recall the sound of Zimbabwe Broadcasting Corporation (ZBC)'s music for the start of the 8pm news. One day soon we may even hear an epitaph for a coup d'état: "Asante sana. Iwe neni tinebasa"