Last week, we analyzed data on 475 coups around the world from the past 66 years and learned that another coup in Turkey wouldn't be shocking. This week, we turn our attention to the economic impact of coups.
Geopolitics and economics tend to have a messy, confusing relationship. And where there is confusion, there's often only one way of getting to the truth: by turning to cold, hard data. Take Turkey's failed coup on July 15th. It is one of the most commented-upon stories of the year, and analysts have churned out reams and reams of speculation on what the coup would do to Turkey's economy.
As Turkey's lira (TRY) fell to an all-time low against the dollar on July 16th, Standard & Poor's demoted Turkey's credit rating to "junk" status, while Fitch warned it might follow suit. This prompted many to predict looming economic trouble. HBSC (reported in Business Insider) predicted that Turkey's coup "will significantly impact the country's economy", while Quartz said Turkey's economy was in "chaos".
Some pessimists argued that the lira's drop was ominous. However, while exchange-rate data can be helpful in the short term, it's far too volatile to use for predicting Turkey's economic well-being. The lira did touch an all-time low after the coup, but compare it with previous lows from the past year, and suddenly it doesn't look so precipitous after all. See the graph below.
The dot on the right side of the graph marks the exchange rate as of July 15, 2016. Graph is from xe.com.
Turkey's Deputy Prime Minister Mehmet Simsek said the lira's decline was just a "knee-jerk reaction" that would subside; in the long term, he claimed, the failed coup will have a limited effect on Turkey's economy. Many sites, including Business Insider and Bloomberg, noted that the lira has bounced back since its decline.
Meyersson's study contains some fascinating insights. For instance, successful coups in autocracies often improve economic performance, since they often replace a poor leader. The story is different in democracies. There, successful coups often institute a military leader and reverse positive economic policies, so they are associated with a 1% to 1.3% point decline in per capita GDP growth rates over the following decade.
However, Meyersson isn't much help if you want to gauge Turkey's prospects. His paper shows that the effect of a failed coup in a democracy is confoundingly mixed, and it is almost impossible to forecast any effect on economic growth within the next decade.
We decided to do our own analysis to hazard a guess about what might happen to Turkey's economy. We looked at data on 475 coups across 95 countries in the past 66 years. We then supplemented this with GDP data (which is a far better indicator of long-term economic well-being than exchange rate) and income classifications from the World Bank, as well as regime-type data from the Center for Systematic Peace. Our final data set ended up covering about 200 coups (many coups had to be excluded because of problematic GDP data or an inconclusive regime type).
What did we find? There were plenty of interesting nuggets about how economies have responded to past coups, but nothing conclusive about Turkey's economic fate. Turkey is an upper-middle income democracy that experienced an unsuccessful coup. Like with Meyersson's results, our data too threw up mixed results for this group of countries. The closest we could come to a decisive statement is this — whatever happens is likely to be only a small deviation from expected economic growth, rather than a huge swing one way or the other. While it's possible that the doomsayers might be proven right, going purely by data they shouldn't necessarily be so pessimistic.
Looking at basic GDP growth isn't helpful
If you look at GDP growth rates after each coup, there's no clear pattern. Many economies (such as Iraq in 1992, Syria in 1962 and 1969, the Dominican Republic in 1962, and Sudan in 1975 and 1976) grew during the year of a coup. However, many others (such as Iraq in 1991, Rwanda in 1994, Guinea Bissau in 1998, and Liberia in 1994) saw declines.
Why? Simple GDP growth rates might not respond well to country-level events like coups.
GDP growth responds both to what is happening within a country and in the world as a whole. For example, if a country had a coup in 2008 (in the middle of a global recession), a low or negative GDP could be caused by the slow global economy, not by the coup. Conversely, if a high-income country had a coup in 2000 (at the height of the dot-com bubble), it wouldn't be clear if positive growth was caused by the coup or despite the coup.
In short, looking at a country's GDP growth rate in isolation isn't enough. It's important to see it in the context of how the world at large is faring.
To separate out the effect of global trends that might influence a country's growth rate, we adjusted the GDP growth rate data.Here's how we did it:
1. For every coup, we found the GDP growth rate for that country in the year of the coup.
2. We then found the average GDP growth rate in each year for countries within each income bracket (high, upper middle, lower middle, and low), as classified by the World Bank. This is essentially the expected GDP growth rate for any country in a given income bracket.
3. For each coup, we looked at the difference between that country's actual GDP growth rate and the expected GDP growth rate (for countries in the same income bracket) in the year of the coup. For simplicity, we call this difference the "adjusted GDP growth rate".
If adjusted GDP growth is positive, it means that a country's GDP growth rate after a coup was better than expected. If adjusted GDP growth is negative, it means that a country's GDP growth rate after a coup was worse than expected. If adjusted GDP growth is zero, it means that the country performed as expected.
Magnitude of adjusted GDP growth changes has muted over time
First, we figured we'd start with the big picture trends. We know from our previous blog that the number of coups has been decreasing in recent decades. So it's reasonable to question whether the adjusted GDP growth rate after a coup has also changed over time.
When we organized the coups by decade, we discovered that the economic change (either positive or negative) after a coup appears to be somewhat muted in recent years. Countries experienced the greatest deviation from expected GDP growth rate in the 1990s, and the lowest deviation in the 2010s.
Since Turkey's latest coup just happened this year, we could expect any economic effects to be muted as well.
Adjusted GDP growth rates are more resilient in democracies
Next we looked at another key question — how democratic and authoritarian countries respond to coups.
Autocratic countries seem to have experienced a greater deviation from the expected GDP growth after a coup. However, in democratic countries, the adjusted GDP growth rate appears to be muted.
Since Turkey is a democratic country, any economic effects it sees as a result of its coup might be more muted.
How adjusted GDP growth rates changed after different types of coups
While the big picture is important, we know from Meyersson's research — and common sense — that all coups are not alike. A stable, rich country will respond to a successful coup far differently than an unstable, poor country. A democracy and autocracy will respond to a coup differently.
Our next step was to examine the interplay between adjusted GDP growth and these factors (country income, type of regime in power at the time of the coup, and success of the coup) after a coup.
There's a lot of data in that graphic, so let's break down some of the big insights.
All other factors aside, upper-middle income countries and low-income countries are hit hard most frequently after a coup. Lower-middle income countries have more mixed experiences. (Our insights ignore the data from high-income countries, since that sample of just 2 countries is far too small.)
Within low-income countries, unsuccessful coups cause more trouble than successful ones, which produce more mixed results. This means that, for low-income countries, the regime in power doesn't seem to matter as much.
This is very different from upper-middle income countries, where success alone is unimportant. Failed coups in autocratic countries and successful coups in democratic countries seem to produce similar economic results — both are associated with economic slowing. Meanwhile, successful coups in autocratic countries and failed coups in democratic countries are both associated with more mixed responses.
For lower-middle income countries, one type of coup seems the most problematic — successful coups in autocratic countries. Nearly 90% of these coups led to economic slowing. All other types of coups produce more mixed results; sometimes the economy actually accelerated after a coup.
How does Turkey fit into this picture? Turkey is currently classified as a upper-middle income democracy. If its recent coup was successful, history suggests that Turkey would be in for some level of economic slowing. However, its coup failed, for which the data is more mixed. While 57% of upper-middle income, democratic countries with a failed coup experienced economic slowing, 43% actually saw some level of economic acceleration after the coup.
More detail, murkier insights
Data on the economic impact of coups is fascinating — the farther we went into the data, the murkier the insights became.
At a high level, it was easier to draw relevant insights, such as the adjusted economic effects of coups being muted in the last decade and in democracies. We can draw a straight line from these insights to Turkey's economic future. However, once we went deeper into more relevant data — like upper-middle income democratic countries with an unsuccessful coup — the insights became less clear.
While we aren't certain about Turkey's economic future, we can be certain about one thing: commentators should not be predicting Turkey's future —or the future of any country that goes through similar turbulence — with unqualified confidence. Historical data shows that Turkey's economy could respond in many different ways within the next year, and any effects that do show up will possibly be muted.
Meanwhile, we wish the Turkish people a peaceful, happy future. Wouldn't it be awesome if Turkey did so well that our next data story was an analysis of Turkey's amazing peace and its effect on economic growth?