Thoughts from a Renaissance man Economics & Strategy

6 September 2016

Charles Robertson Thoughts from a Renaissance man +44 (207) 005-7835 [email protected] The revolutionary nature of growth (2016) Mobile +44 7747 118 756 Big data allows us to quantify political risk and we must do so, in order to @RenCapMan avoid pointless attempted coups and excessive political repression. It may also be useful for investors. Vikram Lopez +44 (207) 005-7974 In any given year, there is an 86-100% chance of no change in political [email protected] regime type The attempted coup in in July 2016 resulted in a reported death toll of over 240 people. This is a tragedy that could have been avoided if the perpetrators had recognised that the coup was doomed to fail – please see our report The revolutionary nature of growth, published 22 June 2011. With a broader 7,776 data points, we reiterate and update our conclusions from that report. The underlying thesis – using political ratings by Polity IV and per capita GDP (2011 dollars) from the Penn tables – is that as countries get richer, they all democratise unless they export a lot of oil per capita or are called . Here we examine all emerging, frontier and beyond-frontier countries within CLICK HERE TO SEE THE VIDEO that theme to highlight the political risks looming on the horizon. Our most important conclusion is that in any given year, our data show there is an 86-100% chance of no change in political regime type. But democratisation is coming: from to The Polity IV database provides 200 years of political rankings, from the ‘full ’ of modern at +10 to ‘democracy’ at +6 to +9, autocracy at -6 to -10, with that have characteristics of both, which can be open anocracies (0 to +5) or closed (-1 to -5). The data tell us that we can confidently assume countries with a negative score will shift to a positive score as they get richer. While it could happen earlier, we expect autocratic China to change in the , and autocratic Vietnam to follow. The Taiwanese or Korean examples suggest this will be a voluntary and relatively peaceful ceding of monopoly power by the ruling parties. Meanwhile, the ‘closed anocracies’ of and Thailand have a remarkably high 12% annual chance of becoming more democratic, and Morocco and Jordan have a 9% chance of doing the same. The greatest two-way risk is in low income countries regained ‘democracy’ in 2015 and has a 96% chance of retaining this rating each year. But lower income countries from Zimbabwe to carry a 1 in 7 risk of political change, and are more likely to shift towards autocracy than democracy. , Turkey, - the middle income trap and a new democratisation wave We find that autocracies are not significantly better than at delivering growth in low income . But our data imply that democracy might be essential if a country is to escape the middle income trap. Russia, Turkey and Malaysia retain positive Polity IV ratings, but all lost ‘democracy’ in the past decade and became ‘open anocracies’. This has never happened before at a $15-25k per capita GDP level (2011 PPP). The world has also never seen an open with per capita GDP above $25k, which suggests to us that these countries will either regain ‘democracy’ and get richer, or stagnate economically and politically, or become the world’s first rich open anocracies. Malaysia’s 2018 could see it regain ‘democracy’. For Turkey and Russia, the data imply a change of president might be required to avoid stagnation over the medium term. Curiously, when oil prices fell in 1985, we saw a democratisation wave engulf , the USSR, and Algeria in 1988-1990. We wonder if the 2014-2015 oil price fall may provoke similar democratisation in Russia, Iran and in 2018-2020. The data sample is too small to give us high confidence in this scenario. To conclude, in any given year, we think political change is unlikely. But when the risk of change can be as high as 1 in 7, we think investors should be prepared for the unlikely to happen.

© 2016 Renaissance Securities () Limited. All rights reserved. Regulated by the Cyprus Securities and Exchange Commission (Licence No: KEPEY 053/04). Hyperlinks to important information accessible at www.rencap.com: Disclosures and Privacy Policy, Terms & Conditions, Disclaimer.

Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Friday 15 July, was meant to include a relaxed evening, eating flame-grilled steak, drinking Malbec and melting marshmallows around a campfire, followed by the discomfort of sleeping in a tent.

However, a faction of the Turkish army that had never read our 2011 report, The revolutionary nature of growth, mistakenly thought it could overthrow the Turkish government. As we read the news at around 9pm, we instantly tweeted that “Turkey is on the cusp of having an immortal democracy so this apparent coup is extremely unlikely to succeed”. This tweet was arguably too generous to Turkey by calling it an immortal democracy, and too generous to the plotters by suggesting there was any chance of the coup succeeding. A few minutes later, before anyone had heard from President Recep Tayyip Erdogan, we were suggesting that the most concerning consequence of this coup attempt was that he might end up with even more power.

This is not a ‘we told you so’ piece. What we think we proved in 2011 was that data can predict some political changes – specifically that coups do not succeed in countries with Turkish wealth levels. We can no more take the credit for this, than a scientist can take the credit for suggesting that the sun rises each day.

Yet these facts are sadly still not understood. More than 250 died in Turkey because of a coup that all of history tells us could not succeed. Since 15 July, President Erdogan has purged more than 60,000 people. This is 10x less than the 600,000 arrested after the 1980 military coup1, but looks like an over-reaction to a coup that was doomed to fail. US investors told us in mid-July that they think Erdogan is overplaying his hand. Meanwhile many political commentators have sought to explain what happened on 15-16 July with reference to social media, but too few have stated the basic truth that “Turkey was too rich for the coup to succeed”.

What we think is most important is that the coup attempt proves how bad we all are at understanding when societal change has occurred. Some in the Turkish military presumably thought the coup could succeed, because coups have so often succeeded in Turkey. They failed to recognise that Turkey in 2016 has changed fundamentally from the country it used to be.

We see the same misunderstanding of change in Russia and China. Some believe that President Vladimir Putin holds as much power as he does in Russia, because Russians have always had autocratic leaders and always will do. China has never been a democracy so some (such as the rulers in Beijing) suggest that China is not suited to democracy. Yet similar arguments could have been made (and probably were) in the run up to the decision to let Swiss women have a vote in 1971 – yet change came, women got the vote and got richer. It does take time, but change does come and is inevitable for almost all of us.

Given this, we think it is worth re-visiting The revolutionary nature of growth. We can also do better than we did in 2011, because:

1) We can capture the Arab spring, as we can now extend the 1950-2010 political dataset to 2014. We are also using 2011 PPP dollars, which is replacing the former 2005 PPP benchmark. As a result, the data are more accessible to us, and to you,

2) We wound up a lot of people by simplistically calling Russia and Turkey democracies. We will be more nuanced in this piece, and also look at the erosion of democracy, a process that has been under way in key markets such as Russia and Turkey in recent years, and where they go next.

1 1980 data courtesy of this piece.

2 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

3) Lastly, we can update our forecasts. In 2011 we pointed out that if China grew at 2005-2007 rates of growth, it would be a democracy in 2017. But China has not grown that fast, so it’s date with democracy has changed.

‘All’ countries become democracies

We encourage you to re-read our original report, where we discussed Abraham Maslow’s hierarchy of needs, and how tax-payers will eventually demand a say in how their taxes are spent. At low income levels, the priority is food and shelter, but the need for Hegelian recognition becomes increasingly important after basic needs are met. Francis Fukuyama made this point in The end of history and the last man with the core thesis that all will end up as capitalist democracies. We pointed out in 2011, and reiterate now, that neither Islamic theocracy nor one-party communism are proving to be alternative systems with universal appeal.

However, as we saw in the 1930s after the 1929 stock market crash, and in the 1970s after the 1973 crash, and now again after the 2008 crash, the road to democracy is not a straight one. In each case, the political status quo was challenged. In the 1930s, when European countries were poor and democracy was vulnerable, we saw a resurgence of autocracy, and at the same time a Keynesian challenge to income inequality and classical economics. After the 1970s crash, we saw a rise of radical terrorism but this could not dislodge the democratic system in rich countries. What did change was the political , with a successful neo-liberal challenge to Keynesian orthodoxy and what was seen as excessive income equality. Today again we see changing, and we expect a swing back towards Keynesian policies (once politicians learn that markets will let them borrow) and income equality2. Already new prime ministers in and the UK have set out these goals. But democracy itself is not under threat in rich western countries. Where big political change can occur is in lower income countries, and it is this which we focus on here.

In this first section, we use very broad definitions. We call any country with a positive Polity IV score a democracy, and any country with a negative score we label an autocracy. We change these definitions and get more precise, in later sections.

When does this shift from autocracy (we all start with autocracy) to democracy take place? The most risky period for an autocracy is when per capita GDP is in the $8-15k per capita GDP range, using PPP dollars. China in 2014 stood at $12,472. We use PPP dollars because they are the best way of measuring middle class values, in our view. From this point on in this report, we will tend to round down these per capita GDP numbers, as PPP figures themselves are not accurate enough to warrant spurious precision.

This risk lessens to 1.7% in the $15-25k range, and above $25k, no autocracy has ever transitioned to democracy.

2 See Thoughts from a Renaissance man: The war on inequality was won in 2008, published 17 March 2016.

3 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 1: The annual % chance of a country above a certain per capita GDP level moving from a negative (autocracy) to a positive (democracy) score, 2011 PPP dollars To democracy 6%

5%

4%

3%

2%

1%

0% 0 1,000 2,000 4,000 8,000 15,000 25,000

Source: Penn tables, Polity IV, Renaissance Capital

As the quotation marks in the title of this section suggests, there is a caveat which is that energy exporters don’t become democracies. Countries that are net oil exporters do not tend to tax their people, so they don’t give them a vote. Below we show all the world’s largest net oil exporters, and seven of the top eight are labelled autocracies. The exception is which was a democracy before it began exporting energy – and as we show below, rich democracies are immortal, even if they do discover oil.

Figure 2: Countries ranked by oil exports per capita, 2015 data 646 624 400 356 329 Left-hand scale capped at 400 350 313 300 When countries are large net 260 259 exporters of energy, then they 250 tend to be autocracies as 203 defined by Polity IV (2015) or Not Free 200 as defined by . The exceptions are those like 134 150 Norway that were rich democracies 93 before they exported energy. 100 79 79 68 63 60 58 54 53 50 30 29 25 20 18 14 12

Oil exports (bpd) 1,000per in people2015 7 5 0 Iraq (6) Iran (-7) UAE (-8) Chad (-2) (na) Gabon (3) Oman (-8) Russia (4) Algeria (2) Nigeria (7) Mexico (8) (-7) Qatar (-10) Angola (-2) Brunei (NF) (5) Norway (10) Canada (10) (7) Venezuela (4) Eq. (-6) Azerbaijan (-7) Trinidad Tob.…& Kazakhstan (-6) Congo, Rep. (-4) Turkmenistan (-8) Saudi Arabia (-10) Source: BP, IMF, Polity IV, Renaissance Capital

Below we group all the autocracies in the world in per capita GDP groups. Singapore is the single non-energy exporter in this group, and is often the exception in all these types of comparison. There is virtually a zero chance (historically) of autocracies transitioning to democracy if their per capita GDP is above $25k3.

In the $15-25k group are four oil exporters and Belarus, and as oil exporters tend to remain autocracies, the implication is that Belarus is the only country vulnerable to a shift to democracy (with a headline 2% chance each year). We are not sure Belarus is really this rich. Its nominal per capita GDP in current dollars is similar to which has a per capita GDP in PPP terms of $16k, or 75% of the Belarus level cited by the Penn tables. Belarus also has the same nominal per capita GDP as and Ecuador,

3 did, but only for one year, before returning to autocracy. Note we class it as an energy exporter due to its high production per capita.

4 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man which both have PPP GDP of $11k, roughly half the Belarus level. But Belarus is still subsidised by Russia by having access to cheap gas, and is a re-export conduit to sell sanctioned goods into Russia, which might distort the comparison with other countries.

The really interesting group is the next set of six countries, all of which are at maximum risk of transitioning to democracy with a 5% chance in any given year. One exception might be Uzbekistan; it is a very marginal oil exporter so might be at less risk of transitioning to democracy. Investors should be aware of the relatively high risk that Thailand, China, Jordan, Egypt and Swaziland transition to a positive Polity IV score in any given year.

Below that group are countries ranging from Morocco to Tajikistan with a 3% chance of transitioning each year. Morocco is the country which is on course to enter the $8-15k bracket in coming years where the risk of political transition (hopefully relatively peacefully, like ) should rise to 5%. Vietnam might, we imagine, prove more like China, Korea or Taiwan and end up taking longer to make that transition.

Figure 3: All countries with a negative Polity IV rating in 2015, ranked by per capita GDP (2011 PPP dollars), with % annual chance for countries in each income grouping to shift to a positive Polity IV rating in any given year (net oil exporters in red) 0% Polity IV (2015) Per cap GDP in 2014 (2011 PPP $) x x 3% Polity IV (2015) Per cap GDP in 2014 (2011 PPP $) Qatar -10 144,340 Angola -2 7,968 Singapore -2 72,583 Morocco -4 7,163 UAE -8 64,000* Laos -7 5,544 Kuwait -7 63,886 Vietnam -7 5,353 Saudi Arabia -10 48,025 Congo Brazzaville -4 4,426 Bahrain -10 41,626 Syria -9 4,200 Equatorial Guinea -6 40,133 Oman -8 38,527

2% Polity IV (2015) Per cap GDP in 2014 (2011 PPP $) 3% Polity IV (2015) Per cap GDP in 2014 (2011 PPP $) Kazakhstan -6 23,450 Sudan-North -4 3,781 Turkmenistan -8 20,953 Mauritania -2 3,409 Belarus -7 20,290 Tajikistan -3 2,747 Azerbaijan -7 15,887 Cameroon -4 2,682 Iran -7 15,547 Chad -2 2,013

5% Polity IV (2015) Per cap GDP in 2014 (2011 PPP $) 2% Polity IV (2015) Per cap GDP in 2014 (2011 PPP $) Thailand -3 13,967 Uganda -1 1,839 China -7 12,473 Rwanda -3 1,565 Jordan -3 10,456 Gambia -5 1,544 Egypt -4 9,909 -2 1,384 Uzbekistan -9 8,195 Ethiopia -3 1,323 Swaziland -9 8,029 Burundi -1 771

Unknown % chance Cuba -7 n/a Eritrea -7 n/a Afghanistan -1 n/a Korea North -10 n/a

*UAE per capita GDP is a Renaissance Capital estimate. Source: Penn tables, Polity IV, Renaissance Capital

Rwanda, Uganda, Burundi and Ethiopia have just a 2% chance of transitioning to democracy each year, but this will rise to 3% when their per capita GDP in 2011 PPP dollars reaches $2k. Successful economic development will undermine their leaders’ attempts to remain in power. The Robert Mugabe approach can help keep an autocrat in power, by forcing voters to concentrate more on putting food on the table than on middle- class issues such as taxation and representation. Since 1980, Zimbabwe’s per capita GDP in PPP terms has fallen from $2,496 to $1,867 in 2014. If Zimbabwe was classed as an autocracy (we address why it is not in the next section), this decline in GDP would have reduced the annual risk of democratisation from 3% to 2%.

We can see where these percentage chances come from by looking at previous countries that have transitioned from autocracy to democracy. The biggest outlier is Taiwan, which

5 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man in 1992 broke all records by shifting to democracy at $21,847, which at the time was far above the previous record of $13,619 set by in 1976. We recognise that in 1991 we might have been saying there is zero chance of Taiwan becoming a democracy, because it had never happened at that income level before. Since then only Venezuela has also achieved democracy in the same $15-25k range.

Figure 4: All countries which above $2k per capita GDP (2011 PPP dollars) have shifted from a negative to a positive Polity IV score

Autocracy to Democracy - GDP per Capita at time of transition (constant 2011 PPP $)

25,000

20,000

15,000 China today $12k 10,000 Egypt today $10k

5,000

0 Fiji (2014) Fiji (1990) Iran (1997) (1979) Peru (1993) Peru (1963) Peru (1956) (1989) (1985) Spain (1976) (1982) Gabon (2009) (1970) Turkey (1983) Turkey (1973) Algeria (2004) (1989) Bhutan (2008) (2012) Taiwan (1992) (1999) Tunisia (2011) (1974) (1990) Djibouti (1999) (1990) (1975) (1989) Ecuador (1979) Ecuador (1968) (1998) Panama (1955) (1989) (1985) (2007) Pakistan (1988) Thailand (2008) Thailand (1992) Thailand (1978) Thailand (1974) (1990) (1990) Colombia (1957) Suriname (1990) (1983) Argentina (1973) (1989) (1999) (1980) (1990) Mauritania (2007) Azerbaijan (1992) Venezuela (2013) Venezuela (1958) (1986) (1986) Guatemala (1966) (2005) (2000) Bangladesh (2009) Cape Verde (1991) Korea South (1987) (1978)

Source: Penn tables, Polity IV, Renaissance Capital

When do democracies become immortal?

We were confident that Turkey would not lose democracy on Friday 15 July because no country has (yet) lost a positive Polity IV rating above $15k and Turkey’s per capita GDP was $19k in 2014.

Figure 5: The annual % chance of losing democracy above key per capita GDP levels (2011 PPP dollars) To Autocracy 7% 6% 5% 4% 3% 2% 1% 0% 0 1,000 2,000 4,000 8,000 15,000 25,000 Source: Penn tables, Polity IV, Renaissance Capital

The only four countries to have ever lost democracy above $8k, were Greece ($8,022 in 1967), Thailand ($10,203 in 2006), Iran ($11,852 in 2004), Venezuela ($12,712) and top- ranked Thailand again ($13,967 in 2014). Note this is a big shift from our 2011 report, when the old (2005 PPP) per capita GDP series suggested Greece’s coup happened at a higher income level ($9,800 in 2005 PPP dollars) than Iran or Venezuela. Measuring PPP GDP over time is clearly something we must take with a pinch of salt.

6 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 6: The per capita GDP level (2011 PPP dollars) at which countries turned from positive to negative scores on Polity IV – above $1k only – countries can, and do, feature more than once in this chart Democracy to autocracy - GDP per Capita at time of transition (constant 2011 PPP $) 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 Fiji (2006) Fiji (1987) Iran (2004) (2000) Haiti (1991) Peru (1992) Peru (1968) Peru (1962) Chile (1973) Brazil (1964) (1963) (2002) Sudan (1989) Sudan (1970) Ghana (1972) Ghana (1981) Turkey (1980) Turkey (1971) Nigeria (1984) Nigeria (1966) Greece (1967) Gambia (1994) Panama (1968) Armenia (1996) Ecuador (1961) Uruguay (1972) Thailand (2014) Thailand (2006) Thailand (1991) Pakistan (1999) Thailand (1976) Thailand (1971) Pakistan (1977) Pakistan (1958) Comoros (1976) Comoros (1999) Suriname (1980) Argentina (1976) Cambodia (1997) Zimbabwe (1987) Mauritania (2008) Azerbaijan (1993) Venezuela (2009) Philippines (1972) Guatemala (1974) Guatemala (1954) Bangladesh (2007) Bangladesh (1974) Korea South (1972) Korea South (1961) (1971) Sierra Leone (1967) Guinea-Bissau (2003) Congo Brazzaville (1997) Congo Brazzaville (1963)

Source: Penn tables, Polity IV, Renaissance Capital

Note we can see there is a cultural or local aspect to these shifts between democracy and autocracy. We saw a wave of democratisation in eastern Europe in 1989-1990, and an attempt at similar in the Middle East and North Africa in 2011. Peru and Pakistan have often gained and lost democracy. Nigeria suffered more coups than most until former president Obasanjo made operational changes that weakened the army’s ability to conduct a coup.

Figure 7: Number of coups by country in Africa since 1950 (via French ) – note their definition of a coup is most probably NOT the same as Polity IV’s

The continent with 87 coups.

Number of coups since 1950, by country.

Source: Jeune Afrrique

7 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Thailand has a predilection for coups that stand in stark contrast to , which despite having low per capita GDP for generations, has remained broadly democratic. The graph below highlights how much some countries have changed since 1950 while India’s rating has remained largely steady.

Figure 8: Shifts in Polity IV ranking among five emerging/frontier markets since 1950

Not all countries are as politically calm as India

India Thailand Nigeria Pakistan Turkey

10 DEMOCRACY 8 6 4 2 0 -2 -4 AUTOCRACY -6 -8 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

But despite these cultural factors, economics appears to be the key determinant to whether coups succeed or not, and win out over cultural preferences in the end. Since a British ambassador said 100 years ago “I have been in Russia, Berlin, Constantinople and Persia which are all popularly supposed to be autocratic governments”, the only country to become a full democracy is the one (Germany) that is far wealthier than the others.

Refining the definition of democracy and autocracy

Everything written above is using a very crude definition of democracy and autocracy. These are not the definitions that Polity IV itself would use. We have ended up irritating a number of readers who argue that countries like Russia, Turkey or Zimbabwe are not really democracies, and Polity IV definitions support that criticism. Polity IV would define all three as ‘open anocracies’, compare with ‘closed anocracies’ like Egypt, Morocco, Singapore or Thailand, or autocracies like China and the UAE. So below, we start again, with a more nuanced approach, where ‘democracy’ means a score of 6 to 9 as in the Polity IV definitions below:

Full democracy = 10

Democracy = 6 to 9

Open anocracy = 0 to 5 (ie a system which combines democratic and autocratic elements with more semblance of democracy like elections)

Closed anocracy = -1 to -5 (ie a system which combines democratic and autocratic elements with more autocratic elements)

Autocracy = -6 to -10

8 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

When we categorise all the countries in MSCI Developed, Emerging and Frontier markets by their political regime, we are not surprised to see that all the autocracies are oil exporters, except for lower-income China and Vietnam.

Figure 9: Polity IV political ratings of developed (red), emerging (blue), frontier (green), beyond frontier (yellow) in 2015 with 2014 per capita GDP in $000 (2011 PPP) Autocracy xx x Closed anocracy xx x Open anocracy xx x Democracy xx x Full democracy xx Azerbaijan $15.9 -7 Angola $8.0 -2 Bangladesh $2.9 1 Argentina $20.2 9 $43.1 10 Bahrain $41.6 -10 Egypt $9.9 -4 Ivory Coast $3.4 4 $43.7 8 $47.7 10 China $12.5 -7 Ethiopia $1.3 -3 Malaysia $23.2 5 Botswana $16.2 8 Canada $42.4 10 Iran $15.5 -7 Jordan $10.5 -3 $1.1 5 Brazil $14.9 8 Chile $21.6 10 Kuwait $63.9 -7 Morocco $7.2 -4 $2.2 3 Bulgaria $17.5 9 $44.9 10 Kazakhstan $23.5 -6 Thailand $14.0 -3 Russia $24.0 4 Colombia $12.6 7 $40.4 10 Oman $38.5 -8 Rwanda $1.6 -3 Turkey $19.2 3 Croatia $22 9 Germany $46.0 10 Qatar $144.3 -10 Singapore $72.6 -2 $10.3 4 Czech Rep $31.9 9 Greece $26.0 10 Saudi Arabia $48.0 -10 Uganda $1.8 -1 Venezuela $14.1 4 $28.5 9 Hungary $25.8 10 UAE $64? -8 Zimbabwe $1.9 4 $39.4 9 Ireland $48.8 10 Vietnam $5.4 -7 Ghana $3.6 8 $35.8 10 $9.4 7 $35.4 10 *RenCap estimate India $5.2 9 $28.2 10 Indonesia $9.7 9 $17.9 10 Iraq $12.1 6 $34.7 10 $33.3 6 Norway $64.3 10 $7.4 9 $47.2 10 $2.8 9 Poland $25.2 10 $14.0 6 Portugal $28.5 10 Mexico $15.9 8 $44.6 10 Nigeria $5.5 7 Switzerland $58.5 10 Pakistan $4.6 7 Taiwan $44.3 10 Peru $11.0 9 $30.5 10 Philippines $6.7 8 Spain $33.9 10 Korea South $35.1 8 UK $40.2 10 Romania $20.8 9 USA $52.3 10 $12.1 9 $13.4 8 $10.3 6 Tunisia $10.4 7 $3.7 7

Source: Politity IV, Penn Tables, MSCI, Renaissance Capital

We can see the benefits of EU membership, which has helped propel six EU member states of the eight EM and FM markets into the ‘full democracy’ category, which is otherwise dominated by rich developed markets.

We can all of course argue with the rankings. We are surprised that Hungary still has a score of 10 and a ‘full democracy’ rating, after reading Anton Shekhovstov’s piece Is Transition Reversible? The Case of Central Europe, published 19 January 2016. So a little further below we look at all those countries that have ever lost that perfect 10 score.

Anocracies are inherently unstable

As countries get richer, they all end up becoming democracies, unless they are oil exporters or are a city state called Singapore. The table below includes all the data points we have since 1950. One interesting aspect in the table below is that ‘open anocracies’ become less frequent as income levels rise, and cease to exist once countries hit the $25k per capita GDP level. This tells us that the 10 countries in the ‘open anocracy’ list (see Figure 9 above) have inherently unstable regimes that will change over time, or that they will get stuck in the middle income trap and fail to grow.

9 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 10: The at each per capita GDP level over 1950-2015 – ‘anocracies’ get squeezed out of existence as wealth levels rise (brackets show number of countries at each income level since 1950)

Autocracy (-10 to -6) Closed Anocracy (-5 to -1) Open Anocracy (1 to 5) Democracy (6 to 9) Full Democracy (10) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% $00-$999 (607) $1,000-$1,999 (1,411) $2,000-$3,999 (1,402) $4,000-$7,999 (1,384) $8,000-$14,999 $15,000-$24,999 $25,000+ (819) (1,141) (766)

Source: Penn tables, Polity IV, Renaissance Capital

Very few countries have ever lost the ‘full democracy’ label

No country with a ‘full democracy’ score of +10 has, since 1950, subsequently had a negative score. However, two countries (France, Malaysia) have lost ‘democracy’ and become an ‘open anocracy’ for at least a decade. Another four have seen their rating fall from ‘full democracy’ to ‘democracy’. Curiously none of these six countries has yet recovered its ‘full democracy’ score once it has been lost.

Figure 11: The % annual chance of a ‘full democracy; losing its (2015) +10 score above a per capita GDP level (2011 PPP dollars), with key countries per capita GDP levels on the right (rounded to the nearest thousand, 2014 data) To To closed To open To To To Data Key countries and their autocracy anocracy anocracy democracy occupation failure points per capita GDP in 2014 0 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0 1,000 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0 2,000 0.0% 0.0% 2.9% 0.0% 0.0% 0.0% 35 4,000 0.0% 0.0% 0.0% 0.6% 0.0% 0.0% 158 8,000 0.0% 0.0% 0.3% 0.3% 0.0% 0.0% 347 15,000 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 471 Mauritius 18, Chile 22 Poland 25, Hungary 26, Greece 26, 25,000 0.0% 0.0% 0.0% 0.3% 0.0% 0.0% 607 Lithuania 28, Slovenia 30, Taiwan 44

Source: Penn tables, Polity IV, Renaissance Capital

The two richest countries to lose ‘full democracy’ were Belgium, with its score slipping from 10 to 8 in 2007 and the slipping to 9 in 2006. Both have remained there since. Israel slipped to 9 in 1967 and then to 6 in 1981, where it has remained since. Jamaica edged marginally down to 9 in 1993 (with a per capita GDP of $5.2k) where it has remained since – and this country is interesting, as it is perhaps the poorest to have sustained a full democracy for as long as it did (1959 to 1992).

France, during the rule of Charles de Gaulle, fell from 10 to an ‘open anocracy’ score of 5 in 1958 (per capita GDP, $9.5k) until 1969, which is relevant today given Erdogan’s ambition to create a similar system in Turkey. It then returned to 8 where it remained until 1986 when it rose again to 9.

Malaysia collapsed from 10 to 1 in 1969 which is the deepest fall of any ‘full democracy’ since 1950 but occurred at a per capita GDP level of just $3k – the lowest level where change occurred of any in this table. The score rose back to 4 in 1971, slipped to 3 in

10 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

1995, returned to a ‘democracy’ score of 6 in 2008 then lost ‘democracy’ again with a slip to 5 in 2014. We thought Malaysia might be an interesting example for those looking at Russia, as it was an oil exporter. However, it only exported more than 100k b/d from 1979-2005, and at its peak in 1989, it exported less than half of what Russia exports on a per capita basis today. Curiously though it did regain a democracy ranking in 2008, just three years after it stopped being a net exporter of oil.

For the sake of comprehensive data disclosure, Vichy France, the regime imposed by Nazi Germany in 1940 gets a score of -9 after being +10 in 1939. We don’t have comparable per capita income data (it was surely less than in 1958), and don’t believe this example is very useful to investors. Of course we cannot entirely rule out democracies being conquered by neighbouring autocratic states, but this piece is focused on domestically driven regime change, not externally imposed regime changes.

Conclusion

Mauritius and Chile will remain full democracies until per capita GDP reaches $25k.

Of Poland, Hungary, Greece, Lithuania, Slovenia and Taiwan there is a 0.3% chance each year they are downgraded to ‘democracy’ status – echoing what we previously saw in Belgium and the Czech Republic. We see Hungary and Poland as being the most likely candidates of the six to experience this. We doubt any such downgrade would have great market significance, and doubt the small sample of similar shifts noted above would be a reliable guide.

‘Democracy’ scores have two-way risk

The status quo is very dominant, but if there is change then deterioration is more common than a shift to ‘full democracy’, particularly at low income levels

We have 1,700 data points of countries with a political score of +6 to +9. Overall we have seen 14 shifts to full democracy and 48 shifts to a less democratic system. Losing democracy is more common than reinforcing it, except at the highest income levels. Below $8k of per capita GDP we see numerous examples of countries becoming an autocracy within a year. Only above an income of $2k do democracies have a chance of becoming a full democracy. There is two-way risk in democracies.

It is worth noting how few data points can change the percentages below. At an income level of $15-25k, we have seen Russia, Turkey and Malaysia shift from democracy to become ‘open anocracies’, while Chile and shifted from ‘democracy’ to ‘full democracy’. So it appears that if for example Mexico changes regime, it is more likely to follow Russia than Chile at this income level. We think that percentage difference is in fact trivial. Just one more country joining Chile and Slovakia’s example, would be enough to even those ratios out.

Figure 12: The % annual chance of a ‘democracy’ losing its (2015) +6 to +9 score above a per capita GDP level (2011 PPP dollars), with key countries per capita GDP levels on the right (rounded to the nearest thousand, 2014 data) To To closed To open To full To To Data Key countries and their autocracy anocracy anocracy democracy occupation failure points per capita GDP in 2014 0 3.6% 1.2% 1.2% 0.0% 0.0% 0.0% 83 1,000 3.7% 0.4% 1.2% 0.0% 0.0% 0.8% 245 2,000 1.1% 0.4% 1.8% 0.7% 0.4% 0.4% 275 Kenya 3, Ghana 4, Zambia 4 4,000 0.7% 0.9% 1.4% 0.7% 0.0% 0.2% 444 India 5, Pakistan 5, Nigeria 6, Jamaica 7, Philippines 7 Georgia 9, Indonesia 10, Sri Lanka 10, Tunisia 10, Peru 11, Iraq 12, 8,000 0.0% 0.5% 0.7% 1.5% 0.0% 0.0% 411 SA 12, Colombia 13, Serbia 13, Lebanon 14, Brazil 15 15,000 0.0% 0.0% 1.9% 1.3% 0.0% 0.0% 154 Botswana 16, Mexico 16, Bulgaria 17, Argentina 20, Romania 21, Croatia 22 25,000 0.0% 0.0% 0.0% 1.1% 0.0% 0.0% 88 Estonia 29, Czech 32, Sth Korea 35

Source: Penn tables, Polity IV, Renaissance Capital

11 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

In terms of history, we are not going to dwell on very low ($0-2k) income countries – where there is a 4% chance each year of a democracy becoming an autocracy, because no democracies we follow are at that income level.

The six instances of coups between $2k and $8k are Argentina (1976), Chile (1973), Ghana (1981), Nigeria (in 1966 and in 1984 due to General Muhammadu Buhari’s coup) and Pakistan (1999).

No country has gone from ‘democracy’ straight to ‘autocracy’ with a per capita GDP above $8k.

There are nine instances of democracies at any income level shifting pretty significantly to closed anocracies, and six of these are in the $4-15k per capita GDP range. These are Fiji (1987, 2006), Turkey (1971, 1980) and Thailand (2006, 2014). Given the double coup events in each – and the failed attempt in Turkey in 2016 – we must accept the notion that local specific factors do make some countries more susceptible than others to coups.

There are 21 instances of democracy erosion, where ‘democracies’ become ‘open anocracies’. Of these, 12 show only a marginal fall from a democracy score (between 6 and 9) to the highest possible ‘open anocracy’ score of 5.

We think EEMEA investors are acutely aware of democracy erosion in Russia at $18k (2007) and Turkey (at $4k in 1954 and again at $19k in 2014) which have both lost ‘democracy’ and become ‘open anocracies’, while Malaysia ($23k in 2014) did this at the highest income level yet seen for this change. Until a decade ago, no ‘democracy’ had slipped back to ‘open anocracy’ in this $15-25k range.

The other 18 instances, at varying income levels, include countries as diverse as Sri Lanka (1982, 2003, 2009 – in each case from a score of 6 to 5), Ukraine (1993, 2014) and Brazil (1961). In a later section, we look at what happens subsequently to these countries. We do not have income data, but this loss of democracy to ‘open anocracy’ has also been seen in Portugal (1926), Finland (1931), Estonia (1934), Brazil again (1947) and Greece (1949). We think all were poorer than Russia or Turkey today.

There are 14 instances of a shift to a strong democracy, with most of these above the $8k mark. They range from Cape Verde to Uruguay. There are also four cases of democracies becoming state failures, Cyprus at $5k in 1963, the Dominican Republic at $3k in 1963, at just under $2k in 1998 and at $1k in 2012. Singapore is counted as an ‘occupation’ due to its temporary merger with Malaya in 1963. State failures and occupations are extremely rare events.

Democracy conclusions

$2-4k countries: There has historically been a 5% chance of a shift in democracies, including a 3% chance of a shift to a less democratic system and less than a 1% chance of a shift to ‘full democracy’. There is more downside than upside risk in Kenya, Ghana and Zambia. But our 95% base case in any given year should be continued ‘democracy’.

$4-8k countries: Within the 4% chance of overall change, there is a 3% chance of a shift to a less democratic regime and just under a 1% chance of a shift to a more democratic regime. There is more downside than upside risk in India, Pakistan, Nigeria, Jamaica and the Philippines. We would note their individual histories, which suggest India and Jamaica are least vulnerable, while Pakistan is most vulnerable. Nigeria’s recent long run without a coup suggests it may have altered its institutions enough to pull away from being in the Pakistan camp. Note the above data suggest there is a tiny 0.3% risk of either Nigeria or Pakistan becoming a failed state. This is considerably less than news

12 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man coverage might lead you to expect. The Philippines may be a candidate for a shift to open anocracy under its new president.

$8-15k countries: There is low, two-way, risk in this group, and we can highlight that a sudden shift to autocracy would be unprecedented. These countries are marginally more likely to become full democracies than anocracies. Some are concerned that SA will become like Zimbabwe (an open anocracy) – but there is a less than a 1% chance of this each year. The 97% base case is no change each year.

$15-25k countries: Argentina, Botswana and Mexico are at the income level where Russia, Turkey and Malaysia stood before falling back to open anocracy status, and where Chile and Slovakia were before reaching full democracy status. The base 97% chance each year is they stay just where they are. The precedent of other EU member states suggests Bulgaria, Romania and Croatia have a better chance of eventually achieving full democracy than the headline figure in our table.

$25k+ countries: We believe Estonia, the Czech Republic and will either remain as democracies or achieve full democracy. We have yet to see a country like the Czech Republic which once had full democracy, regain it.

Open anocracies – what happens next to Russia and Turkey?

History tells us ‘democracies’ are more vulnerable to moving away from, rather than towards, ‘full democracy’. So at first glance, it is a little surprising that for ‘open anocracies’ the reverse is true. There are 56 instances of shifts to democracy (55 to democracy, 1 to full democracy) and only 38 of a deterioration (17 to closed anocracy and 21 to autocracy). More interesting still, we have never seen an open anocracy above a per capita GDP level of $25k. However, until Russia, Malaysia and Turkey broke some data records in the past decade, we had never seen a democracy at $15-25k become an open anocracy either. These three are already trailblazers.

At any given income level, a shift to ‘democracy’ is more likely than a change to ‘closed anocracy’ or to ‘autocracy’. If we add up the latter two, we can say that democratisation (a shift to the right in our table) only becomes more likely than a shift away from democracy (to the left of the table) at a per capita GDP level above $4k.

Ukraine is nearly 10x more likely to become a democracy than become a closed anocracy, and 4x more likely to become a democracy than become an autocracy. We are inclined to think the same is true of Venezuela too, but have to caveat this as it is an oil exporter (see the section on Russia below).

Figure 13: Open anocracies - the % chance of change above per capita GDP levels (2011 PPP dollars) To To closed To To full To To Data Key countries and their autocracy anocracy democracy democracy occupation failure points per capita GDP in 2014 0 4.8% 2.4% 9.5% 0.0% 0.0% 1.2% 84 1,000 2.6% 3.9% 5.2% 0.0% 0.0% 2.6% 153 Mozambique 1, Zimbabwe 2 2,000 4.5% 1.8% 4.9% 0.0% 0.0% 0.9% 223 Tanzania 2, Bangladesh 3, Ivory Coast 3 4,000 0.5% 2.2% 8.6% 0.0% 0.0% 0.5% 186 8,000 1.7% 0.9% 8.6% 0.9% 0.0% 0.0% 116 Ukraine 10, Venezuela 14 15,000 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 20 Turkey 19, Malaysia 23, Russia 24 25,000 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0

Source: Penn tables, Polity IV, Renaissance Capital

Evidently, the four African countries in the above table, and Bangladesh, are in a different place. Change in political regime at this low income level is surprisingly common. Mozambique and Zimbabwe are not only more likely to slip towards closed anocracy and autocracy than to become democracies, they also carry a non-negligible 2.6% (1 in

13 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

40) annual risk of becoming failed states. Previous examples of ‘open anocracies’ becoming failed states at this income level were Comoros in 1995, Guinea-Bissau in 1998, Haiti in 2010 and Sierra Leone in 1997.

We suspect Mozambique is at greatest risk due to its low levels of education. Only around one-fourth of secondary school age children (11-17), attend secondary school. We find that this not only dooms a country to poverty for the next generation, unless that country has high levels of mineral wealth (Mozambique luckily has gas), but it is also a useful predictor of political crises. Please see the footnote for more information4.

Figure 14: Secondary school gross enrolment (rhs) vs nominal 2016 per capita GDP in $ (lhs), and energy exporters in pink Per capita GDP in 2016 ($) vs gross secondary school enrolment in 2011 (rhs) Per capita GDP (2016) Gross secondary school enrolment 2011 3,500 40 Until 2031, we should expect plenty of bad news (coups, civil unrest, default and poverty) from countries that did not 3,000 35 educate 25-28.5% of children at secondary school in 2011. Energy exporters (marked in pink) may get richer, but 30 2,500 this may benefit elites more than most 25 2,000 20 1,500 15 1,000 10 500 5 0 0 Eritrea Angola Uganda Burundi MozambiqueBurkina Faso Chad MauritaniaCentral African Rep. South Sudan (att)Somalia (2007)

Note: Somalia GDP is a RenCap estimate Source: BP, Unesco, IMF, Renaissance Capital

These figures do not mean that if you are doing a 10-year DCF, you should assume a 26% chance of Zimbabwe becoming a failed state and writing down the value of your company to zero. We would normally assume that per capita GDP will tend to rise over 10 years, meaning Zimbabwe or others should gradually move up the per capita GDP ladder, so failed state risk should disappear. But Zimbabwe is actually a bad example for us to pick. In 2014, Zimbabwe’s per capita GDP was 29% below its 1980 per capita level. As noted earlier in this report, it has dropped down the ladder since Mugabe came to power, not risen.

Tanzania, Bangladesh and Ivory Coast have a 5% chance of becoming a democracy each year, but are marginally more likely to shift towards a closed anocracy or an autocracy. We think a 1 in 10 chance of significant political change (roughly half and half between positive or negative) means we need to watch local politics carefully.

4 See pages 17-19 of Fixit, published on 7 June 2016.

14 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 15: Polity IV scores have been volatile for four of the five poorest open anocracies we look at (Mozambique has not changed in 20 years) The volatility of low income open anocracies Mozambique Zimbabwe Tanzania Bangladesh Ivory Coast 10 DEMOCRACY 8 6 4 2 0 -2 -4 -6 AUTOCRACY -8 -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Note: Pre-1972 data for Bangladesh use Pakistan (West and East) data, a score below -10 suggests civil war. Source: Polity IV

The data purport to show that Turkey, Malaysia and Russia can stay where they are or have a 10% chance annually of becoming democracies. As we have never seen an ‘open anocracy’ with a per capita GDP of $25k or more, the data imply that the political status quo is incompatible with higher wealth levels for these countries.

Unfortunately, the data is insufficient to justify confidence in the above statement. There are just 20 data points for open anocracies at this income level, of which most are Russia and Malaysia themselves (the others are France in 1969, Gabon in 2011-2012 and Suriname in 2012-2014). Within this, there are two examples of countries becoming democracies from open anocracy, France in 1969 and Malaysia in 2008 (before Malaysia then fell back to open anocracy).

Nonetheless, we think Malaysia may repeat its democratising 2008 feat. If the governing coalition is defeated at the August 2018 (after effectively being in power since 1955), then Malaysia may well return to a ‘democracy’ category ranking. The trigger may be the current corruption scandal and would amusingly fulfil the Rahman prophecy, which suggests the current prime minister will be the last ruler from the Barisan Nasional (BN).

In Turkey, we believe that it is President Erdogan who is primarily responsible for Turkey’s ‘open anocracy’ ranking. We believe that Turkey’s electorate is the reason that Turkey has not seen a larger deterioration in its Polity IV score. Turkish voters refused to give the governing AKP a big enough victory to change the constitution in either of the 2015 parliamentary elections. We think Turks will resist a shift to a closed anocracy or autocracy, which some fear would come from Erdogan amassing more power, and our data show no country has let this happen at Turkey’s per capita GDP level. The reaction of the and opposition parties to the 15 July coup attempt is additional proof of modern Turkey’s resistance to autocracy. Without Erdogan, we believe Turkey could regain a ‘democracy ranking’. In the meantime, we assume it remains an open anocracy – and may remain in a middle income trap, which we talk about below. It will be unprecedented if Turkey at this income level is pushed into a ‘closed anocracy’ or ‘autocratic’ rating by Erdogan’s post-purge moves.

We, and many others, see similarities between Russian and Turkish politics. Without President Putin at the helm, we think Russia can again become a democracy as it was in 2000-2006. It is rich enough to avoid a further shift towards autocracy. While Putin is in place, we assume Russia stays an open anocracy, although its score of 4 might fall to Turkey’s level of 3 or a little lower. The parliamentary elections in September 2016 may well influence the score.

15 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

A new democratisation wave for oil exporters?

What makes Russia more complicated to analyse is that it is a net oil exporter. It does not export enough oil to sustain an autocratic ranking like the Gulf states, nor does it export so little oil that democracy is inevitable in the long -run (Nigeria, Mexico, Colombia). Russia was never a rich democracy before exporting oil – unlike Norway or Canada5 – which might have protected it from becoming an open anocracy. Russia did democratise when oil prices last fell, and that example is worth looking at below.

A number of countries export a similar amount of oil per capita to Russia (see Figure 2). Below we look at five net oil exporters which were low or middle income in 1965 – Algeria ($4.6k), Iran ($2.6k), Mexico ($7.6k), Venezuela ($7.6) and Russia (using estimates for the Soviet Union, we estimate its per capita GDP was around $10k).

As the oil price declined after 1980, and fell to low levels after 1985, most became less autocratic. For example, Mikhail Gorbachev became leader of the Soviet Union in 1985 and introduced policies of glasnost (political openness) and perestroika (economic reform). The democratisation in these countries would not have just been a function of the oil price – countries are anyway likely to democratise due to rising per capita GDP but oil exporters are often an exception to this rule. Since the oil price began rising again both Iran and Russia have seen a reversal of their Polity IV direction. Venezuela has seen a significant deterioration in its score relative to the 1990s. There has been no improvement in the Polity IV scores of Mexico or Algeria either.

Now that the oil price has fallen from over $100/bl to around $45/bl, some of the countries that export oil may be forced to tax their people more or cut back on government spending. As a result, they may feel forced to grant more political rights to their . After oil prices fell sharply in 1985, we saw Mexico’s Polity IV score shift significantly three years later in 1988, Algeria’s in 1989 while Russia moved significantly over 1988-1990. The oil price in 2015 fell similarly to 1985, implying change in Algeria, Iran or Russia in 2018-2020.

Figure 16: Oil in 2015 prices compared to Polity IV scores for countries that export some (but not Gulf levels of) oil per capita

Oil price (lhs) and democratisation Oil price (2015 prices, lhs) Polity IV score (USSR, Russia) Polity IV score (Mexico) Polity IV score (Iran) Polity IV score (Algeria) Polity IV score (Venezuela) 140 -10 -8 120 Autocracy -6 High oil 100 -4 -2 80 0 60 2 Low oil 40 4 6 20 Democracy 8 0 10 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017F

Source: BP, Polity IV

5 If was larger, we might consider this as an interesting comparison point for Russia. It was an oil exporter when it became independent with a democracy score of +8 and per capita GDP of $8k in 1962. Its score rose to +9 at $17k in 1984, and to a full democracy score of +10 at $10k in 1997.

16 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Curiously, Moscow’s economic liberals see Mexico as a comparison point for them when it comes to inflation targeting and managing the oil/manufacturing mixed economy6. Comparison with Mexico suggests Russia will become more democratic than it is today.

Open anocracies such as Russia and Turkey – and the middle income trap

Russia will be in unprecedented territory if it can grow its per capita GDP in PPP terms above the $25k level, while remaining an open anocracy. But this may be impossible. We may discover in coming years that if a country is not sufficiently politically open then it cannot escape this $15-25k middle income trap (unless it exports Gulf levels of oil per capita or is called Singapore). Reinforcing that view is the likely fall in Russian per capita GDP in PPP terms that we expect to see in 2015 and 2016 data.

Perhaps open anocracies must change politically or they can’t grow, as it may be that growth beyond $15-25k requires an independent judiciary, political openness and freedom. If this is right, then Russia has the choice of economic stagnation or political change. Stagnation is implicit in consensus economic forecasts suggesting Russia will grow at 1.5-2.0% over the long term. We have yet to meet anyone who believes Russia can achieve the sustainable rate of 4% growth that Russia’s economy ministry would like to see (unless oil prices soar of course). There is also a strong consensus that Russia will stagnate politically. No-one we know expects President Putin to retire before 2024.

What we know then is the following

1) There are no open anocracies above $25k of per capita GDP. The implication is that Russia, Turkey and Malaysia do not have a sustainable political and economic model.

2) In the $15-25k range, open anocracies either see their incomes fall, or they stagnate politically and economically, or they become democracies. We see Malaysia as particularly likely to democratise. We think consensus assumes Russia stagnates politically and economically. We believe this is also a risk for Turkey. 3) Russia is an oil exporter and many countries (including Russia itself) became more democratic after oil prices fell in 1985. Because oil fell similarly in 2014-2015, we may see a new wave of democratisation in 2018- 2020, in Algeria, Iran, Russia and Venezuela. 4) In any given year, there is a 90% chance of no change, based on the (inadequate) data we have.

Closed anocracies – tend to get more democratic

For this table, we have stripped out oil exporters, which cuts our total sample from 934 data points to 812 data points, but we think this offers a better guide to all the countries we focus on except for oil exporting Angola.

6 See Thoughts from a Renaissance man: Russia - 4% CPI before Putin’s fourth term? 23 July 2015.

17 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 17: The % annual chance of an ‘closed anocracy’ shifting its (2015) -1 to -5 score, above certain per capita GDP levels (2011 PPP dollars), with key countries per capita GDP levels on the right (rounded to the nearest thousand, 2014 data) – the data in this table excludes oil exporters To To open To To full To To Data Key countries and their autocracy anocracy democracy democracy occupation faliure points per capita GDP in 2014 0 7.1% 3.6% 0.0% 0.0% 0.0% 1.8% 112 1,000 4.2% 1.7% 1.7% 0.0% 0.0% 0.7% 286 Ethiopia 1, Rwanda 2, Uganda 2 2,000 4.6% 6.1% 1.5% 0.0% 0.0% 0.4% 262 4,000 3.6% 4.3% 5.0% 0.0% 0.0% 0.0% 139 Morocco 7, Angola 8 8,000 0.0% 10.1% 2.2% 0.0% 0.0% 0.0% 89 Egypt 10, Jordan 10, Thailand 14 15,000 0.0% 4.2% 4.2% 0.0% 0.0% 0.0% 24 25,000 4.5% 0.0% 0.0% 0.0% 0.0% 0.0% 22 Singapore 73

Source: Penn tables, Polity IV, Renaissance Capital

Closed anocracies are pretty common at around $1-4k of per capita GDP, but largely disappear by the time per capita GDP reaches $15k. All 21 data points for a closed anocracy above $25k are accounted for by exceptional Singapore, and it also accounts for 11 of the 16 data points in the $15-25k range. Taiwan accounts for the other five – it then shifted to democracy in 1992 at $22k.

Our base case assumption is that all closed anocracies with a per capita GDP above $8k will become more democratic, unless they export oil or are called Singapore.

This is particularly important for investors in Egypt, Jordan and Thailand which are currently in the $8-15k range. Non-oil exporters that have shifted from closed anocracy to open anocracy are Portugal (1975), Spain (1976), South Korea (1987), Hungary (1989), Croatia (1999), Thailand (2008) and Tunisia (2011). We believe this was simply a function of rising per capita GDP and a middle class demanding more political rights.

Excluded from our table are the oil exporters that have also made the same shift at this income level: Mexico (1988), Gabon (2009) and Algeria (2004). We have one example of a country shifting all the way to democracy, which was Turkey (1983). Overall, excluding the oil exporters, there is a 1 in 8 chance each year of a significant shift towards democracy at this income level. This surprises us, but means we should expect democratic change in Thailand, Jordan and Egypt within a decade. None is likely to become an autocracy.

One income bracket down at $4-8k is Morocco – currently the second-largest market in MSCI Frontier. There is a 5% chance of shifting to democracy, as seen in Brazil (1985), Chile (1989), Dominican Republic (1978), Ecuador (1979), Indonesia (1999), Thailand (1992) and Turkey (1973). There is also a 5% chance of a shift to open anocracy, as seen in Bhutan (2008), Fiji (1990 and 2014), Romania (1990) and Suriname (1990), and also in oil exporting Azerbaijan (1992). Even excluding Azerbaijan, this gives us a 1 in 10 chance of a significant shift towards democracy each year. There is just a 3% chance of a shift to autocracy, as seen in Panama (1984), Suriname (1982) and Uruguay (1973). We believe Morocco has a good chance of making a relatively peaceful transition to a more democratic regime, thanks to the relatively liberal approach of the current ruler.

Angola is also in this bracket, but as an oil exporter, its future is likely to be different. We have seen shifts to autocracy in Algeria (1992 at $7.8k) and Kazakhstan (2002 at $7.1k) at around this income level. Azerbaijan initially became more democratic as it left the Soviet system but in 1995 Azerbaijan also shifted to autocracy (with a lower income of $2.3k). A relatively low level of education in Angola (see Figure 14) also encourages us to assume a shift to autocracy is more likely than a shift towards democracy – although of course, no change at all is by far the most likely annual event.

Lower down the income scale we have Ethiopia, Rwanda and Uganda. The base case should be to assume no change, but each is marginally more likely to become autocratic at this income level. Rwanda recently changed its constitution to allow the president to run for a third term in the 2017 elections. Previous examples of autocratic shifts include Benin

18 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

(twice), Bangladesh (twice), Boliva, Chad, Mauritania, , and Syria (in 1963). Shifts to ‘open anocracy’ were seen in (2001), Comoros (2001), Ghana (1996), Guinea (2010) and Haiti (2004). Shifts to democracy at this income level were experienced by Bangladesh (1991), Guinea-Bissau (2005), Kenya (2002), Thailand (1969), Zimbabwe (2009), South Korea (1960) and (2008).

Yes, Ethiopia and Rwanda are roughly 55 years behind South Korea in terms of per capita GDP in PPP terms.

However, both Rwanda and Uganda (and soon Ethiopia given its rate of growth), will enter the $2-4k range, which will lead to a 6.6% chance annually of a shift towards open anocracy with plenty of examples ranging from Ghana in 1970 to Ivory Coast in 2000. Pakistan, Thailand and Guatemala did this twice each and Peru did this transition three times at this income level. These countries with multiple occurrences suggest too much political change may hold back growth and keep a country stuck in GDP range (an idea we will not explore further in this piece). Full shifts to democracy were seen in Pakistan (1998), Nicaragua (1990), Cape Verde (1991) and Colombia (1957). A shift towards democracy becomes more likely still in the $4-8k range.

Autocracies – oil exporters, China and Vietnam

We’re in interesting territory already with the rich (>$25k) energy exporting autocracies, who are virtually immortal autocracies based on the data we have so far. Amidst 128 data points for six countries, all were autocratic except when Bahrain tested the waters of closed anocracy in 2010, before rushing back to autocracy in 2011. Until 2010, we would have said there is no chance of any change occurring.

Figure 18: The % annual chance of an ‘closed anocracy’ shifting its (2015) -1 to -5 score, above certain per capita GDP levels (2011 PPP dollars), with key countries per capita GDP levels on the right (rounded to the nearest thousand, 2014 data) To closed To open To To full To To f Data Key countries and their anocracy anocracy democracy democracy Occupation aliure points per capita GDP in 2014 0 3.7% 1.5% 0.3% 0.0% 0.6% 0.9% 328 1,000 1.9% 1.0% 0.7% 0.0% 0.0% 0.1% 727 2,000 3.5% 1.3% 0.2% 0.0% 0.2% 0.0% 607 4,000 3.1% 0.2% 0.9% 0.0% 0.0% 0.2% 457 Vietnam 5 8,000 2.8% 1.1% 0.6% 0.0% 0.0% 0.0% 178 China 12 15,000 1.0% 0.0% 0.0% 0.0% 1.0% 0.0% 97 Iran 16, Azerbaijan 16, Kazakhstan 23 Oman 39, Bahrain 42, Saudi Arabia 44, 25,000 0.8% 0.0% 0.0% 0.0% 0.0% 0.0% 127 UAE 64*, Kuwait 64, Qatar 144 Note: UAE per capita GDP is a Renaissance Capital estimate. Source: Penn tables, Polity IV, Renaissance Capital

It is possible that we could see some marginal shift towards democratisation among Gulf states. The reforms now under way7 – reducing subsidies, encouraging locals to take on more work at the expense of expatriates – might reduce the ‘bribe’ factor of government spending. Also, since the Arab Spring, some may argue that those who have a political stake in the system will do more to keep it alive, than those excluded from the system, so some democratisation might help. But 127/128 data points tell us we should assume no change in rich Gulf states unless the world stops using oil.

Among the oil exporters like Iran, Kazakhstan and Azerbaijan, data for all autocracies suggest there is only a 1% annual chance of a shift to a closed anocracy. We think this may be reasonable for Kazakhstan and Azerbaijan, but this understates the chances of Iran shifting. Iran exported a net 2mn b/d in 2015. If it lifted this to 3mn b/d, it would still export less per capita than open anocracies like Russia and Venezuela, or democracies like Canada and Trinidad and Tobago, or closed anocracies like Angola and the Republic

7 See Thoughts from a Renaissance man: A wave of reform by oil exporters, published 28 April 2016.

19 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man of Congo. It is a more diversified economy than the autocracies that export more (or indeed) less oil than Iran does. Because Iran exports relatively little oil per capita, we think there is a higher chance of a democratisation shift than in Kazakhstan or Azerbaijan.

China and Vietnam

Taiwan is the only example of a significant political change in any relatively rich (>$15k) autocracy since 1970, except for Bahrain in 2010 and the occupation of Kuwait by Iraq. In 1987, at an income level of $16.2k, roughly 30% above China today, the Kuomintang chose to turn Taiwan into a closed anocracy,followed five years later by an unprecedented shift to democracy.

There are many more examples above $8k.There was the fall of the military dictatorships in Spain and Portugal (both to closed anocracy) and Greece (to open anocracy) in 1974- 1975. There was the fall of communist dictatorships once the Soviet Union stopped propping them up, in Bulgaria and Czechoslovakia8 (straight to democracy in 1990), Hungary (to closed anocracy in 1988) and Poland (to open anocracy in 1989). The only two examples not cited yet were Mexico (1977) and Gabon (1991), both to closed anocracies.

For China the most relevant model is probably Taiwan. That model suggests that divisions within the party itself, and bottom-up pressure from the growing middle class, will become too much for the Communist Party to manage. It will accede to political change, in the hope that it will remain the dominant party. The example of the Kuomintang suggests if there are competitive elections, the Communist Party would win them – meaning China might democratise in the 2020s but not see a change of ruling party until the 2030s.

In both China and Vietnam, a shift to closed anocracy is likely to precede a shift to open anocracy and/or democracy. Their communist parties are unlike those of eastern Europe, which were externally imposed. Nor are they long-lasting strongman dictatorships like Spain or Portugal, which collapsed with a change of leader (as Vietnam’s January 2016 party congress showed with the ousting of the premier). So a sudden collapse of Communist rule is unlikely in either.

When? If we strip out the oil exporters because they never democratise (and China is not an oil exporter), we have a 4% chance annually now of China shifting towards democracy. This increases to 10% (on an inadequate sample), as China enters the $15-25k per capita GDP range. Based on our assumption that Chinese GDP growth slows by 0.5 ppt a year every year, we see China entering this $15-25k range in 2018, crashing temporarily in 20239, then growing by 3% annually after that. China could end up being stuck in a middle income trap from 2018 until 2035, until it sees the democratisation we expect.

We suspect then that Chinese democratisation will be a post-2022 19th National Congress theme. It could of course come earlier – like the often-heralded Chinese banking crash – but data tell us China will be at maximum risk of democratisation in the 2020s due to the size of the middle class. The economy we think will be struggling with Japan 1990-type symptoms of excessive real estate valuations and potentially inflation. Middle income trap stagnation might be a threat too by this time, if we are right in our view above that may be necessary to escape it.

Vietnam is the only other autocracy we cover that does not export oil on a net basis. It has a 4% chance of moving towards democracy in any given year, rising to 7% once its per capita GDP (now $5k) rises into the $8-15 range, where China sits today. Shifts in the current $4-8k range within Asia include Indonesia (1998), Bhutan (2005) and South Korea

8 Note Czechoslovakia was not included in this table, we mention it only for reference. 9 See Fixit, published 7 June 2016.

20 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

(1981), but there are plenty of other episodes too, including double events in Algeria, Argentina, Iran and Panama.

Again, Taiwan may be the most appropriate model for Vietnam too, suggesting that a political shift could be a late 2020s or 2030s story. But once again it comes down to the size of the middle class, and the potential for division within the ruling party, which is already seen as factional between pro-China and pro-US lines.

Do countries grow faster if they are autocracies when they are poor?

China’s leaders – and many others – believe Russia made a mistake in democratising before completing economic reform. Some argue autocracy is a better system for boosting growth, after which democracy comes into its own. The section above, talking about Russia, Malaysia and Turkey being potentially stuck in a middle income trap, implies that democracy will win out in the end, but what about lower down the income scale?

Among the countries in our tables, the poorest closed anocracies of Rwanda, Ethiopia and Uganda, along with the poorest autocracies, Vietnam and China, have produced some of the best growth rates in the world since 2000. This group has governments which prioritise infrastructure and economic development and are pretty good at it.

Yet there is a sample bias here. We do not see the autocracies of Burundi, Cuba or Tajikistan, because they do not have equity or debt markets.

Below we compare growth over 2004-2014 with average Polity IV scores over the same period, excluding the oil exporting countries (which grew rapidly because oil prices were high). We see that the five countries we focus on have among the best growth stories among those countries with a negative Polity IV score. There are a number of others that are much less impressive, including Eritrea, where GDP growth probably failed to keep pace with population growth (its education numbers have historically been poor too).

Figure 19: For countries with a negative Polity IV score in 2015 – we show average real % GDP growth and average Polity IV rating for 2004-14 (EM in blue, frontier in green, beyond-frontier in yellow, developed in red)

Average GDP 2004-14 2004-14 avg Polity IV rating 15

10

5

0

-5

-10 Laos Syria Togo Egypt China Eritrea Jordan Belarus Burundi Uganda Gambia Ethiopia Rwanda Vietnam Morocco Thailand Tajikistan Swaziland Singapore Mauritania Cameroon Afghanistan Sudan-North Source: MSCI, Polity IV, Renaissance Capita, IMF

21 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Of course, it may be that while there is a range of growth for autocracies, that range is higher than the range of growth for democracies – so autocracies are still better? In the chart below, we strip out all countries with a nominal per capita GDP of $18k, because we want to exclude rich countries that inevitably grow slowly, and compare Polity IV scores with GDP growth.

We find autocracies grow a little faster than democracies but the r-squared is an insignificant 6%. This is the case whether we include oil exporters Angola, Azerbaijan and Kazakhstan, or exclude them (they are excluded in the graph below). The data show democracies grow at roughly 4% while autocracies might grow at an average of 6%. Note that some doubt Chinese and Ethiopian headline growth figures, and similar doubts may be justified for Myanmar, Afghanistan and Turkmenistan. But more important, given the very wide range of growth among countries with a negative Polity IV score, it is hard to justify that this form of political system is best at low (or indeed high) income levels. Admittedly, we have not tested this correlation with 20th-century data.

Figure 20: GDP % change (vertical) compared with Polity IV rating (horizontal axis)

Avg real GDP growth, 2004-14 (y-axis) vs avg Polity IV score, 2004-14 (excluding iterregna and state failure) (x-axis) - countries with 2016 nominal GDP per capita <$18,000

12 TKM ETH R² = 6%

CHN 10

MMR MNG UZB TCD AFG8 PAN LAO RWA BTN KHM SLE MOZ IND TJK ZMB GHA UGA VNM TZA LKA COD LBR TLS BGD BLR 6 PNG ARM GEO SLB PER MRT BFA MWI DOM URY JOR NER KEN ARG PHL COG MYS LBN BOL BWAPRY DJI NPL NAM TUR MDA CRI MAR EGY KGZ SUR LSO CHL PAK BDI BEN MUS 4 HND GTM NICALBPOLCPV CMR TGO THA GUY MLI BRA MNE LTU SDN TUN SEN ZAF MKD SVK SWZ GMB CIV MDG UVK LVA GNB ROU BGR TTO SRB FJI YEM GIN MEX CZE 2 SLV COM ERI HTI UKR HUN ZWE HRV 0 JAM CAF

-2

-4

-6 SSD

-8 -10 -8 -6 -4 -2 0 2 4 6 8 10

Source: Polity IV, IMF, Renaissance Capital

22 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Political regime and the legal, business and corruption environment

We have published numerous times on legal systems10, the business environment11 and corruption12. Is there a link between them and Polity IV scores? Measuring this is problematic. Political freedom is one component of the , which accounts for half of the Renaissance Legal scoreboard; the other half of our legal score comes from components in the Ease of Doing Business rankings. Meanwhile, Transparency International’s Corruption Perceptions Index takes data from the World Justice Project. Each of these surveys is influencing at least one other of the surveys.

The key common element to all the comparisons is that stronger political systems – at either extreme – tend to deliver a better business, legal and corruption environment. Note the charts below do not account for the duration of country scores (how long a country has been a democracy, autocracy, etc.).

Legal systems

The legal system has the strongest correlation with political regime type. The autocracies of the UAE, China and Kazakhstan all have reasonably good legal scores. Full democracies have better legal systems, but that may be a function of higher per capita GDP.

Nigeria, Pakistan, India and Kenya have democracy but have legal systems that underperform our regression line. We might expect Nigeria to see an improvement in its legal system since it returned to a ‘democracy’ score in 2015.

10 See for example, Thoughts from a Renaissance man: Are you constrained by the law? published 31 March 2015. 11 See for example Thoughts from a Renaissance man: Post-communist reform and Vietnam, published 1 December 2015. 12 See for example Thoughts from a Renaissance man: Climb into his skin and walk around in it, published 9 December 2013.

23 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 21: Renaissance Capital legal score (y axis) vs Polity IV rating (x axis), 2015 100 FIN NOR AUS DEU DNK SWE SGP KOR JPN R² = 0.4801 AUT NZL GBR NLD 90 BEL FRA CAN EST USA PRT CZE POL ESP 80 SVN

ITA 70 ARE CHL ROU GEO HRV

MYS HUN 60 BWA JAM ZAF KAZ BRA BGR GRC MEX CHN RUS TUN THA 50 SRB ARG PHL MNG PER 40 COL MAR TZA TUR GHA VNM JOR CIV

30 UGA LKA IDN ETH UKR IRN ZMB LBN KEN

EGY PAK 20 IND NGA KHM

VEN 10 ZWE MMR BGD

0 -13 -8 -3 2 7 12

Source: Polity IV, World Justice Project, , Renaissance Capital

24 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

There is also some correlation with the Ease of Doing Business scores. Again, strong political systems, whether democratic or autocratic, seem to produce a better business environment than mixed political systems.

Figure 22: Ease of Doing Business scores (y-axis) vs Polity IV rating (x-axis), 2015 90 NZL R² = 0.3856 SGP DNK KOR USA GBR NOR SWE MKD FIN TWNDEU AUS 80 EST IRL CAN MYS LTU LVA AUT GEO SVN POL PRT FRA CHE SVK ARE ESPJPN ARM ROU CZE MEX BGR NLDITA KAZ BEL HRV HUN BLR MNE MUS THA RUS PER CHL CYP 70 ISR COL MDA TUR GRC SRB LUX CRI AZE RWA BHR JAM MNG QAT KGZ UVK OMN BTN PAN MAR TUN BWA ZAF GTM SAU CHN UKR SLV TTO UZB VNM FJI NPL DOM PRY URY KWT 60 NAM ZMB PHL ALB SWZ LKA KEN HND BRA IDN IRN JOR ECU GHA LSO ARG UGA LBN SLB CPV KHM NIC EGY IND LAO TJK MOZ PAK TZA PNG MWI GUY CIV MLI BFA ETH 50 SLE GMB TGO BDI SEN ZWE SUR COM SDN MDG BEN GAB IRQ BOL MMR DZA GIN MRT NER NGA CMR DJI TLS SYR BGD COG AFG LBR GNQ AGO 40 GNB TCD COD

VEN

30 ERI

20 -10 -5 0 5 10

Source: Polity IV, World Bank

25 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Lastly, the Corruption Perceptions Index suggests, once again, that a definite political system helps reduce corruption.

Figure 23: Corruption Perceptions Index (y-axis) vs Polity IV (x-axis) ranks 100

R² = 0.3911 DNK 90 FIN SWE NZL NOR NLD SGP CHE CAN 80 GBR DEU LUX AUS BEL AUT USA IRL JPN URY QAT FRA ARE 70 EST CHL

BTN BWA PRT POL TWN ISR CYP 60 LTU SVN ESP KOR CZE CRI LVA CPV RWA JOR NAM MUS SAU GEO BHR HRV HUNSVK 50 MYS KWT GHA ROU GRC OMN SEN LSO MNE ITA TUR ZAF MKD JAM BGR 40 SRB TTO ZMB SLV PAN MNG THA BFA TUN BRA IND CHN MAR LBR BEN EGY DZA SUR LKA ALB PER ARM COL MEXPHL GAB DJI MLI NER BOL ETH DOM MDAIDN BLR TGO CIV ECU UVK ARG VNM MRT MOZ MWI HND 30 TZA PAK AZE RUS SLEGUY TLS KAZ GMB MDG LBN KGZ GTM PRY IRN CMR UKR NPL NIC TJK NGA COM LAO UGA BGD GIN PNG KEN COG TCD MMR COD BDI20 KHM ZWE UZB SYR TKM ERI VEN GNB IRQ AGO SDN AFG10

0 -10 -5 0 5 10

Source: Polity IV, Transparency International

Curiously, if we strip out richer countries (those with nominal per capita GDP above $18k), the correlation between corruption and legal systems goes down. We think this is because per capita GDP is a significant determinant of corruption. Rich countries, whether they are Western democracies or autocratic oil exporters, have less corruption than poor countries (within that, oil exporters have more corruption than similarly wealthy non-exporters). The chart below shows this, with some oil exporters highlighted in yellow for their per capita GDP – and this is where corruption scores dip lower too.

26 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 24: Corruption scores (lhs) improve as countries get richer (per capita GDP, 2015 nominal dollars, rhs) – oil exporters (yellow) often have worse corruption than per capita GDP peers GDP per Capita, 2015 CPI 2015 100 50,000 90 45,000 80 40,000 70 35,000 60 30,000 50 25,000 40 20,000 30 15,000 20 10,000 10 5,000 0 0 US Iran Italy Peru India Chile Togo Libya Chad Niger Malta Brazil Benin China Nepal Korea Japan Kenya Ireland Turkey Jordan Austria Angola Nigeria Yemen Guinea Finland Greece Albania Djibouti Zambia Norway Ukraine Bahrain Canada Guyana Gambia Georgia Ethiopia Bulgaria Vietnam Rwanda Jamaica Armenia Panama Uruguay Hungary Slovenia Australia Romania Mongolia Lithuania Germany Tanzania Colombia Sri Lanka Argentina Suriname Honduras Cambodia Singapore Venezuela Azerbaijan Guatemala Kyrgyzstan Bangladesh Cabo Verde Luxembourg Saudi Arabia New Zealand Burkina Faso Turkmenistan Guinea-Bissau Trinidad Tobagoand Sao Tome and Principe Central African Republic Source: Transparency International, BP, IMF, Renaissance Capital

Overall, we think investors seeking a reasonable legal system, less corruption and a better business environment may be better served investing in an autocracy or a democracy, than in an anocracy with a score of around -1 or +1. This helps explain why FDI can still be high in China, Vietnam or oil exporters. It is a point we made in our piece on political leaders who remain in place beyond 10-15 years – corruption may be a problem but this does not appear to deter FDI13.

Conclusions

The table below shows what history tells us to expect on an annual basis in our universe of countries. We show the number of data points, admitting that fewer than 100 points means forecasts are likely to have low reliability, around 100-300 being of medium reliability and over 300 being high reliability. We may be able to increase the sample size once the Madison table of PPP GDP data is released later this year, particularly for low income countries. We think this might show a decrease in the percentage chance of change, because so many low income countries saw so little change through the 19th century, for example.

We also note local cultural factors, such as Thailand or Pakistan’s historical political volatility, or the democratising impact of EU membership, which will affect political risks in ways we cannot reliably quantify.

The likelihood of no change to a political regime in any given year ranges from 85.7% to 100%.

Emerging markets

In EM, Thailand and Egypt appear most likely to change, with a 10% chance of a shift to an open anocracy (to become like present-day Russia or Turkey) or a 2% chance of a full shift to democracy.

13 See Directional economics: The problem with political longevity, published 12 August 2013.

27 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

There is also a 10% chance of Malaysia, Russia or Turkey making the shift to democracy. Unfortunately, the sample size for all five countries is small, so we have low confidence in these percentages.

The limited data we have for Malaysia, Russia and Turkey suggest they must either democratise or stagnate within their current income bracket. It would be unprecedented if they become richer open anocracies, but we must also note it was unprecedented that each should lose democracy in the past decade and become open anocracies.

The next most vulnerable country to change is China, albeit with just a 4.5% chance of some form of democratisation in any given year. There is no rich autocracy in the world except for oil exporters, so at the least, we assume China will democratise to Singapore’s closed anocracy status in the 2020s. But as China is not a city state, we think the South Korean and Taiwan models are more appropriate. Their example implies that after becoming a closed anocracy, there will be a 8% chance of further democratisation to open anocracy or democracy in the 2020s when we think China’s per capita GDP in PPP dollars will be in the $15-25k range. Democratisation is usually managed peacefully at China’s level of per capita GDP.

India and the Philippines are 96% likely to remain democracies in any given year, and SA is 97% likely to do the same. MSCI LatAm countries are even more (97-100%) likely to remain stable democracies.

We see a small chance that Hungary and Poland are downgraded from full democracy to democracy, as the Czech Republic was a decade ago.

Figure 25: The percentage chance of a political regime changing each year MSCI Closed Open Full Failed MSCI EM Autocracy Democracy Occupation Sample Reliability Historically weight anocracy Anocracy Democracy state China 25.7 95.5 2.8 1.1 0.6 0 0 0 178 Medium South Korea 14.7 0 0 0 98.9 1.1 0 0 88 Low Taiwan 12.3 0 0 0 0.3 99.7 0 0 607 High India 8.4 0.7 0.9 1.4 96.1 0.7 0 0.2 444 High Surprisingly stable Malaysia 2.8 0 0 90.0 10.0 0 0 0 20 Low Often volatile Indonesia 2.8 0 0.5 0.7 97.3 1.5 0 0 411 High Thailand 2.3 0 87.7 10.1 2.2 0 0 0 89 Low Often volatile Philippines 1.5 0.7 0.9 1.4 96.1 0.7 0 0.2 444 High xx South Africa 7.6 0 0.5 0.7 97.3 1.5 0 0 411 High Medium oil exporter, democratisation Russia 3.6 0 0 90.0 10.0 0 0 0 20 Low wave 2018-20? Turkey 1.2 0 0 90.0 10.0 0 0 0 20 Low EU member, but vulnerable to Poland 1.1 0 0 0 0.3 99.7 0 0 607 High shift to democracy Qatar 0.9 99.2 0.8 0 0 0 0 0 127 Medium UAE 0.9 99.2 0.8 0 0 0 0 0 127 Medium Greece 0.3 0 0 0 0.3 99.7 0 0 607 High EU member EU member, but vulnerable Hungary 0.3 0 0 0 0.3 99.7 0 0 607 High to shift to democracy Egypt 0.2 0 87.7 10.1 2.2 0 0 0 89 Low EU member, but lost Czech Rep 0.2 0 0 0 98.9 1.1 0 0 88 Low full democracy once xx Brazil 7.4 0 0.5 0.7 97.3 1.5 0 0 411 High Mexico 3.9 0 0 0.7 98.0 1.3 0 0 154 Medium Chile 1.2 0 0 0 0 100 0 0 471 High Colombia 0.4 0 0.5 0.7 97.3 1.5 0 0 411 High Peru 0.4 0 0.5 0.7 97.3 1.5 0 0 411 High

Source: Penn tables, Polity IV, Renaissance Capital

28 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Frontier

Within MSCI Frontier, the countries most likely to change are Bangladesh, Morocco and Jordan.

Morocco has a 5% chance of shifting to full democracy in any given year, a 4.2% chance of a shift to open anocracy but also a 3.6% chance of a shift back to autocracy. The country’s ruler managed the Arab Spring very well, in our view, which supports the view that democratisation is the more likely direction, and it should be relatively peaceful.

Jordan has a 10% chance of democratic shift to open anocracy and a 2% chance of a shift to democracy in any given year.

Bangladesh has significant two-way risk. There is a roughly equal 4.9% chance of a shift to open anocracy,(plus 0.9% to democracy) or a 4.5% chance of a shift to autocracy (plus 1.8% to closed anocracy).

The rest of the frontier universe has a 95% or higher chance of remaining with the current regime in each year. The three countries marginally more vulnerable to change than the others are Nigeria, Kenya and Pakistan. These are sufficiently low income that there is a 0.4% chance of state failure in any given year.

While the data suggest Pakistan and Nigeria are equally (3%) likely to shift towards autocracy, we suspect the risk is actually relatively higher in Pakistan. The military remain politically powerful in Pakistan (as the Turkish military were in Turkey until about 2007). while in Nigeria, the military’s ability to conduct a coup appears to have been minimised over the past generation.

Figure 26: The chance of a political regime being this type in a year's time MSCI MSCI Closed Open Full Failed Autocracy Democracy Occupation Sample Reliability Historically Frontier weight anocracy anocracy democracy state Kuwait 18.1 99.2 0.8 0 0 0 0 0 127 Medium Morocco 8.5 3.6 87.1 4.3 5.0 0 0 0 139 Medium Oman 4.4 99.2 0.8 0 0 0 0 0 127 Medium Lebanon 3.4 0 0.5 0.7 97.3 1.5 0 0 411 High Bahrain 3.2 99.2 0.8 0 0 0 0 0 127 Medium Jordan 1.2 0 87.7 10.1 2.2 0 0 0 89 Low Tunisia 0.6 0 0.5 0.7 97.3 1.5 0 0 411 High xxx Argentina 15.8 0 0 0.7 98.0 1.3 0 0 154 Medium xxx Nigeria 8.4 0.7 0.9 1.4 96.1 0.7 0 0.2 444 High Kenya 5.3 1.1 0.4 1.8 95.2 0.7 0.4 0.4 275 Medium Mauritius 3.2 0 0 0 0 100 0 0 471 High xxx Romania 3.5 0 0 0.7 98.0 1.3 0 0 154 Medium EU member Slovenia 2 0 0 0 0.3 99.7 0 0 607 High EU member Kazakhstan 1.8 98.0 1.0 0 0 0 1.0 0 127 Medium Large oil exporter Croatia 1.4 0 0 0.7 98.0 1.3 0 0 154 Medium EU member Estonia 0.5 0 0 0 98.9 1.1 0 0 88 Low EU member Serbia 0.2 0 0.5 0.7 97.3 1.5 0 0 411 High Lithuania 0.1 0 0 0 0.3 99.7 0 0 607 High EU member Bulgaria 0.1 0 0 0.7 98.0 1.3 0 0 154 Medium EU member xxx Pakistan 10 0.7 0.9 1.4 96.1 0.7 0 0.2 444 High Often volatile Vietnam 4.1 95.8 3.1 0.2 0.9 0 0 0.2 457 High Bangladesh 2.5 4.5 1.8 87.0 4.9 0.9 0 0.9 223 Medium Sri Lanka 1.7 0 0.5 0.7 97.3 1.5 0 0 411 High

Source: Penn tables, Polity IV, Renaissance Capital

29 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Beyond frontier

This is where political uncertainty is at its highest.

We have two oil exporters, Iran and Venezuela (which export similar quantities of oil to Russia), that we think may face democratisation pressures in the wake of the oil price fall, perhaps in 2018-20. The data imply this is far more likely in Venezuela, with an 8.6% annual chance of a shift to democracy, than in Iran which appears to have just a 1% chance of a shift to closed anocracy each year. We believe that because Iran exports only a limited amount of oil per capita – but the dataset mainly consists of the Gulf’s high oil exporters – the real chance of change is higher than this.

Angola also appears to have a fairly high 9.3% chance of some form of democratisation in any given year, against a 3.6% chance of being more autocratic. However, it exports considerably more oil per capita than Venezuela, and in addition has low levels of education (implying a narrow but wealthy middle class), which makes us think Angola is not really that likely to democratise.

There is significant two-way political risk in Ivory Coast, Mozambique, Tanzania and Zimbabwe. There is roughly a 1 in 20 chance each year that they might move towards democracy, and a similar chance that each moves towards autocracy. We think investors need to recognise that the chance of significant political change is 13-14% each year, compared with just 4-5% in more established markets such as Nigeria or Kenya.

There is also reasonably high two-way risk in Ethiopia, Rwanda and Uganda, with a roughly similar 3-4% chance of a shift towards democracy or towards autocracy in any given year. By the 2020s – assuming strong GDP growth – we expect the odds to favour democratisation (7.6%) if change occurs in these countries, but they have to see at least a threefold increase in wealth levels to reach where Morocco is today (with a 9.3% chance of some sort of democratisation), and that may take until the 2030s.

Figure 27: The chance of a political regime being this in a year's time Beyond Closed Open Full Failed Autocracy Democracy Occupation Sample Reliability Historically frontier anocracy anocracy democracy state Medium oil exporter, democratisation Iran 98.0 1 0 0 0 1 0 127 Medium wave 2018-20? Unlike oil exporting Canada or Norway, Iraq 0 0.5 0.7 97.3 1.5 0 0 411 High democracy was imposed

Saudi Arabia 99.2 0.8 0 0 0 0 0 127 Medium Large oil exporter xx

Azerbaijan 98.0 1 0 0 0 1 0 127 Medium Large oil exporter

Georgia 0 0.5 0.7 97.3 1.5 0 0 411 High

Ukraine 1.7 0.9 87.9 8.6 0.9 0 0 116 Medium xx Large oil exporter, so autocracy Angola 3.6 87.1 4.3 5.0 0 0 0 139 Medium risk higher

Botswana 0 0 0.7 98.0 1.3 0 0 154 Medium

Ethiopia 4.2 91.7 1.7 1.7 0 0 0.7 286 Medium

Ghana 1.1 0.4 1.8 95.2 0.7 0.4 0.4 275 Medium

Ivory Coast 4.5 1.8 87.0 4.9 0.9 0 0.9 223 Medium

Mozambique 2.6 3.9 85.7 5.2 0 0 2.6 153 Medium

Rwanda 4.2 91.7 1.7 1.7 0 0 0.7 286 Medium Presidential elections 2017

Tanzania 4.5 1.8 87.0 4.9 0.9 0 0.9 223 Medium

Uganda 4.2 91.7 1.7 1.7 0 0 0.7 286 Medium

Zambia 1.1 0.4 1.8 95.2 0.7 0.4 0.4 275 Medium

Zimbabwe 2.6 3.9 85.7 5.2 0 0 2.6 153 Medium xx

Jamaica 0.7 0.9 1.4 96.1 0.7 0 0.2 444 High Surprisingly stable Medium oil exporter, Venezuela 1.7 0.9 87.9 8.6 0.9 0 0 116 Medium democratisation wave 2018-2020?

Source: Penn Tables, Polity IV, Renaissance Capital

30 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

What we have not done in this report is analyse how a recession might change the data. The sample size tends to get too small, and we would need to look at lag effects, not just single-year effects. This report is long enough already.

We see this report as useful guidance for what might happen, even if we cannot always forecast the trigger for it to happen. In our view, the inevitable democratisation of China is rather less interesting than the trailblazing behaviours we see in Russia, Turkey and Malaysia – all of which have lost democracy at unprecedented levels, and are nearly in uncharted territory.

We are interested by a possible democratisation wave over 2018-2020 in Venezuela, Russia, Iran and a few other oil exporters.

We were surprised to see how likely democratisation is in Egypt in coming years. We were also negatively surprised to see how much political risk still exists in some of the lowest income Africa and Asian countries we follow. Beyond-frontier markets carry more risk than frontier markets.

We found no strong correlation to support the view that autocracies are better for growth at low income levels, but we did find some evidence that autocracies may produce a better business, corruption and legal environment than some anocracies.

Lastly, we want to emphasise again that history can be misleading. Yes, West European states were always ruled by autocrats, but that changed from the 19th century with the industrial revolution. Yes, Turkey and Nigeria have experienced a number of coups, but we think both have changed thanks to rising wealth. To quote John Connor from Terminator 2, “the future has not been written. There is no fate but what we make for ourselves”. It seems to us that rising wealth does ensure a different political fate for many emerging, frontier and beyond-frontier countries, from the bleaker fate some still assume for them.

31 Renaissance Capital A ppendix 6 September 2016

Thoughts from a Renaissance man

Some take seriously the ‘youth bulge’ theory that countries at are more risk of a political change if they have large young populations. The two tables below show there is some truth to this notion. Counties where 25.0-29.9% of the population are aged 15-29 are the most likely (3% annual chance) to see a transition from a positive to a negative Polity IV score, and vice-versa (4% annual chance). This compares with a less-than-1% chance of shifting to a negative rating when the youth bulge only accounts for 20.0-24.9% of the population. But it is curious that when the youth bulge is even higher, at 30.0% or more, the risk also falls to around 1%. We suspect a very high youth bulge is consistent with a country being poor (so political change is limited), while a very small youth bulge is consistent with a country being rich (and political change is limited). It is countries transitioning from poverty to wealth where we assume the youth bulge is in the 25.0- 29.9% range and most likely to change.

Figure 28: Chance of political transition when the share of the youth population is above this threshold

Annual probability of autocratic transition by youth (15-29) share of total population 3.0% to Autocracy 2.5%

2.0%

1.5%

1.0%

0.5%

0.0% 15% 20% 25% 30%

Note: We excluded the very few incidents where the share of the population was below 15% or 35% and above. Source: UN, Polity IV, Renaissance Capital

Figure 29: Chance of political transition to a positive Polity IV rating when the share of the youth population is above this threshold

Annual probability of democratic transition by youth (15-29) share of total population 4% To Democracy 4%

3%

3%

2%

2%

1%

1%

0% 20% 25% 30%

Note: We excluded the very few incidents where the share of the youth population was less than 20% or 35% and above. Source: UN, Polity IV, Renaissance Capital

This might imply that if countries get old before they get rich, then they may be more likely to retain the political status quo. This does suggest Russia may be more stable than Turkey, given its aged population, and that China’s ruling party may last longer in power if it can sustain its rule until the share of the youth population drops below 20% (in 2015 it was 22.8%).

32 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 30: Youth (15-29) share of population in 2015, by country (listed with Polity IV rating) 15-20% 20-25% 25-30% 30%+ EM Qatar Qatar (-10) South Africa South Africa (9) Malaysia Malaysia (5) Philippines Philippines (8) India India (9) Mexico Mexico (8) Peru Peru (9) Egypt Egypt (-4) UAE UAE (-8) Colombia Colombia (7) Indonesia Indonesia (9) Brazil Brazil (8) Turkey Turkey (3) Chile Chile (10) China China (-7) Thailand Thailand (-3) Taiwan Taiwan (10) South Korea South Korea (8) Russia Russia (4) Poland Poland (10) Hungary Hungary (10) Czech Republic Czech Republic (9) Greece Greece (10) xx FM Oman Oman (-8) Pakistan Pakistan (7) Lebanon Lebanon (6) Bangladesh Bangladesh (1) Kenya Kenya (9) Bahrain Bahrain (-10) Jordan Jordan (-3) Nigeria Nigeria (7) Morocco Morocco (-4) Vietnam Vietnam (-7) Kuwait Kuwait (-7) Tunisia Tunisia (7) Kazakhstan Kazakhstan (-6) Argentina Argentina (9) Mauritius Mauritius (10) Sri Lanka Sri Lanka (6) Serbia Serbia (8) Lithuania Lithuania (10) Estonia Estonia (9) Croatia Croatia (9) Romania Romania (9) Slovenia Slovenia (10) xx Beyond Zimbabwe Zimbabwe (4) Rwanda Rwanda (-3) Jamaica Jamaica (9) Zambia Zambia (7) Ivory Coast Ivory Coast (4) Ghana Ghana (8) Senegal Senegal (7) Iraq Iraq (6) Iran Iran (-7) Tanzania Tanzania (3) Saudi Arabia Saudi Arabia (-10) Georgia Georgia (7) Ukraine Ukraine (4) Source: UN, Polity IV, Renaissance Capital

33 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Below we group together a number of countries, showing their Polity IV ratings since 1960.

Figure 31: Eastern Africa: Polity IV rankings since 1960

Ethiopia Kenya Rwanda Uganda Tanzania

10 DEMOCRACY 8 6 4 2 0 -2 -4 -6 -8 AUTOCRACY -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

Figure 32: Southern Africa: Polity IV rankings since 1960

Angola Mozambique Zambia Zimbabwe

10 8 6 4 2 0 -2 -4 -6 -8 -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

Figure 33: West and North Africa: Polity IV rankings since 1960

Morocco Nigeria Ghana Ivory Coast Tunisia

10 8 6 4 2 0 -2 -4 -6 -8 -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

34 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 34: MSCI EEMEA: Polity IV rankings since 1960

South Africa Egypt Greece Turkey

10 8 6 4 2 0 -2 -4 -6 -8 -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

Figure 35: MSCI Emerging Europe: Polity IV rankings since 1960

USSR/Russia Poland Czechoslovakia / Czech Rep Hungary

10 8 6 4 2 0 -2 -4 -6 -8 -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

Figure 36: Frontier – Emerging Europe: Polity IV rankings since 1960

Romania Serbia Kazakhstan Georgia Ukraine

10 8 6 4 2 0 -2 -4 -6 -8 -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Note 1: For Serbia, Yugoslavia (to 2002), Serbia and Montenegro (2003-2005), Serbia 2006-onwards. Note 2: For Kazakhstan, Georgia and Ukraine we use USSR data to 1990 inclusive. Source: Polity IV

35 Renaissance Capital 6 September 2016

Thoughts from a Renaissance man

Figure 37: Frontier – Argentina and South Asia: Polity IV rankings

Argentina Bangladesh Pakistan Sri Lanka

10 8 6 4 2 0 -2 -4 -6 -8 -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

Figure 38: Asia: Polity IV rankings since 1960

Taiwan South Korea Philippines Indonesia 10 8 6 4 2 0 -2 -4 -6 -8 -10

1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

Figure 39: Latin America: Polity IV rankings since 1960

Brazil Mexico Venuezuela Colombia Peru Chile

10 8 6 4 2 0 -2 -4 -6 -8 -10 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

Source: Polity IV

36

Renaissance Capital Renaissance Capital Ltd. Renaissance Capital Renaissance Securities (Cyprus) Ltd. Moscow London Johannesburg Nicosia T + 7 (495) 258 7777 T + 44 (203) 379 7777 T +27 (11) 750 1400 T + 357 (22) 505 800 Renaissance Securities (Nigeria) Ltd. Renaissance Capital Renaissance Capital Renaissance Capital Lagos Nairobi Istanbul Dubai T +234 (1) 448 5300 T +254 (20) 368 2000 T +90 (212) 362 3500 T +971 (4) 409 2000

This Communication is for information purposes only. The Communication does not form a fiduciary relationship or constitute advice and is not and should not be construed as a recommendation or an offer or a solicitation of an offer of securities or related financial instruments, or an invitation or inducement to engage in investment activity, and cannot be relied upon as a representation that any particular transaction necessarily could have been or can be effected at the stated price. The Communication is not an advertisement of securities nor independent investment research, and has not been prepared in accordance with legal requirements designed to promote the independence of investment research and is not subject to any prohibition on dealing ahead of the dissemination of investment research. Opinions expressed therein may differ or be contrary to opinions expressed by other business areas or groups of Renaissance Capital as a result of using different assumptions and criteria. All such information is subject to change without notice, and neither Renaissance Capital nor any of its subsidiaries or affiliates is under any obligation to update or keep current the information contained in the Communication or in any other medium.

Descriptions of any company or issuer or their securities or the markets or developments mentioned in the Communication are not intended to be complete. The Communication should not be regarded by recipients as a substitute for the exercise of their own judgment as the Communication has no regard to the specific investment objectives, financial situation or particular needs of any specific recipient. The material (whether or not it states any opinions) is for general information purposes only and does not take into account your personal circumstances or objectives and nothing in this material is or should be considered to be financial, investment or other advice on which reliance should be placed. Any reliance you place on such information is therefore strictly at your own risk. The application of taxation laws depends on an investor’s individual circumstances and, accordingly, each investor should seek independent professional advice on taxation implications before making any investment decision. The Communication has been compiled or arrived at based on information obtained from sources believed to be reliable and in good faith. Such information has not been independently verified, is provided on an ‘as is’ basis and no representation or warranty, either expressed or implied, is provided in relation to the accuracy, completeness, reliability, merchantability or fitness for a particular purpose of such information, except with respect to information concerning Renaissance Capital, its subsidiaries and affiliates. All statements of opinion and all projections, forecasts, or statements relating to expectations regarding future events or the possible future performance of investments represent Renaissance Capital’s own assessment and interpretation of information available to them currently. Any information relating to past performance of an investment does not necessarily guarantee future performance.

The Communication is not intended for distribution to the public and may be confidential. It may not be reproduced, redistributed or published, in whole or in part, for any purpose without the written permission of Renaissance Capital, and neither Renaissance Capital nor any of its affiliates accepts any liability whatsoever for the actions of third parties in this respect. The information may not be used to create any financial instruments or products or any indices. Neither Renaissance Capital and its affiliates, nor their directors, representatives, or employees accept any liability for any direct or consequential loss or damage arising out of the use of all or any part of the Communication.

© 2016 Renaissance Securities (Cyprus) Limited. All rights reserved. Regulated by the Cyprus Securities and Exchange Commission (Licence No: KEPEY 053/04).