Corruption vs. Country Stock Returns, 1997-2016

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Corruption vs. Country Stock Returns, 1997-2016

Post by SimpleGift » Fri Apr 06, 2018 11:55 am

In a recent thread on investing internationally, an enduring question was raised but left unexamined: Do countries with more corruption have lower or higher stock returns? In this post, we compare Transparency International's average corruption scores for 70 countries worldwide (ranked by "their perceived levels of corruption, from expert assessments and opinion surveys") against their stock market returns for the 20-year period from 1997-2016.

In the two charts below, the data is divided into developed markets at left (n=24) and emerging/frontier markets at right (n=46). Corruption ranks are on the vertical axes and are the most recent 6-year average for each country (0-100, lowest rank = most corrupt). Average stock price returns for the 1997-2016 period are on the horizontal axes.
  • Image
    NOTE: Stock returns are U.S. dollar, price-only for the S&P free-float, cap-weighted country indices.
    Data: Corruption ranks from Transparency International; stock returns from World Bank.
In both cases, the correlations are weak and not necessarily causative. But in developed markets, the least corrupt countries performed slightly better on average, while in emerging/frontier markets the more corrupt countries outperformed. Perhaps bribing politicians and local officials leads to more lucrative deals or reduces competition, improving profit margins?

Bottom Line: With such an indifferent relationship between corruption levels and stock returns, it's hard to make the case that international stocks should be avoided because of higher levels of corruption, whether real or perceived. In fact, in emerging and frontier markets, corruption may actually boost returns a bit, on average, for the broadly diversified indexes.

Thoughts on any of this?
Last edited by SimpleGift on Sat Apr 07, 2018 12:28 am, edited 3 times in total.
Cordially, Todd

User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by SimpleGift » Fri Apr 06, 2018 11:57 am

If interested, these are the countries with the highest and lowest corruption ranks (past 6-year average):
These corruption rankings don't appear to change much over time, and are not too different than 15 years ago.
Cordially, Todd

golfCaddy
Posts: 695
Joined: Wed Jan 10, 2018 10:02 pm

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by golfCaddy » Fri Apr 06, 2018 5:12 pm

What happens if you plot all countries on the same chart, not separating out developed and emerging? Are those arithmetic mean or geometric mean returns? I wouldn't consider arithmetic mean returns useful.

User avatar
nisiprius
Advisory Board
Posts: 36443
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by nisiprius » Fri Apr 06, 2018 6:56 pm

What happens, just for laughs, if you do a Spearman's Rho rank correlation coefficient significance test on those data sets, using, say, this calculator? What results do you get?

My intuition is the same as yours--those correlations are not significantly different from zero, and it is not evidence of a "clean premium" for emerging markets... or a "corruption premium" for developed markets.
Last edited by nisiprius on Fri Apr 06, 2018 6:57 pm, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by SimpleGift » Fri Apr 06, 2018 6:57 pm

golfCaddy wrote:
Fri Apr 06, 2018 5:12 pm
What happens if you plot all countries on the same chart, not separating out developed and emerging? Are those arithmetic mean or geometric mean returns? I wouldn't consider arithmetic mean returns useful.
The regression below shows all 70 countries on one chart. Because we know from the start that emerging/frontier markets, on average, had higher returns than developed markets over the last 20 years, plus the lower corruption ranks (= higher corruption), they will dominate the all-country plot — which is why I thought to separate them out in the OP.
  • Image
Yes, the stock returns are arithmetic means, which is how the data was presented. Agree that CAGR would be more desirable.
Cordially, Todd

golfCaddy
Posts: 695
Joined: Wed Jan 10, 2018 10:02 pm

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by golfCaddy » Fri Apr 06, 2018 7:12 pm

SimpleGift wrote:
Fri Apr 06, 2018 6:57 pm
The regression below shows all 70 countries on one chart. Because we know from the start that emerging/frontier markets, on average, had higher returns than developed markets over the last 20 years
Are you sure about that? In Morningstar VEIEX would have grown to $28861, while VTSMX grew to $45248 from 1/1/1997 to 1/1/2017. It looks to me like emerging markets got killed over those 20 years, at least compared to US stocks.

User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by SimpleGift » Fri Apr 06, 2018 7:17 pm

nisiprius wrote:
Fri Apr 06, 2018 6:56 pm
What happens, just for laughs, if you do a Spearman's Rho rank correlation coefficient significance test on those data sets, using, say, this calculator? What results do you get?
From the calculator:
  • Developed Markets:
    Value of R = 0.26024
    Value of P = 0.21940

    Emerging Markets:
    Value of R = -0.23841
    Value of P = 0.10272
According to the calculator, "by normal standards, the association between the two variables would not be statistically significant." Not too surprising, but for future projects, I''m glad to know that this online calculator exists. Thank you.
nisiprius wrote:
Fri Apr 06, 2018 6:56 pm
My intuition is the same as yours--those correlations are not significantly different from zero, and it is not evidence of a "clean premium" for emerging markets... or a "corruption premium" for developed markets.
Agree — but it's also not a reason to completely avoid international stocks due to their real or perceived "higher corruption."
Cordially, Todd

User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by SimpleGift » Fri Apr 06, 2018 7:25 pm

golfCaddy wrote:
Fri Apr 06, 2018 7:12 pm
Are you sure about that? In Morningstar VEIEX would have grown to $28861, while VTSMX grew to $45248 from 1/1/1997 to 1/1/2017. It looks to me like emerging markets got killed over those 20 years, at least compared to US stocks.
Well, in the OP sample of 70 countries taken from all S&P country indices with at least 20 years of return history (24 developed and 46 emerging markets), the average annual returns were 8.0% for developed markets and 11.5% for emerging/frontier markets.

But these are equal-weighted averages (same as the OP analysis), which will differ from Vanguard's cap-weighted returns. Plus, you are comparing emerging markets (VEIEX) to U.S.-only stocks (VTSMX), which is not what the OP analysis is all about.
Cordially, Todd

bayview
Posts: 1536
Joined: Thu Aug 02, 2012 7:05 pm
Location: WNC

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by bayview » Fri Apr 06, 2018 8:09 pm

SimpleGift wrote:
Fri Apr 06, 2018 11:57 am
If interested, these are the countries with the highest and lowest corruption ranks (past 6-year average):
These corruption rankings don't appear to change much over time, and are not too different than 15 years ago.
And the US comes trudging in tied for 16th at 75 (maybe an average, instead of the 74 noted above.)
The continuous execution of a sound strategy gives you the benefit of the strategy. That's what it's all about. --Rick Ferri

stlutz
Posts: 4652
Joined: Fri Jan 02, 2009 1:08 am

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by stlutz » Fri Apr 06, 2018 8:20 pm

If there isn't a "corruption premium", then one could argue that it makes more sense to exclude the more corrupt markets, or just make it easy and skip emerging markets altogether. Why take extra risk if you can't expect to be rewarded for doing so?

Out of curiosity, are the returns you are showing in local currency or in USD?

golfCaddy
Posts: 695
Joined: Wed Jan 10, 2018 10:02 pm

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by golfCaddy » Fri Apr 06, 2018 10:05 pm

For four of the most corrupt countries, I calculated the average return and the CAGR for 1998-2016.

Kenya Average - +16.9%
Kenya CAGR - +8.6%
Bangladesh Average - +11.6%
Bangladesh CAGR - +4.9%
Ukraine Average - +18.2%
Ukraine CAGR - -2.66%(yes it's actually negative)
Russia Average - +27.5%
Russia CAGR - +4.1%

That may not answer the question of how corruption correlates to returns, but it should at least drive home the point that arithmetic mean is an extremely misleading measure of performance.

User avatar
siamond
Posts: 4152
Joined: Mon May 28, 2012 5:50 am

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by siamond » Fri Apr 06, 2018 10:34 pm

I appreciate the intent of this thread, I really do. It would be really great to be able to perform such analysis over a much longer period of time. Trouble is the historical data (e.g. corruption index) isn't available. So we're limited to 20 years of data, two consecutive decades. And I am afraid this is nowhere near good enough to draw any conclusion.

Let's do a simple thought experiment. I think we can all agree that the level of corruption in emerging countries is significantly higher than in developed countries (even if many developed countries still have quite some 'room for improvement' in this respect, shall we say).

Now let's look at what the 2014 Credit Suisse Investment Returns Yearbook indicated (the only data source I know covering more than 30 years of past history for emerging markets).

Let's pick two consecutive decades (80s and 90s). Clearly corruption hampers returns, right? Let's pick two other consecutive decades (60s and 70s). Clearly corruption helps returns, right? See my point?

Image

User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by SimpleGift » Fri Apr 06, 2018 11:25 pm

One additional data point for the relationship between corruption and stock returns. Dimson, Marsh and Staunton in their 2015 Credit Suisse Yearbook reported returns for 47 countries ranked by their corruption scores for the 2000-2014 period (chart below). According to their analysis, the "poor" countries (most corrupt, at left) did outperform, but with the important caveats below.
  • Image
    Source: Credit Suisse 2015 Yearbook, page 23.
Dimson, Marsh & Staunton wrote:Realized returns were higher for equity investments in jurisdictions that were more likely to be characterized by corrupt behaviors. This pattern is time specific, and there are sub-periods when more “saintly” markets did better than the sinners. Because the interval we study is short, our results may simply reflect a period when emerging markets outperformed.
As siamond pointed out in his comments (and DMS agrees), we only have 20 years of historical data on country corruption scores, and this is just not a long enough time period to draw any meaningful conclusions. But the OP analysis seemed at least worth taking a look at, with the data that we have available, however limited.
Cordially, Todd

User avatar
SimpleGift
Posts: 3121
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by SimpleGift » Fri Apr 06, 2018 11:31 pm

stlutz wrote:
Fri Apr 06, 2018 8:20 pm
Out of curiosity, are the returns you are showing in local currency or in USD?
According to S&P, the returns reported for all of their country indices are U.S. dollar returns, not local currency.
Cordially, Todd

CurlyDave
Posts: 685
Joined: Thu Jul 28, 2016 11:37 am

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by CurlyDave » Sat Apr 07, 2018 2:09 am

SimpleGift wrote:
Fri Apr 06, 2018 11:55 am
...
  • Image
    NOTE: Stock returns are U.S. dollar, price-only for the S&P free-float, cap-weighted country indices.
    Data: Corruption ranks from Transparency International; stock returns from World Bank.
In both cases, the correlations are weak and not necessarily causative. But in developed markets, the least corrupt countries performed slightly better on average, while in emerging/frontier markets the more corrupt countries outperformed. Perhaps bribing politicians and local officials leads to more lucrative deals or reduces competition, improving profit margins?

Bottom Line: With such an indifferent relationship between corruption levels and stock returns, it's hard to make the case that international stocks should be avoided because of higher levels of corruption, whether real or perceived. In fact, in emerging and frontier markets, corruption may actually boost returns a bit, on average, for the broadly diversified indexes.

Thoughts on any of this?
It is not at all obvious to me that a straight line is the best fit to the data points presented. Least squares analysis is always tempting, but I think I could do a line drawing of an elephant that would fit those data points and have just as much meaning as the straight lines.

Bud
Posts: 139
Joined: Sun Jul 22, 2007 6:59 am

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by Bud » Sat Apr 07, 2018 5:55 am

This is what Transparency International measured to achieve their data and determine 'corruption'.

"The CPI draws upon 13 data sources which capture the assessment of experts and
business executives on a number of corrupt behaviours in the public sector,
including:
 Bribery
 Diversion of public funds
 Use of public office for private gain
 Nepotism in the civil service
 State capture

"Some of the sources also look at the mechanisms available to prevent corruption in a
country, such as:
 The government’s ability to enforce integrity mechanisms
 The effective prosecution of corrupt officials
 Red tape and excessive bureaucratic burden
 The existence of adequate laws on financial disclosure, conflict of interest
prevention and access to information
 Legal protection for whistle blowers, journalists and investigators"

http://files.transparency.org/content/d ... ote_EN.pdf

golfCaddy
Posts: 695
Joined: Wed Jan 10, 2018 10:02 pm

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by golfCaddy » Sun Apr 08, 2018 12:35 am

It would be interesting if they compared risk levels between emerging and developed. Russia, for example, has a standard deviation of 79.7% from 1998-2016. If emerging has equal returns, but dramatically higher risk, then it's still a bad deal. The Dimson-Marsh yearbook is interesting because over 118 years, developed beat emerging by 1% per year. So if you invested in developed markets at the start of 20th century, it would have grown to 3x an investment in emerging markets.

asset_chaos
Posts: 1343
Joined: Tue Feb 27, 2007 6:13 pm
Location: Melbourne

Re: Corruption vs. Country Stock Returns, 1997-2016

Post by asset_chaos » Sun Apr 08, 2018 3:14 pm

Like higher GDP growth rates must translate into higher stock returns, avoiding international because those foreigners are corrupt appears to be one more investing thesis that may sound plausible but doesn't stand up to scrutiny. I suspect these and many similar ideas---the kind everyone knows is true because so many people repeat them---are talked about less as hypotheses to be tested, rather more as a veneer of justification for our unexamined biases.
Regards, | | Guy

Post Reply