Sovereign Wealth Funds, Cross-border investment and Institutions:
Are Arab Countries Different?Ibrahim El Badawi Raimundo Soto Chahir Zaki
November 26-27, 2017, AFESD, Kuwait
Outline
• Our story
• Stylized Facts
• Methodology
• Empirical Findings
Outline
• Our story
• Stylized Facts
• Methodology
• Empirical Findings
Our Story: motivation (1/2)
• Sovereign Wealth Funds (SWF) have become important players in global financial markets. Traditionally, SWFs save extra-budgetary surpluses (from resource revenues, foreign exchange reserves, or pension contributions) to:• Smooth public revenue volatility and spending; and,
• Ensure intergenerational equity.
• SWF surpassed US$ 5.5 trillion in assets in 2014, growing nine-fold since 2002 due to: • The large number of new sovereign wealth funds that have been formed over
the past couple of years;
• The growing assets of pre-existing sovereign wealth funds.
Our Story: motivation (2/2)
• As some of these SWFs enjoy ample liquidity with deep pockets (over USD30 billion in assets per fund), they have emerged as major global investors(of global equity), seeking long-term investment opportunities.
• SWFs have the choice to pursue either domestic (home) investments orabroad:• However, despite the presence of highly endowed SWFs in the Arab region, relatively
low cross-border SWFs funds are invested in capital importing Arab economies.
• On average, cross-border investments account for 89% percent of totaltransactions and most of them are in the form of FDI (90% or more):• However, the share of the Arab countries in the total amount of transactions does
not exceed 11 percent despite the presence of large SWFs in the region.
Our Story: what we do (1/1)
• Using data on the financial transactions on SWFs and building on therather nascent literature on SWFs, we use a linear probability modelto analyze the choice of the investment decision to be either home orabroad.
• In addition to this analysis of the investment decision at the“extensive margin”, we also assess the determinants of the SWFsinvestment abroad at the “intensive margin”; i.e. the share of theirinvestment broad relative to total investment.
• Moreover, we also analyze the determinants of investment at theaggregate as well as sectoral levels (equity, real estate, infrastructure,etc.).
Our Story: some pivotal questions (1/1)
• Why there is such low intra-Arab world investments by SWFs?
• Could this phenomenon be explained by systematic factors/; or,
• Instead, is there an Arab-specific idiosyncratic bias against investing inthe Arab world that defies explanation by systematic factors?• The systematic factors include gravity model type variables as well as
economic and institutional determinants of investment risks and profitability.
• On the other hand, geopolitical and other idiosyncratic factors will beaccounted for by the Arab dummy.
Our story: what we found (1/1)
• Starting with a gravity-type benchmark model, we find that whileArab destination countries are characterized by a positive bias at theextensive margin level, there is a negative bias against them at theintensive one.
• However, once we control for institutions, such as governmenteffectiveness and control of corruption, the Arab effect is eitherreduced or becomes statistically insignificant.
• On the other hand, at the sectoral level, results are highlyheterogeneous as some sectors are sensitive to institutions such asindustry and consumer discretionary, while others are not such asenergy and real estate.
Our story: policy implications (1/1)
• The gist of the evidence suggests that:• The size of SWFs investments is fully explained by the extended model
• Most notably the role of economic governance institutions: control of corruption and government effectiveness
• Therefore, the starkly low quality of economic governance institutions inthe Arab investment recipient countries is likely to be the main factorbehind the relatively low cross-border investments in the Arab world.
• And, with the expected even stiffer competition for SWFs resources underthe “New Oil Order”, capital scarce Arab economies must significantlyenhance their the quality of their economic governance institutions
Outline
• Our story
• Stylized Facts
• Methodology
• Empirical Findings
Aggregate investment
Sectoral distribution
From Arab Arab
Non-
Arab
Non-
Arab
To Arab
Non-
Arab Arab
Non-
Arab
Average investment By sector (in
billion USD)
Consumer Discretionary 25.4 73.3 18.7 54.8
Energy 45.8 179.5 11.5 64.5
Financials 199.4 196.2 21.1 78.6
Healthcare 0.9 36.9 3.8 36.2
Industrials 1.8 93.9 3.4 18.6
Information Technology 5.0 78.4 1.1 18.6
Materials 0.5 87.5 4.1 37.1
Media and Entertainment 0.0 118.0 0.0 43.2
Real Estate 47.0 359.5 7.6 91.6
Telecommunications 88.3 45.0 10.8 76.6
Utilities 0.0 65.6 19.3 18.0
Infrastructure 0.0 638.6 2200.0 493.3Number of transaction by
investment typeOther 5 185 1 259
Equity 36 1,596 63 12,283
Descriptive Statistics
Number of transaction 41 1781 64 12542
Mean 34.5 164.4 191.8 85.9
Std. Dev. 316.0 553.2 274.3 235.4
Min 0.0 0.0 0.0 -18.6
Max 2000.0 9983.4 2200.0 9760.0
• The vast majority of thisinvestment was done infinance, followed bycommunications, realestate, energy andconsumer discretionary.
• Investments by Arab SWFsoutside the Arab world waslargely accounted for byinfrastructure, followed byreal estate, finance andenergy.
Intra-Arab SWF
ARE BHR KWT LBY OMN QAT SAU Total
ARE 19 1 4 0 0 1 0 25
BHR 0 3 0 1 0 0 1 5
EGY 8 0 1 0 0 0 0 9
JOR 2 0 2 1 0 0 0 5
KWT 1 0 4 0 0 0 0 5
LBY 0 0 0 1 0 0 0 1
MAR 6 0 0 0 0 1 0 7
OMN 1 0 0 0 10 1 0 12
QAT 5 0 0 0 0 18 0 23
SAU 0 0 0 0 0 0 3 3
TUN 2 0 0 1 0 1 0 4
Total 44 4 11 4 10 22 4 99
The low intra-ArabSWFs investmentsseems to becompensate for by thelarge government togovernment flows,which are mainlydriven by strategicconsiderations.
Institutions
However, the quality of institutionsin most non-GCC Arab countrieswas much lower relative to thestandard of the capital surpluseconomies of the GCC
Preliminary correlations
Gov. Eff. Cont. corr.
Residuals1 0.2048*** 0.251***
(0.000) (0.000)
Index2 0.189*** 0.216***
(0.000) (0.000)
A significantly positive correlation between first the residuals of regressingboth institutions and cross border investment on destination, year, sectorand investment type dummies and second the observed values of bothinstitutions and cross border investment.
Outline
• Our story
• Stylized Facts
• Methodology
• Empirical Findings
Baseline Specification
• We posit a linear probability model where the dependent variable is a crossinvestment dummy, which equals one if the target investment is madeabroad and zero otherwise. We also estimate a similar model, where thedependent variable is the real value of SWF’s abroad.
• A negative coefficient of the Arab dummy suggests that, controlling for thebaseline gravity factors, there is a bias against investing in Arab destinationcountries; while a positive estimate would suggest over-investment in Arabdestinations.
Extended model
• We introduce other determinants such as the stock of previousinvestments, which is assumed to reflect the extent of knowledgeabout the investment climate in the destination economy; returns toinvestment in the foreign destination economy relative to the homeeconomy; and measures of institutional quality of relevance toinvestment risks:
Outline
• Our story
• Stylized Facts
• Methodology
• Empirical Findings
Summary of Findings (1/1)
Agg. Inv. Equity Consumption Energy Finance Industry
Real
estate
Probability of the of SWFs investments (extensive margin)
Baseline
model +S +S +S +S +S +S +S
Extended
model +S +S dropped dropped +S NS NS
The determinants of the size of SWFs investments (intensive margin)
Baseline
model -S -S NS NS -S NS -S
Extended
model NS NS dropped dropped NS NS NS
Empirical Findings (1/2)
• The benchmark results of the baseline gravity-type model suggests that:• Foreign investors have a positive bias for the Arab destination countries at the extensive margin
level• However, there is a negative bias against them at the intensive one
• When we also control for relative profitability, past investment and economicgovernance institutions:• The positive bias at the extensive margin disappears or drops out for the consumption, energy,
industry and real estate sectoral destinations• Nonetheless, the Arab dummy remains positive and significant for aggregate investment, equity
and finance
• Therefore, cross-border SWFs investment decisions in Arab economies for the equity andfinance continue to be positively influenced by other, possibly strategic, factors notaccounted for by the model
• Moreover, on view of the dominance of the two as sectoral destinations of SWFs’investments, aggregate investment is also influenced by these idiosyncratic factors
Empirical Findings (2/2)
• On the other hand, at the intensive margin, the Arab dummy becameuniformly insignificant for the aggregate as well as for all types andsectoral destinations of SWFs’ cross-border investments.
Extensive Margin – baseline (1/4)
Intensive Margin – baseline (2/4)
Extensive Margin – extended (3/4)
Intensive Margin – extended (4/4)
Thanks for your attention