Received: 12 February 2021 | Accepted: 23 June 2022
DOI: 10.1111/1467-8268.12656
OR IG INAL ART I C L E
On the allocation puzzle and capital flows: Evidence from
foreign direct investment and real sector growth in Africa
Michael Effah Asamoah1 | Imhotep Paul Alagidede2 | Frank Adu3
1Department of Accounting, University of
Ghana Business School, University of Abstract
Ghana, Accra, Ghana This paper takes a new dimension on the foreign direct investment (FDI)–growth
2Wits Business School, University of the nexus through the lens of the allocation puzzle. We examine how the influx of
Witwatersrand, Johannesburg, South
Africa FDI affects the broader economy through its impact on real economic sectors in a
3Transformation Fellow, African Center panel of 42 Sub‐Saharan African (SSA) countries over 1980–2017. Using a
for Economic Transformation, Accra, dynamic panel model and decomposing the real sector into its parts, we test for
Ghana the possibility of a two‐way bi‐directional relationship between growth in
Correspondence agriculture, manufacturing, industry and service value additions, and FDI. There
Michael Effah Asamoah, Department of is no evidence of an allocation puzzle at the level of the real sector and the
Accounting, University of Ghana
components parts, which suggests that SSA countries with relatively high growth
Business School, University of Ghana,
Accra, Ghana. in the real sector will attract more FDI. While the effect of FDI on the real sector
Email: measamoah@ug.edu.gh and is positive at the disaggregated level, there is a positive bi‐directional effect
iameffah@gmail.com
between FDI and growths in manufacturing, industry and service value additions.
The results are robust to key determinants of the growth–FDI nexus.
KEYWORD S
allocation puzzle, FDI, GMM, real sector growth, Sub‐Saharan Africa, value additions
1 | INTRODUCTION
The association between capital flows and economic growth emanates from the assumptions of neoclassical growth
theory (Solow, 1956). The theoretical basis posits that in a closed economy where foreign capital is restricted, higher
domestic interest rates stifle investment and growth. However, in a liberalized economy, barriers impeding capital
flows are removed. The influx of capital leads to a fall in the domestic interest rates, culminating in increased
investments and high growth. Obstfeld (2012) shows the flow of capital to originate from countries where the marginal
product of capital is low to countries that offer higher marginal rates per capita. Alley (2015) also posits that investors
are likely to earn a relatively higher return on the same amount of capital invested in capital‐deficient countries than
when invested in capital‐abundant countries. Despite the benefits of capital flows to economic growth, critical
questions about the directional flow of capital remain unanswered (see Lucas, 1990; Prasad et al., 2007). Current
arguments have also tended to focus on testing the validity of earlier theories on capital flows and economic growth
(Gourinchas & Jeanne, 2013; MacDonald, 2015; Prasad et al., 2007).
Theoretical conventions of the neoclassical growth framework predict that capital flows enhance economic growth
(Mankiw et al., 1992) and that faster‐growing economies should attract relatively more capital to enhance growth rates
(Gourinchas and Jeanne (2013). The latter proposition is because fast‐growing economies have relatively better
investment opportunities and may be creditworthy. However, some empirical studies suggest a contradiction to the
view postulated above—“the allocation puzzle”. The main thrust of our study is the weak link between theoretical
Afr Dev Rev. 2022;1–11. wileyonlinelibrary.com/journal/afdr © 2022 African Development Bank. | 1
2 | ASAMOAH ET AL.
predictions and empirical evidence on the directional flow of private capital and economic growth. Specifically, we look
at the relationship between real sector growth and foreign direct investment (FDI) on the back of the allocation puzzle.
The Organisation for Economic Co‐operation and Development (2008) notes the relevance of FDI to many economies.
Even though some studies have found that FDI may be affected by economic uncertainties (Asamoah et al., 2022;
Moraghen et al., 2021), Onye et al. (2022) and Levchenko and Paolo (2007) suggest that flows of such nature help
economies avoid the uncertainties associated with dependence on short‐term flows for investment and growth.
The literature remains divided on the relationship between capital flows and economic growth, and more importantly,
the voluminous literature is silent on the effect of FDI on real sector growth or the effect of the real sector growth of capital
flows. Explicitly, we test for evidence of the “allocation puzzle”, and a possible bi‐directional relationship in the association
between real sector growth and FDI in Sub‐Saharan Africa (SSA). Extant studies on the FDI–growth nexus have focused on
gross domestic product (GDP), GDP per capita and productivity catch‐up as the only proxies for economic growth. To the
best of our knowledge, this represents the first shot at taking a critical look at the link between FDI and real sector growth,
in the light of the “allocation puzzle”, and through the lens of SSA. For instance, MacDonald (2015) employed data from
emerging and developing countries with no specific focus on SSA. Gourinchas and Jeanne (2013) employed data on non‐
Organisation for Economic Co‐operation and Development (OECD) countries with selected SSA countries. Secondly,
although an array of studies abound on the specific role of FDI in economic growth, only a handful have thoroughly
examined any bi‐directional relationship between growth and FDI. The literature is inconclusive about any bi‐directional
association between economic growth and FDI (Kholdy & Sohrabian, 2005; Luca & Spatafora, 2012), not to mention FDI
and real sector growth. Although Beri and Nubong (2021) notes that Africa's share of global FDI remains infinitesimal, the
focus on FDI and Africa is necessary as such flows have helped most African countries to move away from being heavily
dependent on aid and natural resource endowment. Ernst and Young (2013) show the FDI has in the past created almost
1.6 million jobs on the continent.
Even though studies on capital flows and the real sector in Africa are gaining attention (Asamoah & Alagidede, 2021;
Taylor, 2020), the issue of the allocation puzzle has received minimal attention. In a related study Asamoah and Alagidede
(2020) examined the issue of the allocation puzzle, but the study primarily focused on portfolio investments. With data on
42 SSA countries, the study therefore examines whether FDI promotes growth of the real sector or whether real sector
growth attracts FDI in the wake of the allocation puzzle. The rest of the paper is structured as follows: Section 2 presents a
review of the literature on the allocation puzzle and capital flows. Section 3 discusses the data and methodology employed.
We present our results in Section 4, with policy recommendations and conclusions in Section 5.
2 | LITERATURE REVIEW
Employing data on a set of 65 non‐OECD countries between 1980 and 2000, Gourinchas and Jeanne (2013) found great
inconsistencies in the neoclassical predictions between the flow of foreign capital and economic growth. They show
that, over the period, countries that have achieved higher growth rates relative to their counterparts attract smaller
amounts of net capital inflows, revealing an inverse association between net capital inflows and productivity growth
rates. They found a significant negative relationship between net capital flows and productivity growth even after
controlling for initial capital abundance, initial debt, population growth, capital controls and financial openness. An
analysis of the puzzle, however, reveals that the allocation puzzle is a function of the savings component of net capital
flows as well as the publicly originated capital inflows.
Although the initial association between capital flows and growth produced an inverse relationship on a sample of
67 countries from 1980 to 2007, Alfaro et al. (2014) show that the negative association was due to the public component
of capital flows. Thus, when we decompose capital flows into private and public flows, there exists a positive
association between private capital flows and growth, a phenomenon that supports the neoclassical view of the capital
flow–growth nexus. It is therefore imperative for governments to focus more on addressing issues relating to public
savings and official flows as well as current accounts and not solely on private savings if the full benefits of capital flow
on growth will be attained.
Similar conclusions by Schroth (2016) show that indeed capital does that not flow to fast‐growing economies in the
right proportions and that although FDI flows seems to follow growth, large savings in the form of sovereign debt
mostly offset the positive relationship between FDI and growth, resulting in the allocation puzzle between total capital
flows and growth in developing countries. However, these studies are conscious of their sample size: while Prasad et al.
(2007) focused on non‐industrialized countries, Gourinchas and Jeanne (2013) focused on OECD countries. Although
ASAMOAH ET AL. | 3
Alfaro et al. (2014) and MacDonald (2015) employed a sample of both developed and developing countries, SSA as an
investment destination has received no attention.
3 | DATA AND METHODOLOGY
We construct a panel of 42 SSA countries to test for evidence of an allocation puzzle and the possibility of a two‐way
relationship between foreign investment and real sector growth. The choice of country was based on data availability.
We extracted data from the World Development Indicators of the World Bank, the Penn World Tables and the Chinn
and Ito (2008) Index.
3.1 | Econometric models of real sector and FDI
We specify a regression model to capture a two‐way association between real sector growth and FDI. This approach has
been used in the FDI–growth literature (Asamoah & Alagidede, 2020; Iamsiraroj, 2016). The econometric models are
specified below:
RSGit = αRSGit−1 + Σβ1Xit + Σβ2FDIit + Ui + εit + λt. (1)
FDIit = αFDIit−1 + Σβ1Xit + Σβ2RSGit + Ui + εit + λt. (2)
The dependent variable RSGit measures annual growth in the real sector for country i at time t. For our study, the
real sector consists of the manufacturing, industrial, agriculture and service sectors. We measure the growth of each
sector by value additions. In line with Asamoah and Alagidede (2020), we construct a real sector growth index, which is
an equal weight of the four components. Consistent with Asamoah and Alagidede (2021) and the definitions of
the World Bank, we define each of the four components. The annual growth of agriculture value‐added (AGVA)
encompasses value additions in forestry, hunting and fishing, as well as the cultivation of crops and livestock
production. Services value added (SERVA) comprises value additions in wholesale and retail trade transport, and
government, financial, professional and personal services such as education, healthcare and real estate services.
Industrial value‐added (INVA) comprises value additions in mining, manufacturing (also reported as a separate
subgroup), construction, electricity, water and gas. Manufacturing value added (MANVA) is the net output of
manufacturing after adding up all outputs and subtracting intermediate inputs. Value‐added is the net output of a
sector after adding up all outputs and subtracting intermediate inputs.
FDIit is the sum of equity capital, reinvestment earnings, other long‐term capital and short‐term capital as shown in
the balance of payment. FDI inflow is a percentage of GDP; RSGit−1 is a lag of growth of the real sector; and FDIit−1 is
the lag of FDI.
Xit denotes a set of control variables known to influence the growth of the real sector and the attraction of FDI. The
controls include M2, inflation, natural resources, exchange rate, gross domestic savings, trade openness, financial
openness, gross fixed capital, human capital and GDP growth. We use GDP growth for robustness.
3.2 | Dynamic panel estimations
The limitations in traditional estimators like the ordinary least squares (OLS) are known to be downwards biased,
especially when there are measurement errors with regressors. Again, OLS does not deal with the issues of endogeneity
of the main explanatory variable and the error term. In explaining endogeneity in its simplest form, Lynch and Brown
(2011) posit that endogeneity arises when a predictor or independent variable (X) in a regression model is deemed to
correlate with the error term in the regression model. This can occur when important variables are omitted from the
model (called “omitted variable bias”) and when the outcome variable (Y) is a predictor of (X) and not simply a
response to X (called “simultaneity bias”).
Furthermore, the OLS also does not account for the possibility of unobserved country‐specific differences. Thus, the
correct way to estimate such relationship is to estimate a dynamic panel regression by including a lag of the dependent
4 | ASAMOAH ET AL.
variable as a regressor in the main equation to avoid misspecifications (Arellano & Bond, 1991). Dynamic panel data
estimations have unobserved panel‐level effects, which are known to be associated with lags of the dependent variable,
thereby rendering standard estimators inconsistent. To deal with such inconsistency Arellano and Bond (1991)
proposed the GMM estimator which can provide consistent estimations for these models by taking the first difference
of the data and using the lagged values of the dependent variables as instruments, this is the basic GMM estimator
known as the difference GMM. The GMM estimator helps eliminate country‐specific bias by taking the difference
between Equations (1) and (2), as shown below:
yit − yi,t−1 = α ( yi,t−1 − yi,t−2) + β (Χit − Χi,t−i) + (εit − εi,t−i). (3)
The purpose of the differencing is to eliminate any potential unobserved country fixed effect. Also, Arrelano and
Bond (1991) notes the relevance of the instruments used under the GMM estimator. First, the use of the GMM
approach allows us to control for the endogeneity of the main explanatory variable. Second, to resolve potential
problems of new error terms (εit − εi,t−i) being correlated with the lagged of the dependent variable ( yi,t−1 − yi,t−2).
Also, to avoid further issues of serial correlation of the error terms and independent variables being weakly exogenous,
the GMM estimator employs additional moment conditions:
Ε[ yit−s . (εit − εi,t−1)] = 0 for s ≥ 2; t = 3, …. T, (4)
Ε[Χit−s. (εit − εi,t−1)] = 0 for s ≥ 2; t = 3, …T. (5)
The moment conditions in Equations (4) and (5) are employed in the GMM estimator to ensure consistent and
efficient parameter estimates. The difference GMM is however not immune to limitations. As noted by Arellano
and Bover (1995), lagged level are weak instruments. Furthermore, Blundell and Bond (1998) note that the
difference GMM also has very poor finite properties when it comes to precision and bias if the independent
variables are tenacious over time. The system GMM as proposed by Blundell and Bond (1998) improves the
limitations of the difference GMM as proposed by Arellano and Bond (1991) and the deviation GMM by Arellano
and Bover (1995), by providing additional moment conditions to deal with the issue of poor instruments in the
difference GMM. Thus, the system GMM relies on the use of appropriate instruments even when the independent
variables are highly persistent. The system GMM relies on the use of additional moment conditions. The overall
validity of the instruments used is central to the GMM estimator as they ensure consistency (Carkovic and Levine,
2005). To test the validity of instruments used, two specification tests as indicated by Arellano and Bond (1991)
and Blundell and Bond (1998) must be satisfied. The first is the test for overidentifying restrictions, which analyzes
the analog of the moment conditions employed in the estimation. The second test examines the hypothesis that the
error term is not serially correlated. The Sargan test is used for difference GMM while the Hansen J is used for
the system GMM estimator. The Sargan/Hansen test measures the validity of the instruments by analyzing sample
analogs of the moment conditions used in the estimation. By construction, the error term could be serially
correlated in the first order [AR(1)]. However, second‐order serial correlation is a sign of misspecification, that is,
the second specification test is to ensure that the error term is not serially correlated in the second order [AR(2)].
We further estimate the long‐run effect of the variable of interest on the dependent variable. For a general linear
equation in the form:
Y = a + bx (6)
the long run effect of the short run coefficient, X, is determined as
_b (x) .
[1 − _b (lagY )] (7)
We thus estimate the long‐run effects of the impact of FDI on the growth of the real sector and GDP growth. Is
it the case the impact of FDI on growth is only in the short run or long run, or both? From Equation (1), we
estimate the long‐run relationships of the impact of FDI on growth on the real sector by the application of
Equation (8):
ASAMOAH ET AL. | 5
_b (FDI) .
[1 − _b (L. RSG)] (8)
4 | EMPIRICAL RESULTS
We first discuss the effect of FDI on the real sector growth index and the individual components. We then discuss how
real sector growth and its components contribute to the attraction of FDI.
4.1 | Real sector growth results
We present the core results of our regression in Table 1. The independent variable of interest is FDI. For the robustness
check, we also regress FDI on GDP growth and the associated controls.
Given the dynamic nature of our estimator, we found the lag of both real sector growth and GDP growth to be
positive and significant. As consistent with Rodman (2009), the coefficient of the lagged dependent variables falls
TABLE 1 Effect of foreign direct investment (FDI) on growth
Real sector growth index GDP growth
Dependent variable Model 1 Model 2
Constant 5.722 (1.604)*** 50.648 (3.204)***
Lag RSG 0.168 (0.028)***
Lag GDP growth 0.464 (0.111)***
Foreign direct investment 0.351 (0.108)*** 0.801 (0.107)***
M2 −0.106 (0.026)*** −5.327 (0.845)***
Inflation 0.371 (0.442) −0.956 (0.312)***
Natural resources −0.340 (0.045)*** −0.303 (0.732)***
Exchange rate 0.885 (0.135)*** 0.011 (0.001)***
Gross domestic savings 0.130 (0.037)*** 0.323 (0.565)***
Trade openness 0.045 (0.009)*** −0.066 (0.236)***
Human capital −1.781 (1.326) −3.021 (0.236)***
Gross fixed capital −0.120 (1.604)* −0.286 (0.075)***
Diagnostics
Observations 626 698
Number of groups (n) 33 31
Number of instruments (i) 31 30
Instrument ratio (n/i) 1.03 1.03
AR (1): z (p‐value) −3.56 (0.000) −3.45 (0.001)
AR (2): z (p‐value) −1.50 (0.133) 0.45 (0.651)
Hansen J: χ (p‐value) 26.63 (0.146) 25.39 0.115
Wald χ: p‐value 0.000 0.000
Note: Standard errors are presented in parentheses. From the table, we estimate the determinants of growth proxy by the growth of the real sector (RSG). FDI is
the variable of interest. AR (1) = Test of first‐order autocorrelation; AR (2) = test of second‐order autocorrelation; Hansen J = test of overidentifying
restrictions. *, **, *** denotes significance levels at 1, 5 and 10% respectively.
Abbreviation: GDP, gross domestic product.
6 | ASAMOAH ET AL.
within the accepted range of 0–1. The robustness of the GMM estimation as examined by the Hansen J test supports the
validity of the instruments used.
From Table 1, Model 1 shows FDI to exhibit a positive and significant relationship with the growth of the real sector
over the study period. The results suggest that higher inflows of FDI are necessary for the growth of SSA's real sector.
Inflows of FDI will aid and fast track the growth of these sectors. Thus, at the lower aggregate level, growth will
increase with an anticipated increase in the flow of FDI. Our results indicate that a unit increase in the FDI will lead to
a 0.351 units increase in real sector growth. Our results show that direct channels of FDI into the real sector should be
encouraged. For the robustness check, we also regress FDI on GDP growth as shown in Model 2 of Table 1. The results
show an overwhelmingly positive association between GDP growth and all FDI at the 1% significance level. Our
conclusions on FDI confirm earlier assertions of a positive association by Okafor et al. (2017), Iamsiraroj (2016) and
Alley (2015). It goes on to suggest that there is no evidence of an allocation puzzle at the overall level of real sector
growth and FDI into SSA. The relevance of our studies is for countries to continue to pursue policies aimed at
enhancing growth as increased growth will attract large volumes of FDI. The relative impact of our control variables
can be observed from the above table.
We further discuss the long‐run impact of FDI on the growth of the overall real sector, and GDP growth. We focus
on growth at the overall level of real sector.
From Table 2, we observe that FDI has a significant long‐run relationship with the overall growth of the real sector.
We observe that the long‐run effect on FDI on the growth of the real sector is positive with a coefficient of 0.42. Thus a
1% change in FDI inflows into SSA is associated with a 0.42% expansion in the growth of the real sector in the long run
at a 1% significance level. The results from GDP growth show a similar trend, where the long run impact of FDI on
growth is significant and positive at a 1% significance level. The results imply that a 1% sugre in FDI inflows will
ehance GDP growth by 1.2%. Again, we observe that the overall impact of FDI inflows on GDP growth in the long run
(1.244) exceeds that of the short run (0.801).
4.2 | Foreign direct investment and real sector growth components
We further assess the effect of FDI on the components of the real sector. The results are presented in Models 1–4 of
Table 3. The coefficient of the lagged dependent variables falls within the accepted range of 0–1.
Table 3, Models 1–4 show a significant positive relationship between FDI and manufacturing, industrial and service
sector value additions but a negative relationship with agriculture value additions. The positive association with both
industrial and service sectors is at a significant level of 1% while that of manufacturing is weakly significant at 10%. The
positive relationship could result from the positive spillover effects and technological transfers that host countries enjoy
through FDI. Large inflows of FDI could be a strong catalyst for the growth of the industrialization agenda given the
positive relationship between FDI and manufacturing and industrial value additions. The results show that a unit
increase in the flow of FDI will lead to a 1.024 unit and a 0.462 unit growth in manufacturing and industrial value
additions, respectively.
Furthermore, the higher coefficient for manufacturing suggests that manufacturing as a component of industrial
value additions could lead the way for SSA industrialization. Our finding corroborates that of Kodongo and Ojah
(2017), who also found FDI flows to positively affect industrial and manufacturing value additions for a set of 19
African countries between 1990 and 2013. They postulate that the positive impact could result from spill‐over effects
attributable to technological transfers from the originating FDI countries. Alfaro (2003) also shows the positive impact
of FDI on manufacturing, and thus supports the assertion that externalities that emanate from FDI in the form of
technological and managerial know‐how tend to favor investments in the manufacturing and service sectors. Our
results also support earlier studies that have found positive spillover effects of FDI in manufacturing and at the industry
TABLE 2 Long‐run results on the impact of FDI on real sector growth
Long run Coefficient Standard error z P> z
Real sector growth 0.422 0.127 3.33 0.001
GDP growth 1.244 0.434 2.87 0.004
Abbreviation: GDP, gross domestic product.
ASAMOAH ET AL. | 7
TABLE 3 Effect of FDI on real sector growth components
MANVA AGVA INVA SERVA
Dependent variable (1) (2) (3) (4)
Constant −16.512 (7.828)** −12.091 (3.586)*** 1.38 (0.881) 8.560 (3.001)***
Lag dependent variable 0.493 (0.081)*** 0.205 (0.064)*** −0.017 (0.007)*** −0.092 (0.469)**
FDI 1.024 (0.599)* −0.878 (0.345)** 0.462 (0.070)*** 2.011 (0.390)***
GDP growth 0.770 (0.753)*** 1.180 (0.185)*** 0.283 (0.018)*** 0.630 (0.067)***
M2 1.090 (1.309) 0.391 (0.072)*** −0.329 (0.279) −1.144 (0.574)**
Inflation 0.086 (0.397)** 2.482 (0.516)*** 0.001 (0.003) 0.001 (0.004)
Natural resources −0.206 (0.075)*** 0.656 (0.360)* −0.191 (0.092)** −0.083 (0.231)***
Exchange rate −0.408 (0.362) −0.058 (0.152) 0.047 (0.026)* −0.319 (0.249)
Gross domestic savings 0.211 (0.053)*** 0.281 (0.049)*** −0.027 (0.009)*** 0.054 (0.020)***
Trade openness −0.056 (0.199)*** −0.179 (0.038)*** −0.005 (0.005) −0.027 (0.137)**
Financial openness −0.353 (0.536) −1.483 (0.658)** −0.547 (0.077)*** −0.893 (0.347)**
Diagnostics
Observations 631 573 620 633
Number of groups (n) 35 39 39 36
Number of instruments (i) 30 34 32 30
Instrument ratio (n/i) 1.17 1.15 1.22 1.20
AR (1): z (p‐value) −3.36 (0.001) −3.91 (0.000) −4.00 (0.000) −2.92 (0.004)
AR (2): z (p‐value) 1.36 (0.173) −0.60 (0.551) −0.61 (0.545) −1.34 (0.180)
Hansen J: χ (p‐value) 20.83 (0.185) 27.36 (0.241) 26.80 (0.178) 18.91 (0.273)
Wald χ: p‐value 0.000 0.000 0.000 0.000
Note: AGVA, agricultural value additions; FDI, foreign direct Investment; INVA, industrial value additions; MANVA, manufacturing value additions; SERVA,
service value additions. *, **, *** denotes significance levels of 10%, 5% and 1% repectively.
GDP, gross domestic product.
level, such as Borensztein et al. (1998) on a set of 69 developing economies at the industry level. The results support
assertions by United Nations Industrial Development Organization (2015) and United Nations Conference on Trade
and Development (2016) on the relevance of the manufacturing and industrial sectors to sustained growth in
developing countries. The findings go against some propositions that have sought to describe SSA as lacking any
industrialization agenda. Gui‐Diby and Renard (2015) postulate that FDI inflows do not have any significant effect on a
country's industrialization. Gui‐by and Renard hypothesize that a major contributing factor could be the government's
inability to create a conducive environment for FDI to affect industrialization. Other studies have also found FDI to
negatively affect or have no significant effect on the growth of the manufacturing and industrial sectors. Some attribute
this to the high level of competition from foreign investors which ends up crowding out local firms (Waldkirch &
Ofosu, 2010; Xu & Sheng, 2012).
On the effect of FDI on the service sector, contrary to Kodongo and Ojah (2017), who found FDI to affect service
value additions negatively, we found a strong positive relationship at the 1% significant level, with a unit increase in
FDI flows leading a 2.0 unit in the growth of the service sector. It shows the potency of FDI in improving the growth of
the service sector in SSA. Our findings also depart from those of Alfaro (2003), who found an ambiguous relationship.
The positive effect on the service sector is likely to be a result of the fact that most foreign investors have realized the
growing size of the services sector in the world economy and are thus turning attention to the sector. United Nations
(2011) notes a change in the direction of FDI to developing economies from manufacturing to the services sector over
the past two decades. By the end of 1989, manufacturing accounted for almost 52% of FDI to developing countries
while services accounted for only 35%. However, by the beginning of 2008, the services sector FDI had risen to 49%
while that of manufacturing had dropped to 37% (United Nations, 2011). Also, the higher coefficient gives an indication
8 | ASAMOAH ET AL.
of the recent direction of FDI into SSA and supports recent assertions that the services sector can lead SSA's growth
(African Development Bank, 2018). The results confirm the supposed interest of foreign investors in the services sector
of most SSA countries or the expansion of existing investors into services in other countries. Inflows of foreign capital
in SSA in recent times is evident in the telecom, banking, insurance and education as well as transportation industries.
The positive relationship is also an indication that the flow of FDI into traditional extractive industries could be
gradually changing. It is therefore essential that the challenges facing the services sector such as infrastructure,
logistics, human capital and finance are resolved on a country‐level basis, as the sector is deemed essential to the
development agenda of the sub‐region. United Nations (2011) and Bhinda and Martin (2009) note a significant shift in
the direction of FDI to the services sector.
We further found a significant inverse relationship between FDI and agriculture value additions, at a 5% significance
level. The effect of a 1% increase in the inflow of FDI will be a decrease of 0.878% in the growth of agriculture value
additions. Just like Kodongo and Ojah (2017), the seemingly negative relationship illustrates the type of FDI that is known
to flow into SSA. Most foreign investments go into the extractive‐ and natural resource‐endowed areas of the continent.
AfDB/OECD/UNDP (2014) notes that more than 95% of FDI that came to Africa went to natural resource‐endowed
countries. In a continent where most farmers do not own lands to farm, farmland is likely to be sold for the extraction of
natural resources; this may account for the significant negative relationship. Agriculture may therefore not be the right
channel to attract foreign capital into SSA. As noted by United Nations (2011), between 1990 and 2007, FDI flows into the
primary sector for most developing countries could be found in the mining, quarrying and petroleum sectors, with little or
decreasing FDI to the agricultural sector. Bhinda and Martin (2009) notes that the agriculture sector has suffered from
underinvestment owing to factors such as poor information available to foreign investors in the sector, issues of land rights,
lack of credit and infrastructure. Again, to document the effect with our data, a 1 standard deviation increase in the flow of
FDI (if standard deviation = 5.0) will lead to a 4.39 percentage point drop in the growth of the agricultural sector
(coeffecient of FDI in Model 2, Table 3 × the standard deviation of FDI; −0.878 × 5.0 =−4.39).
4.3 | Foreign direct investment results
The results from the FDI model are presented and discussed. We examine the impact of the real sector growth index
and its components on the FDI. We do this for two reasons: to test for evidence of the puzzle from the point of capital
flows as in Gourinchas and Jeanne (2013) and to test for any possibility of two‐way relationships between real sector
growth, its components and FDI inflows. We first examine the effect of the real sector index on FDI and then the effect
of each part on FDI. For robustness, we so do the same with GDP growth as an explanatory variable, which was the
same as in MacDonald (2015). The results of the effect of the real sector index and GDP growth on FDI are shown in
Table 4.
Under all models, we observe the reinforcing capacity of FDI. Model 1 shows that there is a strong positive
relationship from regressing real sector growth index on FDI, a situation that is in line with the conclusions of the
neoclassical growth theory. At a 1% significance level, a 1% increase in the growth of the real sector will lead to a 0.4%
increase in the attraction of FDI to SSA. The consistency of our results is supported by Model 2, where at a 1%
significance level, a 1% increase in GDP growth will lead to a 0.3% increase in the flow of FDI to SSA. Our results
confirm conclusions by MacDonald (2015) that growth enhances the attraction of FDI and those by Gourinchas and
Jeanne (2013), who found growth to correlate with all forms of capital flows positively. Overall, the results in Models 1
and 2 of Table 4 suggest that there is no evidence of the allocation puzzle between FDI and either growth of the real
sector or GDP in SSA. Thus, the more a country grows, the more likely it is to be able to attract foreign capital. The
implication thus is that, as stated, faster growing countries will need more funds for developmental and investment
purposes. The anticipated growth also assures investors of possible returns. The result is very relevant as studies are
convinced that Africa's growth hinges on increased inflows of FDI and aid inflows in addition to improved
macroeconomic policies (McKinsey, 2012). Again, the findings are robust to various controls of FDI. From the point of
FDI, we found no evidence to support recent claims of “an allocation puzzle” regarding the directional flow of capital
flows about growth, especially regarding SSA countries.
Model 3 shows the effect of value additions in manufacturing, agriculture, industry and services on the attraction of
FDI. We found a strong positive relationship between all value additions and the inflow of FDI. The effect was at a 1%
significance level for value additions in agricultural, industrial and services with a 10% significance for industrial value
additions. The intuition is that when these sectors grow in SSA, they are most likely to attract more FDI. Their growth
ASAMOAH ET AL. | 9
TABLE 4 FDI, real sector growth and its components
FDI FDI FDI
Dependent variable Model 1 Model 2 Model 3
Constant −2.292 (1.654) −0.428 (1.704) −10.198 (1.856)***
Lag dependent variable 0.818 (0.067)*** 0.549 (0.071)*** 0.480 (0.044)***
Real sector growth index 0.401 (0.066)*** — —
GDP growth — 0.311 (0.055)*** —
Agricultural value added — — 0.221 (0.040)***
Industrial value added — — 0.172 (0.058)***
Manufacturing value added — — 0.058 (0.031)*
Service value added — — 0.141 (0.053)***
M2 −0.075 (0.024)*** −0.020 (0.020) 0.075 (0.014)***
Inflation 0.002 (0.0006)*** 0.002 (0.0003)*** −0.0002 (0.0002)
Natural resources 0.143 (0.035)*** 0.203 (0.018)*** 0.065 (0.017)***
Exchange rate −0.006 (0.002)*** −0.001 (0.015) 0.001 (0.0002)***
Gross domestic savings −0.257 (0.040)*** −0.197 (0.021)*** −0.020 (0.018)
Trade openness 0.018 (0.235) 0.003 (0.023) 0.042 (0.010)***
Financial openness 0.459 (0.272)* 0.586 (0.247)** 0.634 (0.237)***
Human capital 1.072 (1.102) −1.340 (1.223) —
Gross FCF 0.384 (0.088)*** 0.372 (0.777)*** —
Diagnostics — — —
Observations 738 729 838
Number of groups (n) 31 31 37
Number of instruments (i) 30 30 37
Instrument ratio (n/i) 1.03 1.03 1.00
AR (1): z (p‐value) −2.94 (0.003) −2.87 (0.004) −2.16 (0.031)
AR (2): z (p‐value) 0.87 (0.384) 1.51 (0.132) 1.04 (0.299)
Hansen J: χ (p‐value) 23.76 (0.126) 19.31 (0.311) 25.16 (0.289)
Wald χ: p‐value 0.000 0.000 0.000
Note: AGVA, INVA, MANVA, SERVA, FDI, and all of the variables are as defined in Table 3. *, **, *** denotes significance levels at 10%, 5% and 1%
respectively.
is thus a bait for foreign investors to inject fresh or additional capital into these sectors and the larger economy. The
results show that a 1% increase in the growth of the agriculture, manufacturing, industry and service sectors will lead to
a 0.2, 0.1, 0.2 and 0.1% increase in the flow of FDI in SSA. The effect of controls on the attraction of FDI is mixed under
the various models. Consistent with GMM estimations, the Hansen J‐test shows that all conditions relating to
orthogonality have been met in all estimations, meaning that our models are properly estimated. AR (2) also shows that
conditions relating to second‐order autocorrelation have been satisfied.
5 | POLICY RECOMMENDATION AND CONCLUSION
From our studies, our results show that policy‐makers should not be worried about growing the real sector amidst fears
of low inflows of FDI, as stipulated by the allocation puzzle. Our results support earlier studies that have recommended
that policies be instituted aimed at attracting more FDI. The puzzle does not even exist at a disaggregated level of
10 | ASAMOAH ET AL.
growth as we found a positive bi‐directional relationship between FDI and manufacturing, services, and industrial
value additions. The only evidence of the puzzle is seen when we regress FDI inflows on growth of the agricultural
sector, meaning that high growth of the sector deters the attraction of FDI.
On the issue of a bi‐directional relationship, we have established that there is a two‐way link between FDI
and overall growth in the real sector. We recommend that policies aimed at improving the growth of the real sector
should be strengthened. Among them are included a dynamic and resilient agricultural productivity, modernized and
mechanized farming and access to credit by all sectors, in terms of the agriculture sector, a vibrant manufacturing
and industrial sector, strong pursuit of SSA's industrialization agenda, and increased efficiency in mining and
construction for the manufacturing and industrial sectors. Government should increase budgetary allocations to all
sectors and laws to sustain and protect the real sector as well as create direct linkages among all sectors will enhance
the inflow of private capital and especially FDI to the sub‐region. In the same way, policies aimed at attracting FDI
such as capital account liberalization, a stable macroeconomic environment (controlled inflation, money supply,
exchange rate and budget surplus), trade and financial openness, human capital development and a favourable
investment and business climate may go a long way to enhance the growth of the real sector. Also, given that the
impact of FDI on the real sector is much greater in the long run, policy‐makers should not be myopic when drafting
FDI‐real sector related policies. Such policies should be a long‐term focus more than a short‐term one.
Again, FDIs have a strong positive two‐way relationship with growth in services, industry and manufacturing. It is
essential that issues affecting these sectors are resolved to help attract more FDI. Direct efforts aimed at improving the
agriculture sector, such as issues of land, credit and access to markets, among others, are essential in attracting FDI, as
the sector was found to be detrimental to the attraction of FDI, but FDI boosts the growth of the sector. We expect that
the outcome of this study will aid policy‐makers to embark on varied growth measures, especially at a less
disaggregated level. Such consented growth efforts will yield a commensurate amount of FDI into SSA.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon request.
REFERENCES
AfDB/OECD/UNDP. (2014). African Economic Outlook 2014: Global Value Chains and Africa's Industrialisation. OECD Publishing.
https://doi.org/10.1787/aeo‐2014‐en
African Development Bank. (2018). African economic outlook report 2018. The African Development Bank Group. Abidjan, Cote d'Ivoire.
https://www.afdb.org/fileadmin/uploads/afdb/Documents/Publications/African_Economic_Outlook_2018_‐_EN.pdf
Alfaro, L. (2003). Foreign direct investment and growth: Does the sector matter (pp. 1–31). Harvard Business School.
Alfaro, L., Kalemli‐Ozcan, S., & Volosovych, V. (2014). Sovereigns, upstream capital flows, and global imbalances. Journal of the European
Economic Association, 12(5), 1240–1284.
Alley, I. (2015). Private capital flows and economic growth of Sub‐Saharan African countries. African Development Review, 27(4), 469–483.
Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment
equations. The Review of Economic Studies, 58(2), 277–297.
Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error‐components models. Journal of Econometrics,
68(1), 29–51.
Asamoah, M. E., & Alagidede, I. P. (2020). Exploring the causal relationships and allocation puzzle between portfolio investments and real
sector growth in Sub‐Saharan Africa. Research in International Business and Finance, 52, 101187.
Asamoah, M. E., & Alagidede, I. P. (2021). Foreign direct investment, real sector growth and financial development. International Journal of
Finance and Economics, 2021, 1–18.
Asamoah, M. E., Alagidede, I. P., & Adu, F. (2022). Exchange rate uncertainty and foreign direct investment in Africa: Does financial
development matter? Review of Development Economics, 26, 878–898.
Beri, P. B., & Nubong, G. F. (2021). Impact of bilateral investment treaties on foreign direct investment in Africa. African Development
Review, 33(3), 439–451.
Bhinda, N., & Martin, M. (2009). Private capital flows to low‐income countries: Dealing with boom and bust. FPC CBP Debt Relief
International.
Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1),
115–143.
Borensztein, E., De Gregorio, J., & Lee, J. W. (1998). How does foreign direct investment economic growth? Journal of International
Economics, 45(1), 115–135.
Carkovic, M., & Levine, R. (2005). Does foreign direct investment accelerate economic growth? In T. H. Moran, E. M. Graham, &
M. Blomstörm (Eds.), Does foreign direct investment promote development? (pp. 195–220). Institute for International Economics,
Peterson Institute.
ASAMOAH ET AL. | 11
Chinn, M. D., & Ito, H. (2008). A new measure of financial openness. Journal of Comparative Policy Analysis, 10(3), 309–322.
Ernst & Young. (2013). Ernst & Young's 2012 attractiveness survey: Africa 2013. Getting Down to Business, Ernst & Young Global Limited.
Gourinchas, P. O., & Jeanne, O. (2013). Capital flows to developing countries: The allocation puzzle. Review of Economic Studies, 80(4),
1484–1515.
Gui‐Diby, S. L., & Renard, M. F. (2015). Foreign direct investment inflows and the industrialization of African countries. World
Development, 74, 43–57.
Iamsiraroj, S. (2016). The foreign direct investment–economic growth nexus. International Review of Economics & Finance, 42, 116–133.
Kholdy, S., Sohrabian, A. (2005). Financial markets, fDI, and economic growth: Granger causality tests in panel data model. SSRN Electronic
Journal. https://doi.org/10.2139/ssrn.676085
Kodongo, O., & Ojah, K. (2017). Cross‐border capital flows and economic performance in africa: A sectoral analysis. In E. Wamboye & E.
Tirune (Eds.), Foreign capital flows and economic development in Africa (pp. 163–190). Palgrave Macmillan.
Levchenko, A., & Paolo, M. (2007). Do some forms of financial flows help protect against “sudden stops”? World Bank Economic Review,
21(3), 389–411.
Lucas, R. E. (1990). Why doesn't capital flow from rich to poor countries? American Economic Review, 80(2), 92–96.
Luca, M. O., & Spatafora, M. N. (2012). Capital inflows, financial development, and domestic investment: determinants and inter‐
relationships. International Monetary Fund. 1–22.
Lynch, S. M., & Brown, J. S. (2011). Stratification and inequality over the life course. In R. H. Binstock & L. K. George (Eds.), Handbook of
aging and the social sciences (7th ed., pp. 105–117). Academic Press.
MacDonald, M. (2015). Patterns of international capital flows and productivity growth: New evidence. Review of International Economics,
23(5), 846–872.
Mankiw, N. G., Romer, D., & Weil, D. N. (1992). A contribution to the empirics of economic growth. The Quarterly Journal of Economics,
107(2), 407–437.
McKinsey, G. I. (2012). Manufacturing the future: The next era of global growth and innovation. McKinsey Global Institute.
Moraghen, W., Seetanah, B., & Sookia, N. (2021). Impact of exchange rate and exchange rate volatility on foreign direct investment inflow
for Mauritius: A dynamic time series approach. African Development Review, 33(4), 581–591.
Obstfeld, M. (2012). Does the current account still matter? (No. w17877). National Bureau of Economic Research.
Okafor, G., Piesse, J., & Webster, A. (2017). FDI determinants in least recipient regions: The case of sub‐Saharan Africa and MENA. African
Development Review, 29(4), 589–600.
Onye, K., Bassey, G., Iriabije, A., & Effiom, L. (2022). Does aid matter for FDI in ECOWAS: Investigating the infrastructure channel. African
Development Review. 34, 1–15. https://doi.org/10.1111/1467‐8268.12621
Organisation for Economic Co‐operation and Development. (2008). Organisation for economic cooperation and development benchmark
definition of foreign direct investment—4th Edition. Report Published by Organisation for Economic Co‐operation and Development.
https://www.oecd.org/daf/inv/investmentstatisticsandanalysis/40193734.pdf
Prasad, E. S., Rajan, R. G., & Subramanian, A. (2007). Foreign capital and economic growth (No. w13619). National Bureau of Economic
Research.
Rodman, D. (2009). “How to do xtabond2: An introduction to difference and system GMM in Stata”. The Stata Journal, 9(1), 86‐136.
Schroth, J. (2016). Capital flows to developing countries: Is there an allocation puzzle? (No. 2016‐53). Bank of Canada Staff Working Paper.
Solow, R. M. (1956). A contribution to the theory of economic growth. The Quarterly Journal of Economics, 70(1), 65–94.
Taylor, R. S. (2020). Foreign direct investment and economic growth. Analysis of sectoral foreign direct investment in Tanzania. African
Development Review, 32(4), 699–717.
United Nations. (2011). Towards human resilience: Sustaining MDG progress in an age of economic uncertainty. United Nations Publications.
United Nations Conference on Trade and Development. (2016). Economic development in Africa report, 2016: Debt dynamics and
development finance in Africa. United Nations.
United Nations Industrial Development Organization. (2015). Industrial development report 2016. The role of technology and innovation in
inclusive and sustainable industrial development overview.
Waldkirch, A., & Ofosu, A. (2010). Foreign presence, spillovers, and productivity: Evidence from Ghana. World Development, 38(8),
1114–1126.
Xu, X., & Sheng, Y. (2012). Productivity spillovers from foreign direct investment: Firm‐level evidence from China. World Development,
40(1), 62–74.
How to cite this article: Asamoah, M. E., Alagidede, I. P., & Adu, F. (2022). On the allocation puzzle and
capital flows: Evidence from foreign direct investment and real sector growth in Africa. African Development
Review, 1–11. https://doi.org/10.1111/1467‐8268.12656