^{1}

^{1}

^{2}

This study investigates whether foreign aid (AID) has a significant influence on economic growth in WAEMU’s (West African Economic and Monetary Union) countries. We use two (2) types of aid data: aggregate aid and disaggregate aid (aid in education, aid in agriculture, aid in trade policies and regulations and humanitarian aid) to run two (2) different regressions. Both the within-dimension and between-dimension estimators reveal that in the long run, the effect of AID on economic growth is heterogeneous across sectors and aid in agriculture, aid in trade policies and regulations as well as aid in education encourages economic growth.

Debate concerning the relationship between foreign aid and economic growth in the receiving country is subject to different points of view. Several studies have been carried out in order to estimate the impact of foreign aid on the economic growth.

The famous work of [

[

Then, the study of [

Moreover, [

The paper of [

In the same context [

As contribution, the present study considers the heterogeneous nature of aid by disaggregating aid’s data into different sectors (aid in education, aid in agriculture, trade policies and regulations aid and humanitarian aid) because different types of aid may have different effects on growth. It also provides new results on the long-run effect of the foreign aid on the economic growth using a large data set on WAEMU’s countries. All the eight (8) countries are taken over the period of 2002-2013 (see Appendix for the list of countries). These data are essentially based on availability, which nevertheless provides a greater number of observations hence better and more reliable estimates. Finally, it makes use of robust econometric techniques, to solve problems such as endogeneity as well as possible biases. These techniques take into account the possibility of cross-country heterogeneity in the coefficients of AID.

The rest of the paper is structured as follows: Section 2 deals with the data and model; the empirical results are discussed in Section 3 whereas Section 4 concludes.

Our dataset contains annual data on GDP growth rate (GDPR), aggregate aid (Aid), aid in agriculture (AGR), aid in trade policies and regulations (TPR), humanitarian aid (HUM) and aid in education (EDU), from all the eight (8) WAEMU’s countries covering the period 2002-2013 (see Appendix for the list of countries). Foreign aid disbursement data are from the creditor reporting system (CRS) on the OECD website while the data on GDP growth rate and GDP are from the UNCTAD.

In order to investigate the long-run impact of AID on GDPR, we consider the following equation:

where Aid, AGR, TPR, HUM and EDU are shares of GDP; α_{i} is a country-specific fixed effect, β_{1i}, β_{2i}, β_{3i} and β_{4i} may or may not be homogeneous across i and ɛ_{it} stands for error term.

As is a requirement, the first step is to check the univariate properties of the variables, precisely whether the series are stationary at level. In case they are non-stationary, their first difference series should be.

The next step is to test for cointegration since we previously found first order stationary variables. Thus, we use the panel and group ADF and PP t-tests developed by [

In order to estimate the long run effect of the different types of aid on GDPR, and assess the robustness of the results; we employ both Within-dimension Fully Modified OLS (FMOLS) and Dynamic OLS (DOLS) panel estimators which allow for homogeneous coefficient across countries and Between-dimension FMOLS and DOLS panel estimators that consider heterogeneity of the coefficient across countries.

However, before performing DOLS in both cases; Equations (1) and (2) are augmented with leads and lags of the first-differenced aid variables. This serves to control for endogeneity problem. Hence, the new equations are given as follows:

where β_{.i} = β_{1i}, β_{2i}, β_{3i} and β_{4i}; AID stands for AGR, TPR, HUM and EDU; and

As developed by Pedroni (2001); the between-dimension estimators (FMOLS and DOLS) are constructed as:

Variables | Deterministic terms | IPS statistic | ADF Fisher statistic |
---|---|---|---|

Levels | |||

(Aid/GDP)it | Constant | −2.51302 | 111.603 |

(AGR/GDP)it | Constant | 3.77751 | 88.1204 |

(TPR/GDP)it | Constant | −114.884 | 144.765 |

(HUM/GDP)it | Constant | 0.16035 | 137.521 |

(EDU/GDP)it | Constant | 2.13143 | 73.1593 |

First differences | |||

Δ(Aid/GDP)it | Constant | −13.0403^{*} | 293.747^{*} |

Δ(AGR/GDP)it | Constant | −2.19512^{*} | 194.380^{*} |

Δ(TPR/GDP)it | Constant | −120.375^{*} | 180.313^{*} |

Δ(HUM/GDP)it | Constant | −3.00551^{*} | 187.147^{*} |

Δ(EDU/GDP)it | Constant | −11.2402^{*} | 103.301^{*} |

Note: two lags were selected to adjust for autocorrelation. ^{*}Indicate significance at the 5% level of significance.

where ^{th} country. Likewise, the associated t-statistic is:

Since the main interest is to estimate the coefficient of the long-run effect; first, we perform the pooled FMOLS proposed by Pedroni (1996) and the pooled DOLS of Kao and Chiang (1997). The results are reported in

Then, the group-mean panel FMOLS and DOLS estimators developed by Pedroni (2001) were run; the outputs are reported in the third and fourth column of

Pedroni (1999) | Aggregate aid | Disaggregate aid |
---|---|---|

Panel PP t-statistic | −11.23613^{*} | −4.301020^{*} |

Panel ADF t-statistic | −2.01403^{*} | −6.948675^{*} |

Group PP t-statistic | −13.01415^{*} | −18.91068^{*} |

Group ADF t-statistic | −0.73197^{*} | −5.101243^{*} |

Kao (1999) | ||

ADF t-statistic | 4.73521^{*} | −9.241825^{*} |

Note: the number of lags is based on the Schwarz information criterion with a maximum number of five. ^{*}Indicate a rejection of the null hypothesis of no cointegration at the 5% level of significance.

Variables | Within-dimension estimators | Between-dimension estimators | ||
---|---|---|---|---|

FMOLS | DOLS | FMOLS | DOLS | |

Aid/GDP | −0.008018 | −0.014976 | −0.205418 | −0.202830 |

(0.0811) | (0.0605) | (0.0697) | (0.0718) | |

AGR/GDP | 0.0361 | 0.0236 | 0.0451 | 0.0252 |

(0.0098)^{*} | (0.0084)^{*} | (0.0250)^{*} | (0.0338)^{*} | |

TPR/GDP | 0.0207 | 0.0187 | 0.0572 | 0.0265 |

(0.0362)^{*} | (0.0484)^{*} | (0.0173)^{*} | (0.0405)^{*} | |

HUM/GDP | 0.0649 | 0.0265 | 0.0433 | 0.0678 |

(0.9905) | (0.9582) | (0.6930) | (0.9330) | |

EDU/GDP | 0.0663 | 0.0487 | 0.0672 | 0.0531 |

(0.0079)^{*} | (0.0182)^{*} | (0.0434)^{*} | (0.0067)^{*} |

Note: P-values are in parentheses. The DOLS regressions were estimated with one lead and one lag. ^{*}Indicates significance at the 5% level.

indicate a heterogeneous effect of the foreign aid on the economic growth across sectors. More precisely, the point estimates for the coefficients indicate that, on average, any increase (decrease) in the AGR-to-GDP (TPR- to-GDP; EDU-to-GDP) ratio by one percentage point increases (decreases) the GDPR by approximately 0.0361, 0.0236 (0.0207, 0.0187; 0.0663, 0.0487) using the Within-dimension estimators (pooled FMOLS and pooled DOLS respectively) and by 0.0451, 0.0252 (0.0572, 0.0265; 0.0672, 0.0531) percentage points using the Between-dimension estimators (group-mean panel FMOLS and group-mean panel DOLS). In addition, the coefficients are slightly different in terms of magnitude. We therefore, argue having found robust results based on different estimation techniques since the within-dimension and between-dimension estimators produce similar results.

This study examined the effect of aid on economic growth in WAEMU’s countries. We found that the impact of aid on growth depends on the sector in which it is allocated. All the estimation techniques used point out that in the long run, aggregate aid has no significant effect on economic growth while aid directed to specific sectors such as agriculture, trade policies and regulations and education enhances economic growth. Thus, these findings lead us to notice the heterogeneous nature of aid across sectors which can help policy-makers to channel properly foreign aid into significant sectors of the recipient countries. The possibility that some additional factors such as institutional, social and economic, specific to each aid receiving country could be considered as this can shape targeted policies toward enhancing the aid-growth nexus. Moreover, the degree of heterogeneity across countries and the related reasons would be deeply investigated in future researches.

Benin | Niger |
---|---|

Burkina Faso | Senegal |

Côte d’Ivoire | Togo |

Mali | Guinea-Bissau |