_{1}

This article analyzes the relation between trade openness and democracy in Gabon. Unlike traditional approaches based on the estimation of linear gravity models, we first estimate a nonlinear model characterizing the relation between democracy and trade openness. We then determine an optimal threshold beyond which democracy has a positive effect on trade openness in Gabon.

Under the structural adjustment programs of the International Monetary Fund, there is renewed interest in the analysis of the relation between trade openness and democracy in developing countries, where there may be conflicts between them two objectives [

Indeed, democracy offers the possibility for lobbies and pressure groups to influence the decisions of the state (central government and bureaucrats). In this respect, it may help to reduce the degree of trade openness [

On the theoretical level, the literature reveals two (2) main points of view. The first refers to the work of Mansfield, Milner and Rosendorff (2000) [

On the empirical level, several studies attempt to characterize the relationship between democracy and trade openness from two (2) methods. The first estimates a decline in the openness of a country on indicators of democracy [

In Africa, it seems that the study of the relation between trade openness and democracy has not attracted much interest until a recent past. This is particularly the case in Gabon, a small country open to the outside and dependent on oil, which is experiencing an acute economic and financial crisis since the drop in prices of this raw material. On the other hand, the future direction of its trade relations with the European Union countries, the Economic Partnership Agreements and the political challenges arising from the last presidential elections, suggest that the country must implement both Economic and political reforms.

This paper refers to Gabon, it’s a contribution to the study of the relations between trade openness and democracy in sub-Saharan Africa (SSA) in that it uses a non-linear model to determine an optimal threshold of democracy. The rest of the study is organized as follows: Section 2 presents the model for determining the optimal threshold for democracy. Section 3 is devoted to empirical analysis. Finally, Section 4 concludes.

Like Niyongabo (2008) [

Trade openness (tope_{t}) is the explanatory variable of the model. It is defined by the sum of exports and imports expressed as a percentage of GDP [

Thus, the equation for determining the degree of trade openness is:

o u v t = α + β X t + ε t (1)

where:

-ouv_{t} is the ratio of exports and imports to GDP in%;

-X_{t} includes all economic and geographic variables;

-α, a constant;

-ε_{t}, an error term.

Like Granger and Siroën (2001), we increase the equation (1) of institutional variables (democracy, in particular) to nonlinear effects.

Our equation [

o u v t = α + a 1 d e m o t + a 2 d e m o t 2 + β X t + ε t (2)

where:

-ouv_{t} is the ratio of exports and imports to GDP in %;

-demo_{t}, is the level of democracy, measured by an index of democracy created in 2006 by “The Economist Group”;

-X_{t}, includes all economic and geographic variables;

-α, a constant;

-ε_{t}, an error term.

We retain a priori five economic and geographic variables X_{t}. The first factor is the population level (pop_{t}) which has a negative theoretical sign: the larger the size of the economy, the more it is self-sufficient or the less it exchanges with the rest of the world. The second explanatory variable is the level of GDP (gdp_{t}) which has a positive effect on the degree of trade openness. The more a rich/ developed economy, the more it trades with the outside in relative terms. The level of development would diminish the effect of distance between markets, both by relative cost mechanisms and by an increased circulation of information [_{t}). Price levels (ipc_{t}) and investment (inv_{t}) represent the fourth and fifth explanatory variables of trade openness. The level of domestic prices appears to have an effect on the degree of trade openness in different ways [

The relation between trade openness and democracy is finally determined from the equation below:

l o u v t = a 0 + a 1 d e m o t + a 2 d e m o t 2 + a 3 l p o p t + a 4 l d g p t + a 5 l g d p f t + a 6 l i p c t + a 7 l i n v t + ε t (3)

where:

-louv_{t}, commercial opening;

-demo_{t}, democracy;

-lpop_{t}, the population;

-lgdp_{t}, the gross domestic product;

-lgdpf_{t}, the gross domestic product of major trading partner;

-lipc_{t}, the price level;

-linv_{t}, the level of domestic investment;

-ε_{t}, an error term.

We optimize equation (3) to determine the optimal threshold of democracy ( d e m o t ∗ ). The determination of the candidate point results from the first order conditions (CPO) and those of second order (CSO):

The first order conditions give the following equation:

∂ l o u v t ∂ d e m o t = 0 (4)

The resolution of equation (4) results in the following formula:

a 1 + 2 a 2 d e m o t ∗ = 0 (5)

where:

- d e m o t ∗ , the optimal threshold for democracy;

-a_{1}, the effect of democracy on trade openness;

-a_{2}, the effect of the square of democracy on trade openness.

The second order conditions (CSO) show that if a_{2} is negative and a_{1} is positive, then the function f is concave and the optimal threshold ( d e m o t ∗ ) corresponds to a maximum. On the other hand, if a_{2} has a positive sign and a_{1} is negative, then the function f is convex and the optimal threshold ( d e m o t ∗ ) corresponds to a minimum.

On the first part we present the methodology; we will proceed on the second part to interpret the results.

The data used to enter the variables comes from the World Bank database, with the exception of the Democracy Index, which comes from the annual reports of the economist intelligence unit. The variables have been transformed into logarithms apart from the democracy index, i.e. louv_{t}, lpop_{t}, lpib_{t}, lpibf_{t}, lipc_{t} et linv_{t}. Moreover, the GDP of Gabon’s main trading partner (pibf_{t}) corresponds to that of France. It remains Gabon’s leading supplier, with 21% market share (2014).

We use historical data to estimate the relationship between democracy and trade openness over the period 2006-2015. In order to broaden the sample size, we proceed with the quarterly data.

The nonlinear model of polynomial form (3) is estimated using the nonlinear least squares method (or NLS). This method has two steps:

The first one consists in verifying that the series lend themselves to a nonlinear estimation based on the analysis of descriptive statistics (Kurtosis and Skewness tests, followed by a graphical analysis). If the elements of descriptive statistics confirm the presence of asymmetries, then it becomes possible to estimate the parameters of the model by the least squares nonlinear (or NLS). The second relates to the estimation of the polynomial by the nonlinear least squares method in order to integrate the nonlinearity of the effects of democracy on the degree of trade openness due to the imperfect competition context characterizing international trade. The model thus estimated is globally validated by the correlation coefficient (R^{2}).

_{t}) and investment (linv_{t}). Skweness statistics are significant at the critical 5% threshold for trade openness (louv_{t}), Gabon’s GDP (lgdp_{t}), France’s GDP (lgdpf_{t}), price level (lipc_{t}) and population (lpop_{t}). Consequently, these variables are a priori asymmetric, and can be assimilated to a non-linearity.

Variables | Pr(Skewness) | Pr(Kurtosis) |
---|---|---|

louv_{t} | 0.0000* | 0.0000* |

demo_{t} | 0.7972 | 0.3204 |

lgdp_{t} | 0.0000* | 0.0001* |

lgdpf_{t} | 0.0000* | 0.0000* |

lpop_{t} | 0.0000* | 0.0001* |

lipc_{t} | 0.0000* | 0.0000* |

linv_{t} | 0.3150 | 0.0000* |

a. (*) the coefficients are significant at the 1%; (**) the coefficients are significant at the 5%; (***) the coefficients are significant at the 10%.

In addition to the asymmetric nature of certain variables, non-linearity is also highlighted in the relationship between trade openness and democracy, as shown in

The demonstration of the nonlinearity of the relationship between commercial openness and democracy leads us to estimate the model by the nonlinear least squares method. The results of this estimate can be seen in the following

The estimation of the model reveals the existence of a convex relationship between democracy and the degree of commercial openness in Gabon, which reinforces the results of Granger and Siroën (2001) and contradicts the work of Milner and Kubota (2005) and Yu (2010) that only countries with high levels of democracy, generally the major industrialized countries, have higher rates of openness.

On the other hand, the gross domestic product (lgdp_{t}) and the population (lpop_{t}) have negative (−0.707 and −2.303) and significant (respectively at the 10% and 1% thresholds) Commercial opening (ltope_{t}). This result confirms

louv_{t} | Coef |
---|---|

demo_{t} | −0.542* |

0.059*** | |

lgdp_{t} | −0.707*** |

lgdpf_{t} | 2.545* |

lpop_{t} | −2.303* |

lipc_{t} | 2.306* |

linv_{t} | 0.312* |

R^{2} = 0.99; Adjusted R^{2} = 0.99. a. (*) the coefficients are significant at the 1%; (**) the coefficients are significant at the 5%; (***) the coefficients are significant at the 10%.

Eichengreen and Leblang (2008)’s conclusion that the larger the size of the economy, the more it is self-sufficient or, what amounts to the same thing, the less it exchanges with the outside world. The increase in domestic gross domestic product reduces the rate of trade openness by acting on the denominator of the ratio. The richer the country, the less incentive it is to interact with the outside world. The gross domestic product of the main trading partner (lgdpf_{t}), the general price level (lipc_{t}) and the investment (linv_{t}) have positive effects (respectively 2.545; 2.306 and 0.312) at the threshold of 1%. This result shows that the resulting increase in imports is largely offset by the joint decline in exports.

From these results, it becomes possible to determine a minimum level of democracy beyond which democracy would have a positive influence on trade openness.

Indeed, the minimum threshold of democracy is obtained by applying the formula [

In this article we studied the relation between trade openness and democracy in Gabon from a nonlinear model. After estimating the direct impact of democracy on trade openness by the non-linear least squares method (or NLS), we have determined an optimal threshold beyond which democracy has a positive effect impact on trade openness in Gabon. More precisely, the results highlight a convex relationship between these two variables. Democracy reduces trade openness at first, and then it increases when the country reaches a hybrid democracy threshold higher than 4.576. However, trade openness, in a context of imperfect competition, exposes the country to external shocks with repercussions on the internal macroeconomic equilibrium of the countries and the monetary union. It might therefore be interesting to study the effect of trade openness on growth, unemployment or inflation in a monetary union such as the Economic and Monetary Community of Central Africa (EMCCA), Composed of small economies open to the outside and whose trade with the countries of the European Union is oriented towards economic partnership agreements.

Ondo, A. (2017) Trade Openness and Democracy: The Case of Gabon. Modern Economy, 8, 191-198. https://doi.org/10.4236/me.2017.82013