_{1}

^{*}

This study investigates the determinants of private consumption expenditure in Cote d’Ivoire using time series data from 1970 to 2016. The Autoregressive Distributed Lags bounds testing approach to cointegration is employed to depict the presence of a long run relationship between private consumption and its determinants and an error correction model is estimated to derive short run dynamics. The results show the presence of a long run relationship among the selected variables. In the long run, current income, wealth and government consumption expenditure play a positive role in determining private consumption, with the effect of current income being higher. Furthermore, consumption expenditure is negatively affected by inflation rate and real interest rate on deposits. In the short run, only income and wealth appear to have positive effects on private consumption while the effects of government consumption, inflation and interest rate were found to be insignificant. This study provides evidence for government to improve the level of private consumption.

Household consumption expenditure plays a crucial role in determining the welfare of populations and the dynamic effects of economic shocks. In most countries, it constitutes the largest component of gross domestic product. For instance, in Cote d’Ivoire, the share of household final consumption spending in GDP increased from 57.3% in 1970 to 63.5% in 1983 and reached a pic of 73.4% in 1992 [

As a result, the study on determinants of consumption expenditure has been an area of intensive research since Keynes [

Based on these theories, a growing body of empirical studies have been undertaken to estimate consumption functions for many individual countries and groups of countries around the world. Most of these works examined the effect of one factor such as income or interest rate in a bivariate framework, ignoring the effects of other relevant factors such as government expenditure, inflation and financial depth. For instance, [

This review of the literature shows that studies on determinants of private consumption yield mixed results. A general consistency among the reviewed studies is the inclusion of income, inflation and interest rate as key drivers of private consumption. A general conclusion from this literature is that the impact of income on consumption is positive while that of inflation and interest rate is negative. However the role of government expenditure and wealth in explaining private consumption has not been widely explored. It is the intention of this study to fill the gap in the empirical literature. The paper investigates some key driving factors of private consumption expenditure in Cote d’Ivoire. The empirical framework applies the autoregressive distributed lag (ARDL) bounds testing approach to cointegration and error correction model to annual data covering the period from 1970 to 2016.

The remainder of the paper is organized as follows. Section 2 outlines the econometric methodology employed for the empirical analysis. Section 3 reports the empirical findings of the study. Section 4 concludes the study and provides some policy recommendations.

The objective of this work is to estimate a relationship between private consumption expenditure and its macroeconomic determinants. Based on the empirical literature, the econometric model is built upon the following equation:

log C t = α + β 1 log Y t + β 2 log W t + β 3 Z t + μ t (1)

where C_{t} is private consumption, Y_{t} is income, W_{t} stands for wealth, and Z_{t} represents a set of other determinants which captures fiscal policy, liquidity constraints, substitution effects and macroeconomic uncertainty. In this study, Z_{t} includes government consumption expenditure (G_{t}), inflation rate (INF_{t}) and real interest rate (r_{t}).

The impact of income and wealth on consumption expenditure is expected to be positive and smaller than one. The coefficient on government spending is ambiguous. Government expenditure can in some cases crowd in private consumption while in other cases it can crowd it out. The impact of inflation rate is negative because inflation lowers the real value of money balances and financial assets with fixed monetary values. The real interest rate reflects the substitution effects in terms of preference for current or future consumption. A priori, an increase in interest rate induces people to reduce consumption expenditure and increase saving. This substitution effect tends to decrease household consumption expenditure. On the other hand, higher interest rates increase the financial wealth of positive savers which in turn increases their consumption expenditure. Thus, the effect of interest rate is positive when the income effect dominates the substitution effect. The two effects may cancel each other leaving the total effect to be insignificant.

We begin our empirical investigation by examining the time series properties of all variables. This step is necessary to make sure that we do not run spurious regressions. To this end, we apply the unit root tests of [

Δ log C t = θ 0 + θ 1 log C t − 1 + θ 2 log Y t − 1 + θ 3 log W t − 1 + θ 4 Z t − 1 + ∑ i = 1 m 1 γ 1 i Δ log C t − i + ∑ i = 0 m 2 γ 2 i Δ log Y t − i + ∑ i = 0 m 3 γ 3 i Δ log W t − i + ∑ i = 0 m 4 γ 4 Δ Z t − i + μ t (2)

where Z t = ( log G t , I N F t , r t ) .

The presence of long-run relationship among the variables is tested using the F-test statistic with the null hypothesis of no long-run relationship being H_{0}: θ 1 = θ 2 = θ 3 = θ 4 = 0 . The optimal lag structure (m1, m2, m3, m4) was selected using the Akaike Information Criterion (AIC) with maximum lag length on each variable set to five. The model has been tested by the diagnostic tests that are serial correlation, normality and heteroskedasticity tests. Stability tests have also been used to test the goodness of fit of the model. The computed F-statistic has to be compared with two critical values bounds generated by [

The study uses annual data covering the period 1970-2016. Variables used include private consumption measured as household real final consumption expenditure (C_{t}), real GDP used as a proxy for income (GDP_{t}), real financial wealth (W_{t}), real government final consumption expenditure (G_{t}), inflation rate (INF_{t}) measured by the log difference of consumer price index, and real interest rates on deposits (r_{t}) calculated as the difference between nominal interest rate and inflation rate. Given the lack of suitable data for wealth, monetary aggregates are often used to represent financial wealth (see, [

As indicated in the methodology, we begin our empirical analysis by examining the time-series properties of the variables. The results displayed in

The results of the ARDL bounds test are reported in

Variables | logC | logGDP | logW | logG | INF | r |
---|---|---|---|---|---|---|

Panel A: Summary statistics | ||||||

Mean | 29.404 | 29.828 | 28.556 | 27.916 | 6.100 | −1.392 |

Median | 29.338 | 29.774 | 28.456 | 27.968 | 4.020 | 0.422 |

Maximum | 30.158 | 30.481 | 29.583 | 28.427 | 27.421 | 7.805 |

Minimum | 28.676 | 29.167 | 27.907 | 27.237 | −0.805 | −22.581 |

Std. dev. | 0.309 | 0.278 | 0.359 | 0.231 | 6.451 | 6.205 |

Skewness | 0.057 | −0.124 | 1.078 | −0.433 | 1.613 | −1.684 |

Kurtosis | 3.182 | 3.263 | 4.047 | 3.840 | 5.414 | 6.142 |

Jarque-Bera | 0.090 | 0.258 | 11.261 | 2.855 | 31.815 | 41.568 |

Probability | 0.955 | 0.878 | 0.003 | 0.239 | 0.000 | 0.000 |

Panel B: Correlation matrix | ||||||

logC | 1.000 | |||||

logGDP | 0.985* | 1.000 | ||||

logW | 0.800* | 0.774* | 1.000 | |||

logG | 0.819* | 0.802* | 0.759* | 1.000 | ||

INF | −0.392* | −0.368* | −0.134 | −0.256** | 1.000 | |

r | 0.309* | 0.285** | 0.074 | 0.260** | −0.969* | 1.000 |

Note: C = Household real final consumption, GDP = real GDP, W = real M2, G = Government final consumption, INF = Inflation rate as log difference of consumer price index, r = real interest rate on deposits. The asterisks * and ** denote statistical significance at the 5% and 10% levels, respectively.

Series | Level | First difference | Order of integration | ||
---|---|---|---|---|---|

PP | KPSS | PP | KPSS | ||

logC | −2.144 | 0.093 | −5.096* | 0.157 | I(1) |

logGDP | −2.388 | 0.085 | −4.273* | 0.139 | I(1) |

logW | −0.321 | 0.506* | −6.013* | 0.212 | I(1) |

logG | −2.934 | 0.099 | −7.810* | 0.161 | I(1) |

INF | −3.753* | 0.504* | −9.046* | 0.061 | I(0) |

r | −4.331* | 0.055 | −8.922* | 0.074 | I(0) |

Note: C = Household real final consumption, GDP = real GDP, W = real M2, G = Government real final consumption, INF = Inflation rate as log difference of consumer price index, r = real interest rate on deposits. * denotes the rejection of the null hypothesis at the 5% level.

Model | F-stat. | Diagnostic tests | ||
---|---|---|---|---|

Normality^{1 } | Heteroskedasticity^{2} | Correlation^{3} | ||

ARDL (4, 5, 1, 5, 5, 2) | 4.526^{* } | 0.790 | 0.296 | 0.298 |

Critical value | Lower bound value | Upper bound value | ||

1% | 2.82 | 4.21 | ||

5% | 2.14 | 3.34 | ||

10% | 1.81 | 2.93 |

Note: 1) p-value of Jarque-Bera test, 2) p-value of Breusch-Pagan-Godfrey test, 3) p-value of Breusch-Godfrey Serial Correlation LM Test. Lag structure of the ARDL model was selected using the AIC criterion with maximum lag set to 5. Critical values of the F-test were obtained from Pesaran et al. [

the 5% significance level, all diagnostic tests do not exhibit any evidence of violation of the classical linear regression model assumptions. That is the error terms have a constant variance, and are independently and normally distributed.

Since the variables under study are cointegrated, we estimate the long run consumption function. To that end, we apply the ARDL approach along with the Fully Modified OLS method proposed by [

Regressor | Dependent variable: log of private consumption (logC) | |||||
---|---|---|---|---|---|---|

ARDL | FMOLS | DOLS | ||||

Coef. | t-stat. | Coef. | t-stat. | Coef. | t-stat. | |

Income | 0.618* | 10.923 | 0.867* | 11.030 | 0.317* | 2.024 |

Wealth | 0.106* | 3.859 | 0.061 | 1.460 | 0.111** | 1.697 |

Gov. consumption | 0.290* | 4.259 | 0.167* | 2.310 | 0.413* | 6.857 |

Inflation | −0.035* | −8.458 | −0.013** | −1.660 | −0.081* | −4.137 |

Interest rate | −0.026* | −5.647 | −0.009 | −1.361 | −0.067* | −3.848 |

Note: The asterisks * and ** denote statistical significance at the 5% and 10% levels, respectively.

with other studies ( [

The coefficient on government consumption is about 0.3 and significant, implying that an increase of 1% in government final consumption causes household consumption expenditure to rise by 0.3%. This evidence strongly indicates that government expenditure has a complementary relationship with private consumption. Therefore, government expenditure can be used to induce household consumption growth. This finding supports the Keynesian effects of fiscal policy on private consumption.

The impact of inflation on household consumption is negative and significant, implying that increases in the level of inflation lessen the level of goods and services demanded by households. This finding lends support to the real balance effect of inflation. Finally, the effect of interest rate on consumption is negative and significant implying that higher interest rates on deposits reduce private consumption. This finding suggests that the substitution effect outweighs the income effect in respect of the choice between saving and consumption. Our findings for inflation and interest rate are in line with [

We have estimated the error correction model to obtain short run dynamic relationship and results are reported in

Regressor | Dependent variable: Growth rate of private consumption | ||
---|---|---|---|

Coefficient | t-statistic | Prob. | |

ΔIncome | 0.473* | 3.073 | 0.004 |

ΔWealth | 0.162* | 2.805 | 0.008 |

ΔGov. consumption | 0.061 | 0.966 | 0.340 |

ΔInflation | −0.010 | −1.225 | 0.228 |

ΔInterest rate | −0.005 | −0.807 | 0.425 |

ECM (−1) | −0.424* | −3.277 | 0.002 |

Note: The short-run equation includes no constant. The asterisks * denotes statistical significance at the 5% level.

strong effect of current income is consistent with how the economic system prevailed in Cote d’Ivoire and other African countries. Indeed, many African countries experience low liquidity options such as credit unavailability and low savings and most households are unable to use savings and borrowing to smooth their consumption. As a consequence of being financially constrained, they rely on current income.

The coefficients with respect to government consumption, inflation and interest rate were found not to be statistically significant in determining private consumption in the short run. For policy makers, this suggests that government expenditure cannot be used to stimulate private consumption in the short run.

The aim of this study was to investigate the determinants of private consumption expenditure in Côte d’Ivoire over the period from 1970 to 2016. Our empirical analysis employed the bounds test approach developed by [

The short run results show that household final consumption expenditure is mostly driven by income and wealth. The effects of inflation and real interest rate on deposits were found to be insignificant. Noteworthy, the effect of income on consumption in the short run was found to be less than it was in the long run. Moreover, private consumption has been found to be more sensitive to current income as indicated by a higher elasticity. Overall, the results reported in this study support the Keynesian theory of consumption. The excess sensitivity of consumption to current income indicates the existence of a robust income-consumption relation. In such a context, income tax reduction can have a considerable impact on household consumption expenditure. The positive relationship between government spending and private consumption also suggests that government spending can be used to stimulate aggregate demand and economic growth in Cote d’Ivoire in the long run. Also, lower interest rates and inflation are recommended to increase household consumption.

This study does not explain the reasons of the strong relationship between current consumption and current income. Do credits constraints or myopia explain this result? It is important to investigate the validity of these hypotheses to provide valuable policy recommendations to government. We intend to examine this issue in a future research.

The author declares no conflicts of interest regarding the publication of this paper.

Keho, Y. (2019) An Econometric Analysis of the Determinants of Private Consumption in Cote d’Ivoire. Theoretical Economics Letters, 9, 947-958. https://doi.org/10.4236/tel.2019.94061