Emerging markets

Sociology

Consumption (sociology)

Computer network

Social science

Capital formation

Macroeconomics

Cointegration

Autoregressive model

Econometrics

Distributed lag

Economics

Microeconomics

Profit (economics)

Exchange rate

Computer science

Monetary economics

Financial capital

Depreciation (economics)

International Journal of Finance & Economics

This study re ‐evaluates the relative adjustments of household's consumption expenditure to extremely small and extremely large variations in exchange rate in African emerging economies (AEE). For this purpose, we used a modified version of non‐linear autoregressive distributed lag (ARDL) and multiple threshold non‐linear ARDL models that specifically captures the differential effects of extremely small and extremely large positive and negative variations of the explanatory variable(s) on the explained variable. Consequently, the empirical estimates reveal that changes in exchange rate has asymmetric effects on consumption expenditure in all the countries except in Nigeria. Furthermore, consumption expenditure in Algeria, Egypt and Morocco increases significantly at the upper quantile of exchange rate appreciation, whereas, it declines significantly at all quantiles of exchange rate depreciation, implying that consumption expenditure in these three North African countries are more sensitive to large positive shocks in exchange rate. In Kenya, Nigeria and South Africa, consumption expenditure are exchange rate inelastic as households maintain their consumption levels regardless of the direction and degree of exchange rate deviations. Overall, the diagnostics tests prove the superiority of this enhanced technique over the standard non‐linear ARDL technique. Based on the outcomes of this study, policy recommendations that could ensure optimal exchange rate alignment for the African countries, as well as other emerging economies have been proposed.

Household consumption and exchange rate extreme dynamics: Multiple asymmetric threshold non‐linear autoregressive distributed lag model perspective