An Empirical Test For Consumption Hypotheses:Evidence From Sixteen Countries

Aggregate real private consumption is one of the major components of Gross Domestic Product (GDP) for global economies. In developing countries, the aggregate real private consumption take the majority share of GDP up to60to70percent, while in less developing countries it accounts for65to75percent of GDP. However, it contributes little to GDP in developed countries. Aggregate real private consumption is one of the important elements for decision makers to accelerate aggregate economic activities. The main purpose of the study is to identify the characteristics and features of the economies that will help to satisfy one of the consumption hypotheses between Absolute Income Hypothesis (AIH) and Permanent Income Hypothesis (PIH). However, in existing consumption literature, there is no answer to this question that on average which type of economy will satisfy what kind of consumption hypothesis.This study used annual data from1985to2013and investigates the validity of PIH and AIH for Australia, Korea, New Zealand, Singapore, China, Peru, Tunisia, Venezuela, India, Senegal, Sri Lanka, Vietnam, Ethiopia, Somalia, Madagascar and Togo. We applied two types of strategies to test the AIH and PIH i.e. Hall’s Random Walk Hypothesis (RWH)(1978) hypothesis of consumption and Divadson et al.(1978) modelling of consumption function. The Hall’s consumption model shows that current income is the most important determinant of aggregate real private consumption in case of India, Senegal, Sri Lanka, Vietnam, Ethiopia, Somalia, Madagascar and Togo. In case of Australia, Korea, New Zealand, Singapore, China, Peru, Tunisia and Venezuela, besides current income, permanent income also proved to be an important predictor for households’consumption. Moreover, Campbell and Mankiw (1989) consumption model found that the proportion of backward looking households is bigger than the forward looking households in case of India, Senegal, Sri Lanka, Vietnam, Ethiopia, Somalia, Madagascar and Togo, while in case of Australia, Korea, New Zealand, Singapore, China, Peru, Tunisia and Venezuela the forward looking households exceeds than the backward looking households. Jointly the Hall’s consumption model and Campbell and Mankiw (1989) consumption model support the AIH for India, Senegal, Sri Lanka, Vietnam, Ethiopia, Somalia, Madagascar and Togo, while PIH for Australia, Korea, New Zealand, Singapore, China, Peru, Tunisia and Venezuela.The time series analysis of Autoregressive Distributed Lag (ARDL) model shows that the most important determinants of aggregate real private consumption are real GDP, wealth, real interest rate and unemployment rate. The most important determinants of aggregate real private consumption for Australia, Korea, New Zealand, Singapore, China, Peru, Tunisia and Venezuela includes real GDP and wealth in both short run and long run. In case of India, Senegal, Sri Lanka, Vietnam, Ethiopia, Somalia, Madagascar the real GDP is the most important determinant of aggregate real private consumption.The panel ARDL approach to co-integration highlights the determinants of aggregate real private consumption in case of High Income Countries (HIC), Upper Middle Income Countries (UMIC), Lower Middle Income Countries (LMIC) and Low Income Countries (LIC). Moreover, to estimate the Hall’s consumption function and Campbell and Mankiw consumption function with panel data for this purpose panel Pool Ordinary Least Square (POLS) and Two Stage Least Square (2SLS) methods were applied. The results of panel ARDL show that the most important determinants of the aggregate real private consumption are real GDP and wealth. However, for HIC and UMIC in the short run the coefficients of real GDP and wealth are somehow similar or little variation was observed, which shows the both GDP and wealth are important determinants of aggregate real private consumption. In case of LMIC and LIC the coefficients of real GDP is very high than coefficients of wealth, which indicate the validity of AIH for LMIC and LIC. Almost this is in compliance the results reported by Hall’s consumption function and Campbell and Mankiw consumption function. The results of Hall’s consumption function strongly supported the AIH for LMIC and LIC while weakly supported PIH for HIC and UMIC, in the same way the Campbell and Mankiw consumption model shows that the proportion of forward looking households in HIC and UMIC are60%and50%respectively while LMIC contains20%and LIC consists of15%. The results of Campbell and Mankiw consumption model supported the PIH for HIC and UMIC while AIH for LMIC and LIC.