Indian Saving Behavior: A Reconciliation Of Time Series And Cross Section Evidence

Serious doubts have been raised as to the validity of using cross‐sectional estimates of the average propensity to consume over time. These d⊙ubts are based on empirical evidence for the United States presented by Kuznets and Goldsmith. This paper extends these considerations to developing countries by looking at the evidence for India. Simple statistical techniques, including ordinary least squares regression, Chow tests, and t‐tests, are used in the estimation of the consumption functions and the formulation of hypotheses. Both India and the U.S. are seen to exhibit the same characteristic secular constancy in the average propensity to consume. For India the average propensity to consume is about 0.95 and is maintained in the face of no substantial secular increases in per capita income during the period under study (1919–1960). The same inconsistencies between time series and cross‐section evidence on the average propensity to consume are found to exist for India. The permanent income theory suggested only a partial explanation of these inconsistencies and the reconciliation was achieved by a Duesenberry type explanation based on evidence for a shifting cross section consumption function over time. The date was provided by a set of Family Living Surveys for Industrial Workers: 1926, 1933–1935, 1950–1952, 1958–1959. Finally it was noted that the cross‐section consumption function in India had shifted both upwards and downwards over the period under study, in contrast to the strict upward shift for the U.S. In an economy such as India's, where secular growth is by no means assured, it is not always likely that consumers can avoid lowering previously achieved consumption standards in the face of cyclical economic conditions.