Consumers estimate future consumption and give current consumption set in fraction form. It can be in form of wealth or what is referred to as permanent income (Modigliani). According to Freidman (1978) the fraction should be near one. Robert Lucas (1976), however, criticizes the model due to a short lag that appears between past and present incomes.
Regressions with consumption as their dependent variables are calculated in the framework of permanent income. In this model, its assumed that most models contain exogenous variables because they have a major and significant influence on any income. However, exogeneity is not testable and there is not certain whatsoever. This brings about a state of uneasy understanding.
The hypothesis of permanent income with only a period lagged behind should be having a non-zero co-efficient. In such a test assumptions of any kind of exogeneity should not be included. Procedures for testing any theoretical implications involve implications such as lagged consumption for more than a period does not have power to predict current consumption.
It, therefore, implies that consumption is not related in any way to an economic variable observed in previous periods. Lagged income though good in predicting current income is unreliable in detecting the behavior of consumers of thinking of the future that makes up the hypothesis of permanent income.
However, if lagged consumption is included, no further explanations should be included in the lagged income. This is because if included, it will be entered as a double entry which will not give accurate and precise results.
Finally, the final test is carried out for the predicted values of prices of stocks from corporations. Consumption changes can be predicted by lagged stock prices for a quarter and has a significant and measurable value.
The hypothesis could be modified to reflect a small lag between changes of permanent income to consumer consumption. Since stock prices move in the same manner as stock prices, the modification to the hypothesis of the random walk is supported.
According to the proper life cycle of hypothesis of permanent income, its very reliable and useful in future predictions in macroeconomic field and policy implementation analysis.
With every deviation of consumption from it’s normal abrupt and permanent trend, the best prediction of future consumption adjusted to any trend making future changes in income unreliable to work with. This is because information used to make them is already included in current consumption.
In models used for forecasting, consumption as a variable is regarded as exogenous. Only abrupt and unavoidable changes in policy affect consumption. These changes are expected to show permanent effects and affect only permanent income as viewed in the permanent income life cycle.
Effects incurred on transit of income, cannot have transitory effects that can be experienced in consumption. However, policies that affect income are seen to have some effects on consumption to some extent.
A conditional expectation is used for the scholastic tests to differentiate theory of permanent income from other theories. It is done by making an estimate of an conditional expectation denoted E. The tests are usually F-tests when one excludes a group of variables in a regression. Regression is the best mathematical technique used for estimating expectations that are conditional.
In the permanent income hypothesis, a major contradiction occurs. It suggests that consumers in consumption on transit cannot have smooth consumption because of constraints of liquidity of money. This makes consumption become very delicate or basically sensitive to previous or current income to be able to fit this hypothesis. It seconds the hypothesis out of the rigidity it has acquired.
Another contradiction shown clearly is that permanent income becomes well distributed in a lag of the actual income of the past. A consumption function relating both ideas fits very well. However, estimating them involves determining income together with consumption.
According to Mishkin (1971), the hypothesis is confirmed when consumption together with disposable income follow the exact stochastic process where permanent income and observed income are the same thing. The life cycle of hypothesis of permanent income is automatically rejected when permanent income and observed income appear to be different.
Flavin (1977) reasoned that a regression that involved consumption that was lagged, provided the distributed lag followed by a stochastic process; it could fit the permanent theory hypothesis. If consumers, however, used estimated using lags that were not optimal in estimation of permanent income, then permanent income hypothesis would be false.
The non-optimal lag makes consumption to have added prediction power for current consumption far beyond lagged consumption rejecting the hypothesis. Consumers that use distributed lag at its optimum of present and past incomes to make decisions in consumption do not reject the hypothesis.
In conclusion, simple tests of predicting variables are more essential than lagged consumption to detect deviations from the hypothesis of permanent income.
References
Lucas. E. Robert (1978). Rational expectations and econometric practice, Volume 1. New York: Monthly Review Press