In this hypothesis we check what factors may affect the discretionary spending by households, or rather, we
check which factors may be stronger predictors of discretionary spending. We check for four factors: literacy,
formal employment, current status, and government support for social infrastructure. Since our variables are
discretized into multiple ordinal levels, we choose to examine the mutual information between the pairs to form
an assessment of the strength of relationship between them. The same method is used to calculate mutual
information between each of the four factors of interest in this hypothesis (literacy, formal employment,
current status, and govern-ment support for social infrastructure), and change in each of the four discretionary
variables (asset ownership, bath-room facility, fuel for cooking, and condition of household). The figure to the
right shows the mutual information calculated on these tables. We can see that literacy, formal employment, and
the current status are all more predictive of change in discretionary variables as compared to government
support on the main source of lighting and main source of water.
This is not altogether surprising because literacy and formal employment go hand in hand with each other, and
the wages in formal employment are substantially higher than wages in the unorganized sector, which puts more
disposable income in the hands of people that can be used for discretionary expenditure. The 2018 India Wage
report by ILO reveals that the average daily wages of workers in formal employment (organized sector) are ₹ 513,
while wages in the unorganized sector are ₹ 166. Similarly, wages for men with the highest level of education as
graduates are ₹ 735, while for casual urban workers with the same level of education the wages are ₹ 219.