In this paper, we seek to analyze how quantity demanded for alcoholic beverages changes when income changes. Specifically we want to determine whether the demand for alcohol is sensitive to economy wide expansions and contractions as measured by Japan’s Gross Domestic Product. We also want to determine how in a consumer’s income affects their demand for alcohol. There is a survey of some of the major non-economic factors that affect alcohol demand and a case study of Kirin Co., a major beer producer in Japan.
A normal good is one whose quantity demanded increases as income increases (Investopedia, 2012). In contrast, an inferior good is one whose quantity demanded decreases as income increases in the economy. This occurs because consumers substitute away from the good as their income increases. Alcohol is a substance that can be used to cope with financial strain, and it has been “shown that Unemployment is positively related to alcohol use, consumption rates and binge drinking” (Peirce, et. all 1994). Depending on the relative strength of these effects, alcohol’s income inelasticity could be positive, negative, or close to zero, making it a normal good, an inferior good, or immune to income changes respectively. We are investigating whether alcohol falls into one of these categories or has more complex demand elasticity characteristics. Donald Freeman used the term “vice products” to describe ““defensive” stocks, relatively immune to the business cycle.” The quantity demanded for these goods, in theory, rises with income, but does not fall when incomes contract.
Within Japan, “drinking patterns and problems can be broken down… by gender, age, ethnicity, and socioeconomic status.” Also, different alcoholic beverages are associated with different drinking patterns that could affect the overall demand for alcohol: “the preference for one beverage over another has been show to correlate with significant differences in lifestyle, personality, and consumption pattern.”
Using the GDP figures for Japan from the World Bank in constant US dollars, alcoholic beverages show some tentative evidence of being resistant to changes in GDP. Despite the fact that GDP growth was negative in 1972 and 1974, consumer expenditures on alcohol rose in both of those years. In 1993, there was a contraction of the Japanese economy that corresponded with a fall in the consumption of alcohol. In 1998, growth again was negative, yet alcohol consumption was steady during this period. After 1998, there is a steady downward trend in the consumption of alcohol that seems unrelated to the activity in the movements of GDP until 2006 when the economy and the alcohol demand both reach a trough and begin to rebound (See figure-3 in the appendix). There are 3 downturns that do not coincide with a corresponding dip in the alcohol demand, and two that do. There is a 79.5% correlation between household alcohol consumption and GDP. Comparing the percent change in GDP and the percent change in Alcohol expenditure, reveals that when GDP growth is high, the curves are close, but as the % change in GDP growth approaches zero, there is less coordination between the curves, as seen in Figure-4 around the year 1990.
This evidence allows us to rule out the possibility of Alcohol being an inferior good. In general, if the economy expands, more alcohol is purchased.
The relationship between household income and alcohol expenditure is much closer (See figure-5 in appendix). There is a 96.34% statistical correlation between income and alcohol consumption over the course of the observational period. Comparing the percent change in income and the percent change in Alcohol expenditure, reveals that when income is high, the curves are close, but as the change in % GDP growth approaches zero, there is less coordination between the curves, similar to the % change in GDP (see Figure-6). There is not enough evidence to conclude that alcohol consumption is a function of income. The relationship between household income and alcohol expenditures, does not show evidence that alcohol expenditure is resistant to downward trends in expenditure or income. There are few distinct troughs in the income graph over time, but spending on alcohol falls faster than the gradual fall of total expenditures from 1997 to 2006. There are variations in alcohol expenditure that are not explained by income. Alcohol consumption, overall, follows the same trend as household income.
After looking at several representations for income including GDP, which represents the whole economy of Japan, and average household level income, it is clear that Alcoholic beverages are not inferior goods: as income increases, the percentage of income spent on alcohol increases. The aggregate data does not provide evidence that they are vice goods, which resist falls to quantity demanded when household incomes fall. There are periods when income falls in which alcohol expenditures fall faster than income, the opposite of what you would expect if alcohol were a vice good.
None of the individual beverages has a strong relationship with income. Beer is the most responsive to changes in income. There is no evidence that any individual alcoholic beverage’s expenditure is resistant to decreases in wages and income. Figure-8, which graphs income graphed against each individual drink shows similar trends, shows mostly noise. There is no closely coordinated response from any drink to changes in income.
If we apply this same analysis to a specific company, Kirin Co, we find evidence that alcoholic beverages are indeed vice goods. Over the limited period of data available from 1998 to 2007, Kirin’s alcoholic beverage sales do not respond to falls in either GDP or Income. GDP falls in ’98 and from 2000 to 2002, yet Kirin’s alcoholic beverage sales are flat in both time periods (see figure-9). Income in Japanese Households falls from 1998 to 2003 and again in 2005, yet Kirin’s sales are either flat or increasing over these periods of contraction (see figure-10). When either GDP or Income increases, Kirin’s sales rise. This is the exact behavior we would expect from a vice good.
The aggregate data tells a different story than the firm level data for the Alcoholic beverage demand. The aggregate data reveals that every reported alcoholic beverage is positively correlated with Income. Alcoholic beverages as a whole also show positive relationships with Japan’s GDP, household expenditures, and Household Incomes. In general, the aggregate data showed that alcohol expenditures show more variance than the proxies for income, so it was difficult to associate patterns between the different statistics. There is not enough definitive evidence to distinguish whether alcoholic beverages, as a whole or individually, are normal goods or vice goods. The aggregate data provides convincing evidence that Alcoholic beverages in Japan are not inferior goods.
Kirin’s Data supports the theory that alcohol is a vice good whose quantity demanded rises when incomes increase, but do not fall when incomes decrease. Although there was clear evidence of the sort of behavior we would expect to see from a vice good, we should be cautions when applying these results. Kirin has released just under a decades of data, and their data is not broken down by alcoholic beverage type. Further, Kirin is only one company. Applying conclusions based on such a limited set of data is not advisable.
In an ideal experiment, we would have access to Kirin’s data over a longer period of time, and more than one company to use as a case study. Also the figures would be separated into the different beverage categories including beer, Happoshu, ‘Third Beer’, and etc.
. The data aggregate data available cannot adequately account for population ageing; changes in tax rates that change the effective cost of a beverage; or cultural changes within Japan’s society over time that would affect consumer preferences. Although the World Bank, and Japan’s Ministry of Internal Affairs and Communications release substantial statistics on the Japanese economy, Japanese consumer prices, and spending there are so many factors related to Alcohol consumption that it is difficult to account for them all. To get a true picture of how the demand for alcoholic beverages in Japan reacts to income changes, we would need data that allows us to control for all of these variables.