In May 2012, the ISET Policy Institute piloted a new monthly survey to measure consumer sentiment and expectations. The first pilot included about 60 randomly selected individuals who were asked about their well-being, saving and spending plans, etc. While based on a relatively small sample, the pilot points to a very large gap in consumer confidence among Tbilisi and the rest of the Georgia. Tbilisi households perceive themselves to be marginally better of at present, yet they report a much larger drop in wellbeing during the last 12 months and are less optimistic about the future. The overall Consumer Confidence Index (running from -100 to 100) takes a value of -18 for Tbilisi and -7 for rest of Georgia.
In planning the first full-scale survey we had to decide when to conduct the interviews since survey results may be very sensitive to timing. For instance, people are likely to have a very different perception of their financial status on June 30 and July 1, that is before and after receiving their monthly pay. Likewise, people may provide very different responses if called before and after they had their early morning coffee. Thus, we settled on conducting interviews during normal business hours in the middle of each month.
More importantly, consumer sentiment will certainly change with the time of the year. For instance, many US consumers may feel ecstatic in the midst of the Christmas shopping spree paid for by year-end bonuses. With no year-end bonuses in the cards, for most rural households in Kakheti the best time is September and October when farmers reap the fruit of a long agricultural season, harvest their grapes and work traditional wine presses. By December, wine festivities will be long forgotten. Farmers will be living off their savings and eagerly waiting for the new agricultural season, the return of tourists and lower food prices.
For the vast majority of Georgian households, consumer sentiment may be strongly, albeit negatively, correlated with the seasonal fluctuations in food prices. Higher food prices affect the real income and consumption of the poorer households for which food accounts for a very large share of monthly expenditures. Seasonal food price hikes are stronger in the outlying regions of Georgian, making provincial consumers (and voters) particularly susceptible to the low season doom and gloom.
Which hints at a possible answer to the question we posed in the title of this blog post: if held in early fall (say, in October), when prices are low and consumer sentiment is at its peak, political elections – the ultimate “consumer confidence survey” – are more likely to favor incumbent governments.
Comments
This is a very good analysis and shows the power of simple, clear thinking that is not influenced by a prior theoretical baggage. As a theorist I am tempted to think of the economic theory behind this all.
If I have understood the diagram correctly then people in Tbilisi have suffered more (-24% versus -10%) and yet despite being better off than the rest of Georgia (in per capita terms) are less confident (-18% versus -7%). This indicates strongly the possibility of reference dependence and, hence, indirect support for prospect theory. If people behaved according to expected utility then the absolute final levels of wealth, which are higher in Tbilisi (and not the relative change in the last 12 months), should make people in Tbilisi feel more confident- but this would not explain the data.
A second point relates to the timing of elections. In standard political economy, voters are rational and forward looking. If this were true then they would not be "fooled" by a call for elections in the fall and the incumbency advantage would disappear. Using standard game theory, the politicians would see through this as well and so would not call for elections in the fall. But we do know that this is not in line with the evidence (or the sensible conclusions of this article).
I know little about the Georgian economy so there could be other valid explanations of the data too. But prima facie the data would seem to provide support for behavioral economics.
This is very interesting argument and it has very strong intuition. Dear professor Sanjit, I am interested in what is possible explanations by prospect theory for the less confidence of people in Tbilisi? I think that the higher wealth increases responsibilities and disturbances about the future.In the big cities, people take more credits and borrow more in compared to people who are living in regions.Moreover, the consumption of city's people are higher and their income is more elastic to any fluctuations in economy. In contrast, in the region, the substantial part of people's consumption is from their own garden and their income and consumption less depends on markets. Moreover, in the regions, the most part of people have less wealth and take less credits in compared to the people of big cities. Can it be possible reason for higher confidence in the regions?
Dear Eric, can the weather factor play role in the schedule of elections? because, in winter and summer, especially in rural areas, it is difficult for people to go election places. Moreover, the Fall is the last season of year and the most part of government works and projects are completed during this time and it will be more convenient for the new government to begin in the beggining of year.can it be reason?
Vusal, you ask some very good questions. The results of the survey are coming from a very small sample, thus I would not want to rush to any conclusions, but you are right that life in a city is more risky: people may face many more opportunities to become rich but also to lose their jobs (and, still worse, go bankrupt). In the countryside, there are lesser opportunities but greater stability. Given the generally high level of uncertainty currently facing Georgia (e.g. related to the upcoming elections), the urban population may have a stronger sense of the downside risks from which the rural population is (for now) insulated. Whatever happens to tourism, trade flows and FDI to Georgia, the small farmers would still have their cow, chicken, potatoes and wine (or, rather, that's how they feel).
As for the timing of elections, I would not rule out any other explanation: weather, parliamentary calendar (summer vacations), etc. I believe that the seasonal pattern of consumer sentiment is an interesting explanation, too. I thought about this while looking at consumer confidence data and the Khachapuri index data for Tbilisi and other locations.
Sanjit, I am glad you like the piece. It never occurred to me that it could provide evidence for behavioral economics theory, but am happy it does (if it does). The difference-in-risk-perceptions explanation proposed by Vusal also sounds quite plausible.
As for the timing of elections, I certainly know from personal experience that even a cup of coffee changes my perception of reality (and political preferences, and consumer sentiment). I would vote against any government imposing on me an early morning ballot
Vusal, my suggestion is based on the idea of reference dependence. You could ask any of the students who took my behavior economic course at ISET in the last 2 years (e.g., Illaha Samadova) for my lecture notes and pursue this further. You are also right about controlling for other factors involving risk- a follow up study might find it worthwhile controlling for these.