EconoMonitor

The Kapali Carsi

Google Trends and Fun Facts About The Turkish Economy- 1

I have not had guest posts for a while (well, because I wasn’t blogging for a while either), so here’s finally another guest post from Ali G (not the rapper– sorry for the bad joke.

BTW, as I mentioned before, I am not really editing these guest posts, so apologies for typos, grammatical errors and the like. And if you’d like to guest-blog here, or at the newly-inaugurated Hurriyet Daily News blogs, here’s the invitation, along with the “application procedure”. Anyway, without further chit-chat, on to the guest post:

As anyone who has done some sort of empirical work knows, the quality and quantity of data is extremely important. No matter how statistically fancy your method is, if your data can’t live up to it, your results will either be weak- or misleading altogether. It is like when a regular person tries on a piece of clothing they saw on a runway model. Even if they can get their hands on the item, it does not fit, so one has to try to get into the right shape first. The nerdy metaphor aside, I honestly believe that collecting and cleaning the data is probably 90% of the work.

One of the things I like about doing empirical stuff in finance compared to economics is that by nature you deal with these problems to a lesser extent. Moreover, most financial decisions (e.g. trading decisions, borrowing etc.) are directly observable and data is relatively well-kept and easy to reach even in emerging countries. As a result, generally it is easier to obtain more relevant and robust results. A recent example is the Investor Sentiment Index for Turkey that is being developed jointly by Özyeğin University’s Center for Computational Finance and the Central Registry Agency of Turkey. The index is constructed using asset holdings of investors (provided by the Central Registry) and so far it looks like a pretty promising high frequency indicator for investor sentiment in Turkey.

Measuring sentiment, or how people feel or think about a certain subject is not very straightforward. In the above example, sentiment is directly measured (at a relatively high frequency) by actual choices people make. However, direct measurement is generally not possible and the standard practice is to conduct surveys. These surveys are generally designed pretty well from a statistical point of view, but are not frequent enough, may be lagging or biased due to the fact that what people say and do actually differ. The last point is apparent in the sub-indices of the CBRT’s Consumer Sentiment Index.

In the recent years however, technological improvements and ever growing use of the Internet provide other measurement alternatives. There are numerous examples especially from marketing, but the most obvious ones are probably choice data constantly being gathered by behemoths such as Google and Facebook to provide better ad service among other things. Considering how habits change and the speed with which the use of the Web increases, pretty much every keystroke provides valuable and accurate data. And some of this is even made public.

Google Trends and Google Insights for Search basically provide search trend data on every search item and classify them according to different categories such as regions. I will talk about this in detail in the next post, but for the interested reader I am posting the link to a paper by Google’s Chief Economist Hal Varian (anyone recognize the name?) below that explains how Google queries can help predict economic activity (1).

In the next post, I will be talking about how these Google services are useful, who has been using them and finally I’ll show you some exciting results from Turkey.

(1) http://static.googleusercontent.com/external_content/untrusted_dlcp/www.google.com/en//googleblogs/pdfs/google_predicting_the_present.pdf

Ali Gökhan is the acting economist at a Turkish conglomerate. The views expressed here are his personal views only and do not represent the views of his company.

I don’t have much to add, except that the Investor Sentiment Index Ali G is mentioning is quite interesting. I had not heard about it before (a couple of phone calls ascertained my economist friends had, but not seen it), and it is not in Turkey Data Monitor, but I know the guy who is doing it, so I may ask him about it and then do a blog post or even a column…

One Response to “Google Trends and Fun Facts About The Turkish Economy- 1”

playdateSeptember 7th, 2013 at 12:03 pm

Thanks for the post. The most affected group in the fall in the value of euro is the Germany. Euro crisis has very badly affected the Germans. Other European countries are behind Germany only. Good going blogger, keep posting.