Impact of Google search trends on market price trends of commodities
Introduction
This study attempts to utilize the Google search trends data to investigate the potential impact that search trends might have on the commodities market value trends or vice versa.
In this study, I have used six commodities: traditional and some not so conventional. The traditional commodities include crude oil, gold, and Silver. The non-traditional ones are the top two digital currencies by market cap, namely Bitcoin (BTC) and Ethereum (ETH). One choice was given to the stock value of the GameStop Corporation. Some other public companies’ stock value was also investigated, but the findings were deemed insignificant and uninteresting and therefore are omitted from this report. The decision to include non-traditional commodities was made to explore their potential unique nature regarding the influence of the public’s behavior on their price portfolio.
Methods
The Google search trends data was collected and exported onto an Excel sheet (CSV) for 12 months from mid-March 2021, going backward in time. The market value of each commodity was also downloaded onto an Excel sheet for the same period. The resolution of the data, however, is different for each data set. In Google trends data, we have a weekly value vs. in the market value dataset, we have the daily values present. For this purpose, the Google trends data was adjusted to reflect both the mid-week and end of the week search trends and shifted on the graphs as seen in the results section of the report. Both scenarios (mid-week and end of the week adjustments) had all but the same effective results. Therefore to make the data visually more accessible, the mid-week adjusted version was plotted for this report.
Further, in the Bitcoin and Ethereum data, I have acquired finer resolution datasets around the week of interest that had the most activity and was the most volatile. The new datasets have the data for every day within 90 days around the peak point of interest for the Google trends dataset and the market price dataset.
All search data, except BTC and ETH’s fine resolutions, have been on the US and worldwide. For the fine resolution search data, only the US data was used. The worldwide and the US data for Bitcoin and Ethereum were comparable.
The market price data was extracted from Yahoo Finance.
Results
Here let us provide the results and graphs of our data collection on each commodity one by one.
In Silver’s case, as seen in the chart below, we observe a point of near-perfect overlap between price and search trends data
In Gold data, we can also see instances of near-perfect overlap between price and searches, and in some cases, we see instances that market price changes precede the search trends.
Gold is the only commodity that, in its data, we can find instances of the extreme difference between the search trend in the US and worldwide. This phenomenon remains unexplained in my study.
Oil data shows an instance of search trends lagging behind a dramatic price change at the beginning of the COVID-19 pandemic. We will attempt to explain this phenomenon in the discussion section of this report.
Bitcoin (BTC) shows a few instances that a search trend is preceding the price change. The uniqueness of this phenomenon in cryptocurrencies motivated me to provide a better resolution of 90 days dataset, with one data point for each day for both the search trends and the price trend. The graph for 12 months and the fine resolution can be found below.
Ethereum (ETH), similar to bitcoin, has instances of price trends following the search trends. Same as with Bitcoin, I also investigated a fine resolution for the Ethereum price and search trend. The fine resolution version of the graph spans 90 days and provides one data point for each day for both prices and search trends.
GameStop was the final commodity that I investigated. This commodity was also the only stock commodity of note among those I investigated (namely Tesla, Microsoft, Google, Amazon, etc.). In the discussion section, we attempt a deeper dive into the story behind the graph’s peak moment.
Discussion
In our observation, one pattern appears that more traditional commodities such as gold, crude oil, and Silver do not follow the search trend and population interest. This phenomenon might not necessarily be valid at a better resolution or smaller population; however, given that my search data source is Google Trends, this can only make observations into larger population trends. Therefore, traditional commodities’ market value has a mind of its own, driven by unknown forces to this study and not by mass’s behavior.
Although the cryptocurrencies of choice in this study do follow price trends preceding the search trends at times, the peaks seem to show the opposite behavior, meaning the peak price changes follow a massive surge in searches on Google. Of course, these searches do not hold the context in which the search was done. Ultimately, the digital currencies’ prices seem to initially get their influence from narrower sources, leading to a considerable surge in searches that drive the more significant price surge waves at the end.
GameStop stock is an excellent illustration of gradual few-day price changes leading to an interest in searches, which in turn drastically increases the stock price. The GameStop also had help by being in the news at the time. Some of the investors shorted it. However, as a way of protest, a group of independent individuals started making many small investments into GameStop, creating the early instances of gradual price change. This story makes its way into the news and sudden increase in searches. Following this increase, the price, of course, skyrockets.
Conclusion
The commodity market’s edge cases seem to be influenced by the population (search) trends. Specifically in their own edge cases. In this study, I have defined edge cases as non-established commodities (officially by the major organizations and/or governments) or up and coming commodities such as Bitcoin and Ethereum or commodities in turmoil and/or narrative driven, like GameStop.
It is worth noting that this is far from conclusive with the given datasets, and further in-depth studies are required to determine the validity of this claim. More likely, more data sources with much more sensitive nuances are needed to understand a complex organism such as the commodity market.
This study was done as a first step to determine whether anything of note can be found in the search vs. price trends. And a simple conclusion to that question, indicated in this study based on the results, is yes, there seem to be some relations of interest between market price trends and search trends.
Although I have already addressed Google Trends data limitations, I would like to put a more delicate note on it here. Google Trends data is a relative measure and does not provide a pure number of searchers. Although, in this case, we are mainly interested in the relative changes, it limits us in comparing the commodities with each other given the presented results.
As mentioned before, Google Trends entirely lacks the context in which a search was done. This will hinder our ability to consider, discuss and conclude with a great deal of certainty. We don’t know whether the individuals searching for the keywords are interested in making purchases, following the news, merely checking the prices, or something else.
Nevertheless, Google Trends’ data provided us with the ability to conclude there are relations of potential interest between some commodities price and their search and media trends within certain contexts.
References:
1. Google trends. (n.d.). Retrieved March 15, 2021, from https://trends.google.com/trends/?geo=US
Keywords: ETH, Ethereum, gold, bitcoin, oil, crude oil, Silver, GameStop.
2. Pisani, P. (2021, January 31). For GameStop day traders, the moment they’ve dreamed about. Retrieved March 15, 2021, from https://www.ctvnews.ca/business/for-gamestop-day-traders-the-moment-they-ve-dreamed-about-1.5289040
3. Yahoo finance — stock Market Live, Quotes, business & finance news. (n.d.). Retrieved March 15, 2021, from https://finance.yahoo.com/
Historical price of gold, Silver, ETH, BTC, oil, and GameStop archive used.