Using Google Trends To Predict Stocks

Summary

• Google Trends allows analysts to see how often certain terms are searched.
• By analyzing bullish and bearish term search volume, we can construct an investor sentiment index.
• This "Google Trends" investor sentiment index appears to predict consumer sector outperformance.
• Google Trends data is an exciting frontier in investment research we will continue to develop in.

For those of you who have not heard of Google Trends, it is a tool produced by Alphabet/Google (GOOG) (GOOGL) that allows users to see search volume and geographic search concentration for a wide range of keywords. It is most often used to see whether Kim Kardashian or Taylor Swift is trending more, but we believe it has great use for active investors and research analysts.

We hypothesize that retail investor sentiment is correlated to Google search volume. When more users search for keywords such as "stocks to buy", risk-on stocks should outperform risk-off stocks. When users search terms such as "how to short sell", it is a signal more retail investors are becoming worried and that risk-off stocks should outperform.

Measuring Bullish vs. Bearish Sentiment

When a new investor enters the market, it often begins with a Google search of "what stocks to buy" or "top stocks". Often such investors are responding to the market itself; their friend touts about how much money they recently made in the market and said new investor takes to Google. Doing so will lead to a few sources on "how to invest" and a few stock tips; and, a week or two later they will buy stocks.

Here is the monthly data we collected on these keywords since 2004:

Source: Google Trends

As you can see, there are large spikes both during sharp market rallies as in January of this year and during large sell-offs as in 2008. To create a more accurate "google search sentiment" index, we will gather the same data for the bearish side of Google search volume.

When investors become worried, they will often search "sell stocks", "how to short sell" and "how to sell stocks". Here is the same as above for these "bearish" search requests:

Source: Google Trends

Here is both the bullish and bearish search volume since 2004:

Data source: Google Trends

As you can see, the two data closely mirror each other. There is a large spike during the worst sell-off months in 2008, and again this January as the S&P 500 climbed 8% in a few weeks.

It is clear that Google search volume reacts to the markets, the question remains as to whether we can use Google Trends to predict the markets.

Sentiment Speaks

To create a sentiment index, we will take the difference between the "bull" and "bear" searches and give the data a 6-month moving average to filter out random noise.

Data source: Google Trends, Google Finance

As you can see, our index was not a strong indicator until after the recession as Google search volume was considerably lower then. However, over the stretch in this bull market, the index has become much more useful. The index does not completely trail market performance, rather it reflects retail investor sentiment.

The index data above corresponds to the newly created "Investor Movement Index" by TD Ameritrade that tracks new brokerage accounts and account positioning:

Source: TD Ameritrade

As you can see, there is a corresponding spike toward the end of 2016 that grew in strength before falling back to its long run average following the February correction. This tells us that our "Google Sentiment Index" is measuring nearly the same information as the IMX index, just in a vastly different way.

Google Trends Index Investment Strategy

In order to show how the data can be used, we wrote a simple algorithm. If our sentiment index is negative, we invest in consumer staples (XLP); if positive, we invest in consumer discretionary (XLY).

The concept is that when negative sentiment is growing, it invests in "risk off" stocks like consumer staples in expectation of risk-avoidance, and when bullish sentiment is growing, it invests in "risk on" stocks like consumer discretionary stocks to take advantage of momentum.

Here is the resulting growth of $10,000 under this strategy. (Note, dividends are excluded).

Data source: Google Trends, Google Finance

As shown above, this strategy beats the market with smaller drawdowns. In fact, since 2004, it has had an average annual performance of 10% while the market had an average performance of 7%. Even more, it did so with the same annualized standard deviation of 11%. Thus, the strategy had an "excellent" Sharpe ratio of 0.92 while the S&P 500 was only 0.59.

Bottom Line

Of course, no strategy is fool-proof and Google Trends data is just in its infancy of becoming useful to the investment industry. Measuring investor sentiment this way can become problematic during "market events" that cause people to Google about the stock market without the intent of buying or selling stocks.

As you can see in "bullish vs. bearish search volume", there is also a lot of random noise in the data that can easily cause a trading strategy to misfire. To deal with that, we're left with using a moving average of the data. While this approach is beneficial for analysis, it causes a large 3-month lag in the data and confines any strategy to a long-term investor. Ideally, more research will be completed on creating shorter-term trading strategies with Google Trends.

Analyzing search volume is a useful tool for measuring investor sentiment. By creating a simple "Bull Bear" spread index, we gain insight into investor risk-appetite which can then be used for "risk on" versus "risk off" allocation. We will be looking much closer at "Google Trends" investment strategies in the future as we believe it's the next frontier for investment research.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.