Can Google Trends search queries contribute to risk diversification?

Abstract

Portfolio diversification and active risk management are essential parts of financial analysis which became even more crucial (and questioned) during and after the years of the Global Financial Crisis. We propose a novel approach to portfolio diversification using the information of searched items on Google Trends. The diversification is based on an idea that popularity of a stock measured by search queries is correlated with the stock riskiness. We penalize the popular stocks by assigning them lower portfolio weights and we bring forward the less popular, or peripheral, stocks to decrease the total riskiness of the portfolio. Our results indicate that such strategy dominates both the benchmark index and the uniformly weighted portfolio both in-sample and out-of-sample.

Figure 1. Portfolio performance based on ticker symbols.

Standard deviation (left) and Sharpe ratio (right) are shown for in-sample (full symbols) and out-of-sample (empty symbols) performances of constructed portfolios. The discrimination parameter α ranges between −2 and 2 with a step of 0.1. The middle point (α = 0) represents the uniformly weighted portfolio. The red dashed line represents the benchmark DJI index. Minimum variance portfolio is found for α = 0.1 for both in- and out-of-sample approaches. In general, the resulting standard deviations for varying α practically overlap for both approaches. For the Sharpe ratio measure, the performance differs. For the in-sample, the Sharpe ratio is maximized for α = 0.6, and for the out-of-sample, it is maximized for α = 2.

Figure 2. Portfolio performance based on ticker symbols combined with the word “stock”.

Standard deviation (left) and Sharpe ratio (right) are shown for in-sample (full symbols) and out-of-sample (empty symbols) performances of constructed portfolios. The discrimination parameter α ranges between −2 and 2 with a step of 0.1. The middle point (α = 0) represents the uniformly weighted portfolio. The red dashed line represents the benchmark DJI index. Minimum variance portfolio is found for α = 1 and α = 1.3 for the in- and out-of-sample approaches, respectively. Again, the resulting standard deviations for varying α are very close for both approaches. For the Sharpe ratio measure, the performance again differs. The Sharpe ratio is maximized for α = 0.2 and α = 0.6 for the in- and out-of-sample, respectively.

Figure 3. Evolution of portfolio value based on Google Trends diversification.

Red line represents the evolution of the DJI index, black line shows the performance of the out-of-sample diversification and the grey line illustrates the development of the in-sample diversification approach. Portfolio value is shown on the y-axis. On the left panel, the ticker symbol searches are utilized and on the right panel, the combination of the word “stock” and the ticker symbol is used. The evolution is shown for the discrimination parameters α which maximize the Sharpe ratio for each scenario. For the practical purposes, a comparison of the black and red lines is essential as it shows how much better off we would be if we applied the Google Trends based strategy. For the first case, the search queries based strategy brings the total profit which is more than 4 times higher compared to investing to the DJI index. For the second case, the cumulative profit is more than 0.4 times higher for the search based strategy.