Search Volume Index Research Articles

Measuring Investor Attention Using Google Search

Although investor attention is fundamental to the efficient functioning of capital markets, it is also an elusive construct that researchers struggle to measure. In recent years, the search volume index (SVI) of ticker searches on Google has become a ubiquitous measure of investor attention, but the amount and effects of measurement error in ticker SVI are unknown. We investigate measurement error in ticker SVI using a data set of 2.7 billion website visits following Standard and Poor’s 500 firms’ ticker searches. We find that 69% of searches are unrelated to investing, that this measurement error is highly correlated with firm characteristics, and that this measurement error can easily generate false-positive or false-negative results in common settings. We go on to show that a modified version of SVI using both a firm’s ticker and the word “stock” (e.g., searches for “CAT stock,” which we label “ticker-stock SVI”) not only better captures the search terms that investors typically use, but also has considerably less measurement error that is largely uncorrelated with observable firm characteristics. Ticker-stock SVI produces better specified tests, and although researchers must still carefully consider the effects of measurement error, we recommend that ticker-stock SVI is used in place of ticker SVI in most settings. We provide a data set of ticker-stock SVI to facilitate future work. This paper was accepted by Suraj Srinivisan, accounting. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.02174 . TS-SVI for the Russell 3000 is available for download at: github.com/robinlitjens/GoogleTickerStock-SVI and will be updated annually.

Surge in Eye Drop Safety Concerns and Interest in Alternative Dry Eye Disease Therapy Following Eye Drop Recalls: A Cross-Sectional Google Trends Analysis.

Beginning in January 2023, numerous eye drops were recalled following reports of contamination. To investigate hesitancy in topical therapies for dry eye disease (DED), we analyzed trends in Internet search traffic for alternative DED therapies and eye drop safety concerns following initial recalls. Google Trends (GT) was used to measure the popularity of specific search terms during a given time period. The main GT output metric for gauging interest in a search term is the Search Volume Index (SVI). We collected worldwide SVI data for terms "dry eye vitamin," "are eye drops safe," "stop eye drops," "lipiflow," "ilux treatment," "dry eye mask," "dry eye warm compress," "omega-3 for dry eyes," "dry eye plugs," "dry eye humidifier," "itear," and "IPL dry eye" from April 24, 2022, to October 29, 2023. Statistical analysis was performed with Microsoft Excel v.16.77.1, and comparisons between groups were performed using paired t tests. A significant increase in SVI was observed for most keywords. When comparing the 40-week period before and after the initial recall, 7/10 alternative DED therapies demonstrated a statistically significant surge in Internet search traffic after recall (P < 0.05). Furthermore, 2/2 keywords pertaining to eye drop safety concerns demonstrated an increase in SVI (P < 0.0001). We identified a significant surge in Internet search activity for eye drop safety and alternative DED therapies following eye drop recalls. Public health agencies and physicians may find this information valuable while attempting to address safety concerns and navigating noncompliance of routine eye drop therapy.

A Collection of Wisdom in Predicting Sector Returns: The Use of Google Search Volume Index

This study investigates whether the aggregate investor information demand for all stocks in a sector demonstrated in the Google search volume index (SVI) can predict the sector’s performance. The evidence shows that a sector’s abnormal SVI can predict the sector’s performance next month, even after controlling for the sector’s contemporaneous standardized unexpected earnings and lagged returns on both the market and the sector. Also found is a partial reversal in the sector’s long-run performance that is not completely consistent with the price pressure hypothesis. This indicates that some fundamental information about a sector can be captured by the sector’s abnormal SVI on a timely basis.

Are stock markets efficient with respect to the Google search volume index? A robustness check of the literature studies

The Efficient Market Hypothesis (EMH) is still a debated subject in the financial area. Particularly, no conclusions are drawn to date in link with the Google Search Volume Index (GSVI). To conclude on this question, our paper takes up the work of Škrinjarić (2019) by proposing robustness tests, various econometric improvements and the inclusion of additional explanatory variables. On a database of ten emerging European indices studied by Škrinjarić (2019), a dynamic panel model was applied. Unlike Škrinjarić (2019) who modeled the time-series separately and thus neglected any possible dependence or homogeneity between countries, our study operates within the framework of panel data. Drawing from a robust estimation approach, our findings indicate that the GSVI has no impact on market returns. In essence, this suggests that internet search queries fail to provide avenues for investors to seize arbitrage opportunities. Such findings support the EMH in the studied markets and underline the exposure of prior studies to robustness challenges.

Mental Health Applications of Generative AI and Large Language Modeling in the United States.

(1) Background: Artificial intelligence (AI) has flourished in recent years. More specifically, generative AI has had broad applications in many disciplines. While mental illness is on the rise, AI has proven valuable in aiding the diagnosis and treatment of mental disorders. However, there is little to no research about precisely how much interest there is in AI technology. (2) Methods: We performed a Google Trends search for "AI and mental health" and compared relative search volume (RSV) indices of "AI", "AI and Depression", and "AI and anxiety". This time series study employed Box-Jenkins time series modeling to forecast long-term interest through the end of 2024. (3) Results: Within the United States, AI interest steadily increased throughout 2023, with some anomalies due to media reporting. Through predictive models, we found that this trend is predicted to increase 114% through the end of the year 2024, with public interest in AI applications being on the rise. (4) Conclusions: According to our study, we found that the awareness of AI has drastically increased throughout 2023, especially in mental health. This demonstrates increasing public awareness of mental health and AI, making advocacy and education about AI technology of paramount importance.

Effects of COVID-19 on Investor Sentiment: Evidence from Commodity Futures Using Google Search Volume Index

Effects of COVID-19 on Investor Sentiment: Evidence from Commodity Futures Using Google Search Volume Index

Physical climate risk attention and dynamic volatility connectedness among new energy stocks

Based on a sample of 41 Chinese new energy stocks during the period from 2016 to 2022, this paper uses the improved Diebold-Yilmaz method to analyze the volatility connectedness among new energy stocks. We study the essential characteristics of the volatility connectedness network, including major risk transmitters and risk receivers, and find that large company stocks play an important role in volatility spillovers in the new energy stock network. Furthermore, we use the Baidu search volume index of 20 keywords related to the physical climate risk to measure the physical climate risk attention and study the impact of physical climate risk attention on the dynamic volatility connectedness among new energy stocks. We find that the effect of physical climate risk attention on volatility connectedness among new energy stocks is significantly positive. The results remain robust through multiple tests.

Searching for soil: Elucidating public interest in soil and soil conservation from 20 years of internet search trends

AbstractOver the last two decades, soil science research has undergone rapid expansion. Understanding public interest in soil science is vital for evaluating dissemination efforts and situating it in the broader environmental discourse. Analyzing Google Trends search data from 2004 to 2023, this study investigates spikes in search volume index (SVI) for soil‐related searches and potential influences. Significant spikes in SVI between 2019–2020 and 2021–2022 were observed for a number of soil characteristics and soil conservation searches. Similar spikes were observed for possible influences such as the documentary “Kiss the ground,” and SVI related to climate change and carbon sequestration. Notably, SVI for “sustainable development goals” aligned with similar patterns in SVI for “soil health,” indicating a possible link between soil interest and the United Nations’ sustainability goals. This study underscores the seemingly rising interest in soil science, possibly linked with dissemination events, and broader environmental concerns and policies.

Does news related to digital economy and central bank digital currency affect digital economy ETFs? Evidence from TVP-VAR connectedness and wavelet local multiple correlation analyses

The rapid and widespread digitalization of economies warrants a better understanding of its impact on society and economy. This study contributes to this new but important research agenda by examining how news on digitalization affects digital economy related exchange traded funds (ETFs). For this purpose, we first construct a digital economy attention index by utilizing Google Search Volume Index for several keywords. Then, using this index and two other indices that represent attention and uncertainty related to central bank digital currency (CBDC), we examine the time-varying connectedness and correlations between these three indices and digital economy ETFs. Our TVP-VAR frequency connectedness analysis shows that attention to and uncertainty around the CBDC and digital economy have strong connectedness with the ETFs in the short-term. The analysis also shows that CBDC and digital economy indices are mainly net transmitters of shocks while the majority of the ETFs are the net receivers of the shocks. The results of our wavelet local multiple correlation (WLMC) analysis show that the correlations between ETFs, digital economy and CBDC indices are time- and frequency-dependent. Moreover, both connectedness and correlations are affected by CBDC-related global events and COVID-19 pandemic. The time- and frequency-dependent relation requires active management of the portfolios containing digital economy related assets.

Cry wolf and it appears: Debt ceiling debates and new corporate bond issues

We provide evidence that attention to debt ceiling debates affects the pricing of new corporate bond issues. Measuring public attention by using data from the Google Search Volume Index with the keyword “debt ceiling,” we find an increase in bond yields for non-investment grade short maturity issues at times of high attention. Further analysis suggests that the results are likely not due to changes in uncertainty or risk aversion, rather, to attention spillover to lower quality firm's short-term financing needs.

Forecasting carbon prices under diversified attention: A dynamic model averaging approach with common factors

To improve the predictability of carbon prices under diversified attention, this paper develops a dynamic model averaging approach with common factors (DMA-CF) which uses dimension-reduction techniques to extract factors from all models including the subsets of attention predictors and allows time-varying coefficients and model switching. The in-sample results using univariate models reveal the strong predictive power of the attention measured by the Google search volume index (AGSVI). The out-of-sample results confirm it as the strongest attention predictor, and show the superior performance of DMA-CFs relative to the original DMA and the benchmark. We further investigate how common factors improve the forecasting performance of DMA. The empirical evidence indicates that common factors efficiently aggregate the information and reduce the estimation errors in complicated models which are assigned with higher probability in DMA-CFs than in DMA, thereby digging out more predictive information for forecasting carbon prices. Moreover, DMA primarily depends on AGSVI, with the highest weights, while DMA-CFs slightly downweight the AGSVI and allocate more weight to the attention proxy of abnormal trading volume (AAVol) which provides complementary information. Finally, our DMA-CF methods can improve the economic gains in the portfolio exercise with carbon futures.

Addressing Google Trends inconsistencies

Google Trends reports the evolution of the popularity of internet searches. Its main output is the Search Volume Index (SVI), a relative measure of the popularity of a term computed using a sample of the searches. Due to the sampling, the SVI series are not entirely consistent, as the same query produces different results that can widely change from day to day. This paper investigates the nature of these inconsistencies by modeling and simulating the data-generating process. Simulations are applied to describe how a typical time series is distorted due to the sampling process and to quantify how averaging extractions smoothes the series. Finally, a relationship between term popularity, series dispersion, and the averaged extractions is derived, so recommendations for constructing consistent SVIs can be provided.

Search symbols, trading performance, and investor participation

We investigate the relationships among search symbols, trading performance, and investor participation. We use two specific datasets from Google Trends’ search volume index. The search volume by number ticker significantly predicts high returns and high investor participation when applied by active retail investors investing in large firms. This does not hold true for less active retail investors who use Chinese company name tickers as their search terms. Our results indicate that the heuristic usage of number tickers to search for company information helps active retail investors to obtain better trading performance compared with less active retail investors who use Chinese company name tickers.

The ability of energy commodities to hedge the dynamic risk of epidemic black swans

The outbreak of COVID-19 increased global financial risks and economic shocks. Although epidemic risk is on the decline, the likeliness of a similar black swan event occurring in the future cannot be ruled out. In this study, we constructed epidemic search indices using the Google Search Volume Index (SVI) to examine whether investor sentiment regarding the COVID-19 pandemic affected stocks, gold, silver, copper, crude oil, and natural gas. The results reveal that, over the entire epidemic period, energy assets (crude oil and natural gas) were the most “resistant” to epidemic panic, making them suitable for hedging the risk associated with epidemic panic risk. We employed a time-varying parameter structural vector autoregression model with stochastic volatility (TVP-SVAR-SV) to analyze the time-varying contagion effects of epidemic search indices on various asset prices at different points in time. The findings indicate that in the early stages of the COVID-19 pandemic when epidemic panic risk was highest, natural gas prices exhibited the most risk-averse characteristics. Only during the middle to later stages of the epidemic did natural gas prices begin to be positively affected by epidemic panic sentiment. However, during this period, the epidemic search indices show a stable declining trend, indicating a reduction in risk. We observe that natural gas prices only reflected panic sentiment during periods of stability. The results of this study suggest that energy commodities, especially natural gas, are suitable for hedging against unforeseen high epidemic risks. When epidemic risk decreases, profits can be realized and reinvested in assets with a higher contagion effect due to epidemic risks, further benefiting from the rebound in the prices of these assets. In addition to explaining why energy commodities are particularly suitable for mitigating epidemic-related black swan events, we provide a detailed investment strategy for hedging future epidemic risk and further elaborate on how investors should dynamically adjust their portfolios according to the severity of the epidemic and whether it is under control.

Global google trends for construction demonstrate low search volume index for stress, mental and suicide

The comparison of Google internet searches for English words in construction for “construction stress”, “construction mental” and “construction suicide” with reference to “construction worker” has not previously been undertaken. It is important to understand internet interest across these 3 terms as all are relevant to mental health and stress in construction. Suicide risk is significantly complex and multifactorial. Our aim is to investigate internet search interest across construction with a focus on mental, stress and suicide, and determine whether there is consistent interest across these search terms. Methods: Using Google Trends, data on global search queries we compared “construction mental” or “construction stress” and “construction suicide”. Two time periods were compared, the last 5 years and the last 24 months, both till December 8th, 2022. The relationship between web search interest, reflected by search volume index (SVI) for society and community versus the business and industrial category and health category were evaluated. Results: Open category searches on Google trends for the key words “construction mental” or “construction stress” demonstrated moderate SVI peaks over a 5-year period. Sub-group analyses for the industrial and business category demonstrated consistent low interest in suicide compared to search terms related to stress or mental health. Conclusion: There is limited online interest in construction mental and stress and even less interest in suicide.

From Global Health to Global Warming: Tracing Climate Change Interest during the First Two Years of COVID-19 Using Google Trends Data from the United States

Climate change mitigation depends on actions that affect the public interest and lead to widespread changes in public attitudes and behavior. With the global outbreak of the COVID-19 pandemic, humanity faced a more imminent threat to its well-being and viability. This retrospective cross-sectional study examines how public interest in climate change was attenuated by the severity of the COVID-19 pandemic using Google Trends Search Volume Index (SVI), weather, and climate data on a United States state-level basis during the first two years of the pandemic from 2020 to 2022. To identify channels through which the COVID-19 pandemic affected information demand on climate change, a novel fixed effect regression model of public climate change interest was developed. The measure captures changes in the climate change SVI independent of weather and climate conditions, comprising pandemic-related changes in living circumstances such as COVID-19-related cases and deaths, mask mandates, and the proportion of the vaccinated population. Our results indicate that public interest in climate change was systematically attenuated by the severity of the COVID-19 pandemic. In addition, this study provides an approach for identifying drivers of public interest in climate change.

Investor attention and the predictability of the volatility of CNY‐CNH spreads: Evidence from a GARCH‐MIDAS model

AbstractCombining the four aspects of self‐, macro, environmental, and policy attention, using backward‐looking rolling regressions, we construct novel international and domestic investor‐attention indices using the search volume index from Google Trends together with Baidu Index to investigate how investor attention affects the CNY‐CNH spreads volatility. Moreover, comparing different GARCH‐MIDAS models and conventional GARCH‐type models is conducted concerning the out‐of‐sample volatility forecasting capability. Our results show that: (i) international and domestic investor attention has a positive impact; and (ii) the GARCH‐MIDAS models involving investor attention improve forecast accuracy. In particular, the model with domestic investor attention has an advantage in forecasting.

Herd behavior in cryptocurrency market: evidence of network effect

PurposeThis study examines herd behavior in the cryptocurrency market at the aggregate level and the determinants of herd behavior, such as asymmetric market returns, the coronavirus disease 2019 (COVID-19) pandemic, 2021 cryptocurrency's bear market and the network effect.Design/methodology/approachThe authors applied the Google Search Volume Index (GSVI) as a proxy for the network effect. Since investors who are interested in a particular issue have a common interest, they tend to perform searches using the same keywords in Google and are on the same network. The authors also investigated the daily returns of cryptocurrencies, which are in the top 100 market capitalizations from 2017 to 2022. The authors also examine the association between return dispersion and portfolio return based on aggregate market herding model and employ interactions between herding determinants such as, market direction, market trend, COVID-19 and network effect.FindingsThe empirical results indicate that herding behavior in the cryptocurrency market is significantly captured when the market returns of cryptocurrency tend to decline and when the network effect of investors tends to expand (e.g. such as during the COVID-19 pandemic or 2021 Bitcoin crash). However, the results confirm anti-herd behavior in cryptocurrency during the COVID-19 pandemic or 2021 Bitcoin crash, regardless of the network effect.Practical implicationsThese findings help investors in the cryptocurrency market make more rational decisions based on their determinants since cryptocurrency is an alternative investment for investors' asset allocation. As imitating trades lead to return comovement, herd behavior in the cryptocurrency has a direct impact on the effectiveness of portfolio diversification. Hence, market participants or investors should consider herd behavior and its underlying factors to fully maximize the benefits of asset allocation, especially during the period of market uncertainty.Originality/valueMost previous studies have focused on herd behavior in the stock market. Although some researchers have recently begun studying herd behavior in the cryptocurrency market, the empirical results are inconclusive due to an incorrectly specified model or unclear determinants.

The Influence of Market Dynamics on Retail Investor Attention

Retail investors have become increasingly active in global markets over the past several years. However, the factors that drive retail investors to focus on particular stocks are unclear. Using a sample of UK FTSE 100 stocks, this paper analyses whether stock volatility, liquidity, returns, and trading volume have the power to attract the attention of UK retail investors, measured using the Google Search Volume Index. Overall, this paper finds positive relationships between three of the dynamics (volatility, returns, and daily trading volume) and increased retail investor attention. Greater stock illiquidity also coincides with an increase in the Google Search Volume Index, although this may be due to liquidity-impacting events. When conditioning on stocks by quartiles of market capitalisation, I find that the effects of returns and trading volume are greater in magnitude for the top 25% of stocks.

Firm Performance and Stock Returns: The Moderating Role of Google Search Volume Index: Evidence from Companies Listed in Indonesian Sharia Stock Index

Firm Performance and Stock Returns: The Moderating Role of Google Search Volume Index: Evidence from Companies Listed in Indonesian Sharia Stock Index