Understanding Investor Behaviour
TD Bank Group
September 2021
"Pay less attention to what (people) say. Just watch what they do."
Dale Carnegie
An Introduction
The actual investment decisions of individuals may be the most honest representation of investor feelings and beliefs. By looking at self-directed investor trading activity, we can see how people react to economic and financial market events. From this we can aim to uncover sentiment patterns of self-directed investors.
TD Direct Investing is in a unique position to help further the understanding of Canadian self-directed investor behaviour. As a market share leader in the Canadian market (number of clients and number of trades0), our data available is the richest of any Canadian self-directed brokerage firm.
In this paper, we present the TD Direct Investing Index (DII), statistics and analysis which harnesses this data. The DII is based on aggregated and de-identified trades placed by TD Direct Investing self-directed investing clients, with reference to contemporaneous market data. The result is an index of self-directed investor sentiment (optimism/pessimism) over a particular historical timeframe. We also present a literature review on sentiment indices, compare our index to other sentiment indices, and detail our methodology for determining the index component proxies.
The data within the DII that was used to create the index can also be segmented to observe the activity of self-directed investors by different age cohorts and even geographic regions. By performing these analyses, we reveal the investing behaviour of self-directed investors to so that they can make the most informed investment decisions for themselves. At the end of this paper, we share an example of how we use the rich dataset to present initial findings that contribute to the literature on investor diversification by demonstrating regional bias of Canadians based on their home province or territory.
Assessing Existing Research on Sentiment
Following the pioneering research by Baker and Wurgler (2006) and referring to the study in Delong, Shleifer, Summers, and Waldmann (1990), investor sentiment is an emotional state of an individual investor. This emotion is tied to the investor's expectation about the future return of their investment assets. As this is a non-physical, intangible psychological state, it is incredibly difficult to measure sentiment directly. An alternative to the challenge of directly measuring the emotions of investors is to define proxies which can infer sentiment. The simplest way of doing this is to survey investors. Unfortunately, this is an imperfect method as investors may articulate one sentiment at the time of surveying but may act very differently when they actually come to invest. Therefore, researchers and behavioral economists alike approach survey results cautiously (Baker and Wurgler 2007). In contrast to using the self-reported survey responses to gauge sentiment, we use aggregated trading activity of self-directed investors to reveal sentiment.
Comparison of the DII and Other Models
Nearly all the sentiment indices found in the literature (presented above) focus on abstracted market-level information as opposed to the actual transactions of an investor analyzed in aggregate. Rarely is it seen that a study focuses on the actual trading behaviour of the investors to probe elements of behaviour and sentiment. The most probable reason behind this is that analyzing investor actions and behaviour requires a highly granular data with records for every transaction (e.g. buy/sell) the investor has conducted. Such data is likely difficult to come by. In cases where transaction data are available, the set of traded securities are different in nature from one study to another (i.e. some studies focus on a specific securities market such as warrants). What's more, published studies focus on aggregate trades and do not make the distinction between trades by individual self-directed investor and institutional investors (e.g. pension funds, insurance companies, etc.). Crucially, the TD Direct Investing platform is utilized by an investor population composed of only self-directed investors and no institutional investors1 , and therefore can better represent the behaviour and sentiment of self-directed investors. We are uniquely able to leverage both granular transaction records of self-directed investors and market data to help address the limitations in prior work that focus on aggregate market metrics, specific security markets, and institutional investors.
Finally, it is noted that there is a lack of consensus—in terms of definitions, assumptions, methodologies and results—in the literature describing stock market dynamics. In part, we attribute this disagreement to the breadth and complexity of stock market parameters, for which researchers do not have complete ability to interrogate. All of this makes it difficult to come to a coherent and somewhat universal conclusion about the necessary components of investor sentiment. In other words, there is no clear consensus on how to determine sentiment—either qualitatively or otherwise.
Despite differences, it is possible to generalize principles of investor sentiment to the self-directed portfolio, even if the trader population and dynamics are not perfectly aligned with those in other studies. A fundamental assumption shared by most researchers is that all investors have the common goal of maximizing their return (regardless of the degree of rationality of their actions). Another assumption that we believe would be fair to make is that the amount and quality of information available to the self-directed investor comparable across markets, though it may be utilized differently by investors.
Review of Other Measures of Sentiment
There are three main approaches to measuring investor sentiment. The first approach is survey-based. An example of such an approach includes the American Association of Individual Investors (AAII) Investor Sentiment Survey, which is a weekly poll of its members' opinion of the stock market over the next 6 months. The drawback of such approach, as mentioned above, is that investors' response may not align with how they will act.
The second approach measures sentiment either based on market variables such as stock prices, trading volume (i.e., number of shares), etc., or based on granular (intraday) investor transaction records. Perhaps the most widely referred study in the market variable-based approach is that of Baker and Wurgler (2006, 2007). In this study, several market variables are proposed in constructing the sentiment:
- Trading volume, or liquidity, increases when investor sentiment is optimistic
- Dividend premium is inversely related to sentiment
- Closed-end fund discount increases when investors sentiment is low
- Initial Public Offering (IPO) first day returns often earn remarkable returns that is in part fueled by investor sentiment. When investor sentiment is high, people tend to buy-up and create demand for IPOs, pushing up prices.
- IPO volume is very sensitive to investor sentiment
- Equity issues over total issues is a broader measure of equity financing activity that reflects a firm shifting between equity and debt and often relate to positive investor sentiment
The authors apply a principal component analysis (PCA, see glossary) on the proxies, which have previously been regressed against key macroeconomic factors to remove business cycle effects. The resulting first component is treated as the composite index for sentiment. Note that, some proxies in the above reveal the sentiment in a delayed manner. As such, for these proxies, their lagged values have been considered.
In another study that leverages market variables, Meier (2018) suggests utilizing stock price data to measure overconfidence in the stock market. Meier argues that confidence has two components: strength and weight. The former measures the extremeness of the available evidence (recent performance versus performance in a benchmark period) while the latter measures the actual credibility of the evidence (are the gains because of good market conditions or rather due to investor's skills?). Here, evidence is simply the difference in past gains. Meier follows the argument of Griffin and Tversky (1992) that investors attribute their recent gains to their own skill rather than other factors (good market conditions or luck) and therefore characterizes overconfidence by high strength and low weight. In another study, Lemmon and Portniaguina (2006) suggest that investor confidence can be measured by considering the variation of small stock prices. This is particularly true for non-institutional investors (i.e., self-directed investorself-directed investors) as they tend to favor small stocks to large ones.
In Burghardt's thesis published 2010, the author leverages self-directed investorself-directed investors option transaction data in the European Warrant Exchange at Börse Stuttgart to build a sentiment index. Four activities, namely: buy-call, sell-call, buy-put and sell-put of options are used to measure the sentiment (see glossary). Naturally, buy-call and sell-put contributes positively to the sentiment while sell-call and buy-put negatively. The author also shows that the proposed transaction-based metric has a high correlation with respected market sentiment indices as well as market returns. This metric is also referred to as the options trajectory.
In a separate study that utilizes self-directed investor transaction data, Kumar and Lee (2006) propose a sentiment index for self-directed investors based on the Buy Sell Imbalance (BSI) metric. This metric first calculates the disparity between buy and sell volumes for each stock on a daily basis. The buy-sell imbalance of each stock is then normalized (see glossary) by stock's volume of transaction. Finally, the normalized imbalance for each stock is aggregated to give the buy-sell imbalance of the whole market. Since the buy-sell imbalance for each stock is normalized by the total buy/sell volume prior to aggregation, the overall metric it not sensitive to the potential bias introduced by large moves from wealthy investors. This metric is sometimes referred to as the Net Equity Demand (NED) or simply the demand-shift, interchangeably.
The third approach measures sentiment based on unconventional, and often unstructured data such as textual data from social media, or factors such as seasonality (Kamstra, Kramer and Levi 2003), sporting events (Edmans, Garcia and Norli 2007) or the occurrence of major events in the news (Li et al. 2014). Perhaps one of the most famous such indices is the BUZZ NextGen AI US Sentiment Leaders Index which scrapes the Web to identify the most mentioned stocks (by taking into account the network influence ranking of the users talking about the stocks), scans social media to check what is said about these identified stocks and then uses natural language processing in order to determine whether the sentiment expressed on these stocks is positive, negative or neutral.
The DII adopts the market-based approach where we consolidate the self-directed investor aggregated trading activity with market variables and leverage the combined data to construct a sentiment index. Compared to the studies surveyed so far, our approach captures the sentiment of a sample of Canadian self-directed investors rather than that of the overall market, which would include both self-directed and institutional investors.
Selecting Proxies for the DII
We utilize anonymous, security-specific transactions to build proxies for self-directed investors from different segments such as age group, geographical region and trading style. We aggregate the proxies to build a sentiment index that reflects the sentiment of Canadian self-directed investors by different demographics. In the construction of the DII, we employ a series of procedures to select the relevant proxies, from which the sentiment index is composed.
In the first step, we filter out proxies that are biased to a particular segment of self-directed investors. Specifically, we ensure that the selected proxies reflect the sentiment of the long-term, average self-directed investor. In this preliminary stage, we also consider the availability of the data required to construct the proxies.
In the second step of the procedure, we employed a proprietary statistical process which utilizes dimensionality reduction techniques (PCA) to further reduce the proxy space
In the final step, we use a qualitative approach where we use expert judgement and effective challenge with subject matter experts to assess and validate the rationale for including various proxies in the sentiment index.
Following the three steps, we reduced the pool of candidate proxies2 down to the four final proxies:
- Bought vs. Sold:
Measures net equity demand—whether self-directed investors were buying more or selling more in a specified month. If positive, they bought more than they sold. If negative, they bought less than they sold.
- Flight to Safety:
Measures how much self-directed investors were pulling back into safer, less risky investments. If negative, self-directed investors traded in lower risk items such as cash, bonds, fixed income, and Money Markets. If positive, self-directed investors traded in higher risk items such as equities.
- Bought at Extremes:
Measures if self-directed investors were buying at either the top or bottom of the market. If positive, self-directed investors placed trades on a rolling 52-week high price. If negative, self-directed investors placed trades in a market dip. (A rolling 52-week high or low means the highest and lowest price the security traded at during a one-year period back from today. The lowest price is often referred to as a market dip.)
- Chasing Trends:
Measures if self-directed investors bought securities on a rising or declining market. If positive, it indicates self-directed investors placed trades on increasing share prices. If negative, DI investors placed trades on decreasing share prices.
Once the set of proxies is determined, we construct the sentiment index by combining the proxies.
Performance of the Sentiment Model
As mentioned previously, investor sentiment is an abstract concept that is difficult to quantify. Therefore, there are multiple approaches to evaluate the validity of the sentiment index. One common way is to compare the sentiment index to a broad benchmark index that represents the market and reflects its sentiment (e.g. bearishness or bullishness) in aggregate. Even if the market is composed of both institutional and self-directed investors, it is expected that some relationship exists between the market return and the sentiment of self-directed investors. This is important because if sentiment doesn't move with the equity markets, the effectiveness of the sentiment index is lessened. Because we think there is a relationship, we conduct the following tests as evidence.
The resulting index can be seen in Figure 1. This is the raw outcome of the model, where positive (negative) values mean bullish (bearish). For benchmark index, we have chosen the S&P/TSX Composite Index. This is a popular index of Canadian publicly traded securities. We can see that over the time period analyzed in our graph, our DII aligns with the TSX - our index is bullish while the TSX is bullish.
Figure 1: TD DII Compared to the TSX Index
In developing the DII, we noted there is a tremendous wealth of research on the relationship between sentiment indices and stock market returns (Brown and Cliff, 2004; Smales, 2017; Bathia et al., 2013). From this, we test using a Vector Autoregression (VAR, glossary), whether the lag of our sentiment index impacts the TSX (or vice versa). Here we show the impulse response of the two variables on each other. We also use an Ordinary Least Squares (OLS, glossary) method for the contemporaneous impact, where the TSX can influence sentiment within the same month (rather than with a lag).
Below are the impulse response functions of the VAR estimates. The resulting impulse response shows statistical significance of the TSX’s influence on self-directed investor sentiment but does not show a significant impact of self-directed investor sentiment on the TSX. Given the scale of self-directed investors relative to that of institutional investors, we would not expect a strong response of the market to the impulse of self-directed investors' sentiment. In other words, this data suggests that self-directed investors tend to react to the market, but rarely drive the market.
The above charts show the impulse response of the TSX Index and DII Sentiment (if one variable moves, how much of an impact does it have on other variables). Top left, the impact of the TSX on itself (if the TSX goes up this month, what are the chances it goes up next month). Top right, the impact of the TSX to DII Sentiment (if the TSX goes up this month, what are the chances DII sentiment goes up next month). Bottom left, the impact of DII Sentiment from the TSX. Bottom right, the impact of DII Sentiment on itself. Readings above (below) zero show a positive (negative) relationship of one variable on another.
Given impact of the TSX on the DII sentiment, as demonstrated by the VAR estimation, we further this analysis by conducting an ordinary least squares estimation (glossary). The independent variables, lagging sentiment and a contemporaneous TSX (same time), are regressed against the monthly sentiment score. Here we find statistical significance: the monthly sentiment score is explained by its lagging score and the concurrent month's market performance.
See glossary for definitions.
Taken together, these tests indicate that each month's sentiment is influenced by its historical score as well as past and present market performance. This is consistent with our expectation that a relationship should exist between the monthly sentiment and the concurrent month's market performance; however, the relationship is not always exact given the fact that institutional investors are usually the market movers.
In addition to the sentiment-TSX interdependency validation test, we have also assessed the stability of the model over time (each month we look at the output of the model to make sure it is logical). The stability of the model is important as a degraded model can no longer represent the current sentiment accurately. The results of the stability test suggest that the model performance when trained on the entire dataset is very much aligned with the performance of the model when trained on a fraction of the data (on an ongoing basis). A model monitoring plan is also in place to ensure that in the future, we can identify any model drift in order to recalibrate the model in a timely manner.
Additional Insights from DII3
As outlined at the onset of this paper, one of our key objectives was to understand the degree of diversification within self-directed investor portfolios, and whether bias influences investment decisions. The data within the DII that was used to create the index can also be segmented to observe the activity of self-directed investors by different age cohorts and even geographic regions. Access to this information goes to the spirit of the creation of the DII, as an educational tool where we share insights with self-directed investors so that they can make the most informed investment decisions for themselves.
For the period of December 2020, we use the transactional records of self-directed investors and perform additional analysis to derive insights of self-directed investor behaviour, providing additional context to interpreting the DII.
Activity By Age Group
In Canada, there was an interesting divergence between age categories for certain sectors. In Table 1 we see that Financial stocks were clear favorites for all self-directed investor groups, but we find that the allocation to that sector increased with age. Though we cannot determine the exact cause this, we do know that Financials provide greater dividends than other sectors, and income streams can be an important factor for Canadians in older age categories. Additionally, Financial corporations typically have been in operation for a long time, and the comfort with such storied companies may impact investor preference.
Technology was second highest allocation for Canadians younger than 35 and those 35-to-50. This compares to Energy being the second most popular sector for Canadians 51 years of age and older. This divide may be a result of younger self-directed investors openness to new companies, versus older self-directed investors comfort with historically successful companies.
Table 1: Top Two Portfolio Sector Weights by Age Cohort (December 2020)
| Age |
Sector Allocation |
| Highest |
Second Highest |
| Sector |
% |
Sector |
% |
| Less than 35 |
Financials |
25.7 |
Technology |
15.7 |
| 35 to 50 |
Financials |
27.5 |
Technology |
13.9 |
| 50 to 65 |
Financials |
30.1 |
Energy |
13.5 |
| Over 65 |
Financials |
34.8 |
Energy |
12.9 |
Sector Allocation by Geography
We can also breakdown the sector allocation by geographic region. Here we found significant geographical differences across provinces and territories. In Figure 2 and Table 2, we show the sectors with the largest differences between regional allocations. In Figure 4, we highlight the provinces of Alberta, Ontario, Saskatchewan, Nova Scotia, Quebec, and British Columbia for descriptive purposes. Table 2 shows the complete dataset as of December 2020.
Figure 2: Sector Allocation by Province
Table 2: Detailed Sector Allocation by Province
| |
Mat. |
Com.
Serv. |
Cons.
Disc. |
Cons.
Stpl. |
En. |
Fin. |
Heal.
Care |
Ind. |
Real
Estate |
Tech. |
Utilities |
| TSX |
12.2 |
4.7 |
4.0 |
3.5 |
12.0 |
30.9 |
1.6 |
12.3 |
3.2 |
10.9 |
4.8 |
| AB |
7.9 |
4.9 |
4.9 |
2.4 |
22.5 |
26.4 |
6.6 |
6.4 |
3.8 |
8.7 |
5.3 |
| NTU |
14.8 |
5.4 |
3.9 |
3.9 |
12.6 |
26.5 |
8.0 |
9.0 |
3.5 |
6.0 |
6.4 |
| ON |
6.5 |
6.8 |
5.7 |
2.6 |
11.7 |
32.3 |
7.6 |
6.6 |
4.1 |
11.6 |
4.4 |
| NL |
11.2 |
5.3 |
4.5 |
2.9 |
11.0 |
27.2 |
6.8 |
5.4 |
4.6 |
9.7 |
11.4 |
| NB |
8.9 |
7.7 |
4.7 |
2.7 |
10.8 |
33.3 |
8.1 |
7.4 |
3.6 |
7.4 |
5.5 |
| SK |
12.9 |
4.4 |
4.6 |
2.6 |
17.1 |
30.3 |
5.9 |
6.8 |
3.7 |
7.5 |
4.1 |
| NS |
6.7 |
6.2 |
4.0 |
2.9 |
10.5 |
37.0 |
6.4 |
6.4 |
4.1 |
6.9 |
8.8 |
| PE |
5.8 |
7.0 |
3.9 |
3.4 |
10. |
33.5 |
7.3 |
6.7 |
3.9 |
8.9 |
9.0 |
| YT |
11.5 |
6.9 |
4.7 |
2.4 |
11.1 |
30.1 |
6.4 |
8.4 |
3.8 |
9.2 |
5.4 |
| QC |
6.4 |
7.3 |
5.6 |
4.5 |
7.5 |
30.4 |
8.6 |
10.5 |
3.7 |
11.9 |
3.4 |
| BC |
9.7 |
6.0 |
6.4 |
2.9 |
11.2 |
28.7 |
7.0 |
6.8 |
4.2 |
11.9 |
5.2 |
| MB |
7.6 |
6.3 |
5.5 |
3.5 |
10.6 |
34.2 |
7.0 |
8.1 |
4.3 |
8.9 |
4.1 |
Data are in percent, as of December 2020.
For the Materials sector, we note that self-directed investors across the Territories and Saskatchewan had much greater exposure than self-directed investors in other locations. Materials companies can include those related to metals (such as gold, copper, and iron ore) and non-metals (such as potash and diamonds).
For Energy, investors in Alberta, Saskatchewan, and the Territories were most exposed. Energy companies largely include those related to oil and gas.
For Financials, we saw that self-directed investors in the Maritimes and Ontario had the most exposure. These include companies such as banks and insurance companies.
This regional breakdown revealed a potential home preference/bias. It may suggest that investors from geographic locations where employment and economic production are dependent on a specific sector tend to have an overweight of equity exposure to that sector as well. This is best exemplified by the Energy sector. Here we saw that self-directed investors who live in the geographic areas with the most economic exposure to Energy were also most overweight Energy in their portfolios.
Concluding Thoughts
In this paper we use the aggregated and anonymous trading data of TD Direct Investing self-directed investing clients to help build a measure of investor sentiment. This index helps determine how self-directed investors were feeling about equity markets over historical period.
We have also presented details of the dataset to help improve our clients understanding of risk-taking and investor bias. This has direct ties to help our clients be aware of the benefits of portfolio diversification. Here we show how investment exposure to certain higher risk sectors was influenced by age. We also show that self-directed investors may have a home bias by overweighting sectors that are economically more prominent in the province or territory in which they live. The evidence of age impacting risk-taking and geographic home bias is important in our understanding of self-directed investor behaviour.
Glossary
Normalization: In statistics and its applications, normalization refers to a process whereby values are adjusted to allow for meaningful cross-comparisons. Normalization may be implemented to bring different measures to a notionally common scale to prior to averaging.
Ordinary least squares estimation: This is a statistical technique used to understand the relationship between a dependent variable and one or more independent variables. The coefficient tells us the estimated magnitude of effect of the independent variable on the dependent variable, as well as the directionality (e.g., increase or decrease) of that relationship. On the other hand, standard error tells us the precision in which the coefficient is measured. If a coefficient is large compared to its standard error, then it implies that there is some relationship between the independent and dependent variable.
Principal component analysis: This is a statistical technique used to reduce the dimensionality of data while preserving as much information as possible of the original dataset. This is achieved by creating new variables, or principal components, that contains information of the original variables and maximizes the information or variance.
Proxies: In the DII context, proxies are investing behaviour measures based on trade activity that allow us to make inferences about investor sentiment.
Put options: A put option gives the owner the right to sell an underlying security at a specific price until a certain date. When selling a put option (sell-put), the seller agrees to buy a stock at an agreed-upon price. It's also known as shorting a put. The seller is anticipating that the stock price will rise in value.
Buy-call: a bullish trade that gives the buyer the choice to exercise the option, allowing them to buy the underlying security at the strike price
Sell-call: a bearish trade that if exercised by the buyer, forces them to sell the underlying security at the strike price
Buy-put: a bearish trade that gives the buyer the choice to exercise the option, allowing them to sell the underlying security at the strike price
Vector autoregression estimation: This is a statistical technique used to capture the relationship between multiple quantities as they change over time. This technique is useful for understanding how a variable is a function of past lags of itself and past lags of the other variables.
References
Baker and Wurgler (2006): Investor Sentiment and the Cross-Section of Stock Returns
Delong, Shleifer, Summers, and Waldmann (1990): Noise Trader Risk in Financial Markets
Baker and Wurgler (2007): Investor Sentiment in the Stock Market
Meier (2018): Aggregate Investor Confidence in the Stock Market
Griffin and Tversky (1992): The Weighing of Evidence and the Determinants of Confidence
Lemmon and Portniaguina (2006): Consumer Confidence and Asset Prices: Some Empirical Evidence
Kamstra, Kramer and Levi (2003): Winter Blues: A SAD Stock Market Cycle
Edmans, Garcia and Norli (2007): Sports Sentiment and Stock Returns
Li et al (2014): The Effect of News and Public Mood on Stock Movements
Huang et al (2015): Investor Sentiment Aligned: A Powerful Predictor of Stock Returns
Balcilar et al (2017): Predicting Stock Returns and Volatility with Investor Sentiment Indices: A Reconsideration Using a Nonparametric Causality-in-Quantiles Test
Matthias Burghardt (2010): Retail Investor Sentiment and Behavior – an Empirical Analysis
Kumar and Lee (2006): Retail Investor Sentiment and Return Comovements
Brown and Cliff (2004): Investor Sentiment and the Near-Term Stock Market
Smales (2017): The Importance of Fear: Investor Sentiment and Stock Market Returns
Bathia, Deven, and Don Bredin (2013): An Examination of Investor Sentiment Effect on G7 Stock Market Returns
Disclaimer
The information contained herein has been provided by TD Direct Investing and is for information purposes only. The information has been drawn from sources believed to be reliable. Graphs and charts are used for illustrative purposes only and do not reflect future values or future performance of any investment. The information does not provide financial, legal, tax or investment advice. Particular investment, tax, or trading strategies should be evaluated relative to each individual's objectives and risk tolerance.
The TD Direct Investing Index (DII) provides data and insights relating to historical self-directed investor activity. The DII is for informational purposes only. Any information provided through the DII should not be considered an investment recommendation, nor is it an offer, or solicitation of an offer to purchase or sell any investment fund, security or other product. Particular investment, trading, or tax strategies should be evaluated relative to each individual’s objectives. Investors should not take the historical information as an indication, assurance, estimate or forecast of future values or future performance. The DII should not be used as individual financial, legal, investment or tax advice. Please consult your own legal, investment and/or tax advisor. Information provided through the DII is subject to change without notice.
In April 2022 we began a three-month transition of the DII methodology. During this period we slowly adjusted the proxy logic and data modelling to improve the quality of our analysis. This transition was completed for the July 2022 data, at which time we also adjusted the monthly data to include the full calendar month. Accordingly, comparisons between periods with different methodologies may not be as accurate as comparisons between periods of the same methodology.
The Toronto-Dominion Bank and/or its subsidiaries or affiliated persons or companies may hold a position in the securities mentioned, including options, futures and other derivative instruments thereon, and may, as principal or agent, buy or sell such securities. They may also make a market in, issue, and participate in an underwriting of such securities.
A high degree of risk may be involved in the purchase and sale of options and may not be suitable for every investor. The risk of loss in trading securities, options and futures can be substantial. Investors must consider all relevant risk factors, including their own financial situation before trading. A higher level of market knowledge, risk tolerance and net worth is required.
TD Bank Group means The Toronto-Dominion Bank and its affiliates, who provide deposit, investment, loan, securities, trust, insurance and other products or services.
TD Direct Investing is a division of TD Waterhouse Canada Inc., a subsidiary of The Toronto-Dominion Bank.
All trademarks are the property of their respective owners.
®The TD logo and other trademarks are the property of The Toronto-Dominion Bank or its subsidiaries.
1 The Global Industry Classification Standard (GICS®) was developed by and/or is the exclusive property of MSCI, Inc. and S&P Global Market Intelligence Inc. (“S&P Global Market S&P GICS Distribution Agreement Intelligence”). GICS is a service mark of MSCI and S&P Global Market Intelligence and has been licensed for use by The Toronto-Dominion Bank.
Footnotes
0 Reported by Investor Economics in the “Online/Discount Brokerage Market Share Report" for the quarter ending June 30, 2021.
1 However, note that TD Direct Investing clients have a broad range of investing sophistication – including some whose trading traits are similar to institutional investors.
2 We have explored a diverse set of proxies (including those from the cited papers). We arrive at the most salient proxies using the 3-step proxy selection process.
3 Insights are observations and not forward-looking. They are subject to change based on future data.
Canada
US
