Financial advice and individual investor portfolio performance.This paper attempts to address the question whether financial advisers add value to individual investor portfolio performance by comparing the portfolio performance of advised and selfdirected investors, using a large data set of Dutch investors. (l) Although many individual investors rely on financial advisers to make portfolio investment decisions, until recently, existing literature has largely ignored the added value of financial advice. (2) Recent theoretical and empirical literature suggests an ambiguous contribution of advisers on retail portfolios. In line with predictions of Stoughton, Wu, and Zechner's (2011) model, Bergstresser, Chalmers, and Tufano (2009) suggest a negative relationship between adviser involvement and investor performance in US mutual funds. In addition, Hackethal, Haliassos, and Jappelli (2011) find that risk-adjusted returns are lower for advised portfolios, partly as a result of higher trading costs. Other studies indicate that advisers fail to debias their customers or even exacerbate client biases that are known to hurt returns (Mullainathan, Noth, and Schoar, 2010). In contrast, Bluethgen et al. (2008) find that advisers are associated with better diversified portfolios that are more in line with predefined model portfolios, but with higher fee expenses. Bhattacharya et al. (2011) find that advice taking is associated with an improvement in portfolio performance, though only a fraction of investors are willing to accept and follow advice. Finally, Shapira and Venezia (2001) report that compared with investors who made independent investment decisions, professionally managed portfolios were better diversified and showed better round trip performance due to better market timing. Thus, whether financial advisers improve or worsen portfolio decision making remains an open question to which this paper tries to make a contribution.
Research regarding advised portfolio behavior may be positioned at the intersection of individual and professional portfolio behavior, two research streams that are well established. In early research on the portfolio performance of retail investors, Schlarbaum, Lewellen, and Lease (1978a, 1978b) report risk-adjusted returns of approximately 0% and reasonable levels of skill, though recent empirical studies indicate that average individual investors perform poorly. (3) Within these findings, however, a large heterogeniety in performance can be observed. (4) In addition, the added value of professional money managers has been debated ever since Jensen (1967) first demonstrated that mutual funds do not outperform a buy-and-hold strategy on average (Barras, Scaillet, and Wermers, 2010; Busse, Goyal, and Wahal, 2010; Fama and French, 2010). Yet Binay (2005) argues that institutional investors, including investment advisers, generate excess returns based on their style and stock picking. Other studies that explicitly compare the portfolio performance of individual households with that of professionals find that professionals significantly outperform less sophisticated investors (Grinblatt and Keloharju, 2000; Barber et al., 2009).
This paper differs from the extant literature in several ways. First, in addition to providing a rich set of descriptives that distinguish advised from self-directed investors, I combine analysis of the role of advisers on risk, return, portfolio composition, and timing skills. Additionally, my results likely rely on a more representative data set than previous studies. (5) Moreover, by comparing preand postadvice seeking behavior, I am able to identify effects from advisory intervention and, at least partly, circumvent endogeneity problems that may hinder previously reported results.
Despite differences in investor and portfolio characteristics between advised and self-directed investors, I cannot reject the hypothesis of no return differentials between the two groups. Less idiosyncratic risk exists in advised portfolios because of their greater diversification resulting from more investments in mutual funds, the use of more asset classes, and a lesser focus on domestic equity. The potential for selection effects leaves me to question whether these findings reflect the adviser's influence alone. Less sophisticated investors, for example, may be more inclined to seek advice (Hung and Yoong, 2010). If sophistication and portfolio performance are positively correlated, selection effects may understate the reported results. Evidence from an additional analysis of investors who switch from being self-directed to advised, however, indicates that the results (at least in part) reflect the effect of advisory intervention.
The remainder of this paper is organized as follows. Section I presents the potential costs and benefits of financial advice. After describing the data and summary statistics in Section II, I present the methods and empirical results in Sections III and IV. I provide my conclusions in Section V.
I. Investment Advice and Individual Investor Performance
A. Potential Costs of Investment Advice
When professionals operate in an organizational setting, they are subject to agency relationships that induce incentive-based behaviors (Ross, 1973). The incentives for financial advisers often pertain to different financial concerns, such as producing commissions for their financial institution, generating a performance-based bonus, or enhancing the performance of investors' portfolios (Loonen, 2006). Several theoretical studies model behavioral responses to these incentives and predict that exploitation of unsophisticated clients may occur (Ottaviani, 2000; Krausz and Paroush, 2002; Inderst and Ottaviani, 2009; Stoughton et al., 2011). Bergstresser et al. (2009) provide empirical evidence regarding conflicts of interest between brokers and their clients in the mutual funds market. Broker-sold funds underperform direct-sold funds (before costs). Zhao (2003) reports similar findings. Funds with higher loads tend to receive higher inflows.
Although research indicates that financial professionals tend to be less biased in some ways than laypeople (discussed in the next section), they may be more biased in some other fashion or, given the agency relationship discussed previously, may have an incentive to exacerbate their clients' biases. For example, overconfidence hurts returns (Odean, 1999), but correcting it may be difficult. Overconfidence likely reduces an investor's propensity to seek advice (Guiso and Japelli, 2006). Even when he or she hires an adviser, it is questionable whether that will help. Shapira and Venezia (2001) find more trading activity in professionally managed accounts, which they relate, among other issues, to a possible higher degree of overconfidence for the managed group. Glaser, Weber, and Langer (2010) document that although all participants are overconfident to some extent, financial professionals tend to be more overconfident than laypeople. Kaustia and Perttula (2011) also find overconfidence among a group of financial advisers and some positive effects from debiasing measures. In addition, Kaustia, Laukkanen, and Puttonen (2009) find strong framing effects among a group of financial advisers. Advisers relate higher risk to higher required returns, but to lower expected returns. Thus, while retail investors may suffer from misconceptions related to risk and return (De Bondt, 1998) advisers may not do much better.
Mullainathan et al. (2010) analyze whether advisers tend to debias their clients. They find that although advisers tend to match portfolios to client characteristics, they fail to debias their customers and, in some cases, even exacerbate client biases. That is, the authors find that advisers promote return chasing behavior, encourage holding of actively managed funds, and fail to discourage the holding of their own company stock. In general, advisers tend to support strategies that result in more transactions and higher fees. In addition, Karabulut (2011) indicates that advisers have no influence on stock market participation, but are associated with lower degrees of home bias and less turnover.
B. Potential Benefits of Investment Advice
Hackethal et al. (2011) indicate that economies of scale in portfolio management and information acquisition, as well as advisers' potentially better investment decision making abilities, may help investors improve portfolio performance. Stoughton et al. (2011) rationalize the use of financial advisers by noting that they facilitate small investor market participation by economizing on information costs. It seems likely that, on average, financial advisers are more financially sophisticated than individual investors in terms of investment experience, financial education, and financial knowledge, characteristics linked to improved decision making. Kaustia, Alho, and Puttonen (2008) report that financial market professionals are still biased in their return expectations, but less so than laypeople, while List (2003) finds that the degree of market experience is correlated with the degree of rationality in decision making. Feng and Seasholes (2005) support this finding by reporting that increased sophistication and trading experience are strongly related to the elimination of biased decision making. In addition, Dhar and Zhu (2006) document a negative correlation among financial literacy, trading experience, and the disposition effect. Shapira and Venezia (2001) report that professionally managed accounts exhibit less biased decision making (in terms of the disposition effect) than independent individual investors. These findings all indicate that education and experience reduce behavioral biases that hurt performance, though they may not entirely eliminate them. Finally, Loewenstein (2003) confirms that emotions may have a key impact on decision making, and experience is related to the level of these emotions, such that Lo and Repin (2002) observe significant differences in emotional responses between experienced and less experienced foreign exchange and derivatives dealers.
Beyond these potential benefits resulting from adviser experience, the legal setting provides advised investors some guarantee that financial transactions will fit their characteristics and financial situations. Dutch and European Union (EU) regulations (in particular the Markets in Financial Instrument Directive (MiFID)) require advisers to make recommendations that fit well within an elaborate client profile, whereas for execution-only services, this client profile is much more limited and transactions do not need to be checked against the client profile.
C. Self-Selection of Investors into Advice Taking
In the sample for this study, investors decide whether to hire an adviser. Therefore, differences in behavior and performance between the groups cannot solely reflect the input of the adviser as any difference that emerges is a combined result of investor heterogeneity and adviser influence. Resolving this issue would require running an experiment that assigns participants randomly to an advised or self-directed investor group. (6) Hung and Yoong (2010) similarly implement a hypothetical choice experiment and find that investors with lower financial literacy are more likely to take advice and enjoy better investment performance suggesting a positive effect of advice. Furthermore, they find that older, wealthier people are more likely to use advisers, but are also significantly less financially literate. The notion that less sophisticated investors are more likely to take advice is consistent with the outcomes of theoretical models, such as those proposed by Stoughton et al. (2011), who predict that underperforming active funds sell only though financial advisers to unsophisticated investors and Inderst and Ottaviani (2009), who assume that naive clients do not rationally anticipate advisers' conflicts of interest. Both models imply that advisers mainly service less sophisticated investors.
Therefore, if advised investors are less sophisticated than self-directed investors, assuming that portfolio performance is a function of sophistication in the absence of an adviser, a direct comparison of the two groups would underestimate the added value of financial advice. In addition, Bergstresser et al. (2009) report that clients of brokers are slightly more risk averse. Bluethgen et al. (2008) also find that customers of a German retail bank are older, wealthier, and more risk averse. In this case, risk aversion likely leads to less risky portfolios for investors who take advice and, thus, to lower returns.
D. Account Size
This study compares the results of relatively large portfolios (values exceeding 25,000 [euro] or 100,000 [euro]) with the results of the whole sample since the impact of advisers on portfolios may depend on the portfolios' size. Large portfolios provide a larger profit potential for the bank giving advisers incentive to pay more attention to them. Large portfolios may also contain more complex securities that require more advisory efforts. For example, a large number of (especially large) advised portfolios hold structured products (see Section III.A). Alternatively, since portfolio size is often used as a proxy for sophistication (Anderson, 2008), small portfolios may deviate more from normative recommendations, which may lead to a greater advisory impact even when less attention is paid to it. In Section IV, I formally test the advisory impact on small and large portfolios.
A. The Sample
The primary database comes from a medium-sized, full service retail and business bank that offers an array of financial products. The bank, which advertises itself as a relationship bank, offers services throughout the Netherlands through a network of bank branches, though it has a stronger presence in some regions of the country than others. Customers typically have an account manager who communicates all the financial services the bank offers. For investment advice, clients visit the investment department, although nonclients may also visit this department by making an appointment themselves. Some clients receive advice after they switch from executiononly services. Execution-only and advised investors of the bank receive service from different departments within the bank. Investors with an advisory relationship cannot trade through the execution-only department, nor can investors who use execution-only services trade with the help of an adviser.
During our sample period, all customers were eligible for advice; that is, smaller investors could access advisory services as well. (7) Although most banks require that a minimum amount of money be invested before a client is eligible for advisory services, this was not the case for the bank in this research during the sample period. Note that assignment to a specific adviser is random. Both new and existing investment clients are directed to an adviser depending on availability at the time. Advisers in the sample are paid fixed wages only, so they have no direct personal financial incentive to generate commissions, but career and prestige considerations are likely to play a role.
For all investment clients in the sample, I obtained both position and transaction files for a 52-month period from April 2003 to August 2007. I use only the accounts of private investors with unrestricted accounts excluding any portfolios owned by a business, linked to mortgage loans, or part of a company savings plan. Therefore, the final sample consists of approximately 16,000 investors. To compare the results with those from other empirical studies, I also report results based on common equity holdings, which involves a sample of more than 6,100 investors. Accounts opened or closed during the sample period are included for the months in which they were active. Thus, the data set is free from survivorship bias.
The overall trade file contains the following data fields: 1) account identifier, 2) transaction date, 3) security identification code, 4) transaction type, 5) quantity traded, 6) trade price, 7) currency, and 8) commission paid. The file consists of 535,543 transactions, with a combined market value of 1.6 billion [euro]. Thirty percent of all trades are option trades.8 The position file consists of 2,434,326 investor-security-month positions, which I aggregated into 654,036 monthly individual portfolio statements. The position file also includes information about the type of the client (execution-only or advised), gender, zip code, and date of birth. The six-digit zip code data (representing, on average, 15 households) from Statistics Netherlands (Centraal Bureau voor de Statistiek, 2006) provide information about residential property values and gross incomes.
To obtain an impression about the representativeness of my sample, I compare my sample with the investment portfolios of 1.5 million Dutch households with security investments using data from the Dutch Central Bank (DNB, 2006) in Table I. According to average portfolio size and composition, it seems likely that my sample reasonably represents the average investor in the Netherlands. A 2007 survey (DNB, 2008) suggests that the investment portfolios in my sample represent a significant proportion of financial wealth for most households and cannot be considered a "play account" (Goetzmann and Kumar, 2008). (9) In addition, I compare the portfolios in my sample with samples from other empirical studies of individual investor behavior in the United States, Germany, and the Netherlands (Dorn and Huberman, 2005; Barber and Odean, 2008; Bauer et al., 2009). This comparison reports many similarities in terms of trading style, portfolio composition, and sociodemographics.
B. Measuring Investor Portfolio Returns
In contrast with most empirical studies regarding investor performance, I take a broader perspective to consider all portfolio holdings including mutual funds, bonds, and derivatives, and explicitly account for both the size and the timing of deposits and withdrawals including intramonth trades. For comparison, I provide a separate analysis of returns on common equity positions for the sample.
To calculate portfolio and common equity returns, I use the modified version of the Dietz measure (Dietz, 1968)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (1)
where [R.sup.gross.sub.it] gross is the gross monthly return of investor i in month t, [MV.sub.it] is the end-of-month market value of the whole (or common equity only) portfolio, [NC.sup.gross.sub.it] it is the net contribution (deposits minus withdrawals) in month t before transaction costs, and [w.sub.it] is the weight attributed to this net contribution. This weight is determined by the timing of the contributions. The earlier in the month a contribution takes place, the greater is the weight.
To calculate the net return ([R.sup.net.sub.it]), both transaction costs and custodial fees (including 19% VAT) need to be deducted. Since I use market values in the calculations, I underestimate the actual costs as some market values are observed on an after-cost basis, such as mutual fund market values that are observed after the deduction of various fees (e.g., management fees). For withdrawals that result from a dividend payment, dividend withholding taxes are added back. (10) Bond transactions are net of accrued coupon interest. For every month that a portfolio holds a fixed income security, the coupon (recalculated on a monthly basis) is included in the transaction file. Monthly turnover is calculated by dividing all purchases and sales by the beginning of the month portfolio value. These calculations provide a sample of 604,831 investor-month portfolio return observations and 217,129 common equity return observations. Any missing values indicate investors who invest for less than the whole sample period of 52 months or the elimination of extreme outliers. (11)
The gross and net monthly returns of the average advised and self-directed investors in every month, for use in the time-series regressions, are calculated as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (2a)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (2b)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (3a)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (3b)
where [N.sub.t] is the total number of investors at time t and the subscripts ADV and SD denote advised and self-directed investors, respectively. Thus, I have four time-series of equally weighted returns that serve as the basis for the time-series analysis in Section III.A.
C. Control Variables
Several variables may influence returns. Thus, this research includes the following controls: 1) gender, 2) age, 3) turnover, and 4) wealth (three wealth proxies: portfolio size, residential value, and income, the latter two observed at the six-digit zip code level). Barber and Odean (2001) find that men trade 45% more than women causing them to underperform by almost 1% per year. Komiotis and Kumar (2011) confirm that older, more experienced investors exhibit greater investment knowledge, but they seem to have poorer investment skills, perhaps due to cognitive aging. Portfolio turnover also may hurt net returns (Barber and Odean, 2000) such that the most active traders outperform in gross terms, but underperform in net terms (Bauer et al., 2009). Finally, with regard to wealth, portfolio size is a widely used proxy for investor sophistication. Anderson (2008) reports a positive association between portfolio value and trading performance, and Bauer et al. (2009) indicate that large portfolios outperform small portfolios. However, Barber and Odean (2000) find no significant differentials between the largest and smallest portfolios. Moreover, Dhar and Zhu (2006) report that income, age, trading experience, and portfolio size are all correlated negatively with the disposition effect, a bias that lowers returns (Odean, 1998).
III. Analysis and Results
A. Univariate Results
Table II presents a comparison of investor and portfolio characteristics of advised and self-directed investors. According to Panel A, of the more than 16,000 investors in the sample, approximately 70% are registered with the advisory department for at least one month during the sample period. (12) For portfolios with a value exceeding 100,000 [euro], the percentage increases to more than 90%. The advised group contains more women (27% vs. 24% for self-directed investors) and joint accounts (40% vs. 36%). The average advised investor is somewhat older (56 years vs. 52 years) and the portfolio size is considerably larger than for the self-directed group (70,000 [euro] vs. 15,000 [euro]).
Panel B of Table II further indicates that advised investors perform much worse in terms of gross and net raw portfolio returns. The results also indicate (see Table II, panel D) that advised investors invest a considerably smaller fraction of their wealth in equity, which may explain their lower portfolio returns given the favorable market conditions for equity during the sample period. For equity-only portfolios, the net return differences are much smaller and better for advised investors in the largest portfolios. Return volatilities for both the whole and the equity portfolios are considerably smaller for advised portfolios.
The average portfolio turnover is 4.7% per month (Table II, panel C), less than the 6% reported by Barber and Odean (2000) and much less than the 9% and 24% reported for option and equity traders, respectively, by Bauer et al. (2009). This result likely occurs because the other samples are from Internet brokerage firms, whereas my sample includes investors who use full service or telephone-based, execution-only brokerage services. Although advised investors execute almost twice as many trades (0.27 per month vs. 0.14 per month), they are less active in terms of turnover (4.4% vs. 5.5% per month). Since advised portfolios are generally better diversified, changes require more trades. Furthermore, there is great heterogeneity in trading activity: 45% of the investors never trade and the 1% of the most active investors turn their portfolio over approximately 1.5 times annually. (13)
Panel D of Table II contains the asset allocations indicating large differences in the asset mixes of average advised and self-directed investors. For both groups, equity and bonds represent the main assets (approximately 85% of portfolio value), while advised investors have less risky portfolios. Their asset mix consists of less than 50% equity, whereas self-directed portfolios allocate almost 70% to this asset class. (14) For larger portfolios (exceeding 100,000[euro]), the average equity allocation drops to almost 50% and the difference between advised and self-directed portfolios becomes smaller. The average number of common equity positions is 4.4, but it is higher for advised portfolios (5.3 vs. 3.3), although this difference is mainly due to the higher average portfolio size of advised investors. (15) Larger portfolios hold more common equity positions (almost nine for portfolios over 100,000 [euro]). Well-diversified portfolios may also be obtained by means of mutual funds. In advised portfolios, 66% of wealth is allocated to mutual funds, whereas self-directed investors allocate 48%. Similarly, fund allocation in equity exposure is 63% for advised investors, considerably more than that for self-directed investors (40%). Additionally, less advised investors own options (4.5% vs. 6%). Of the portfolios with average values greater than 100,000 [euro], almost 13% contain options. Although structured products are much less important than equity and fixed income in terms of value, the number of portfolios holding structured products is considerable (23%) and much more prevalent for advised portfolios (28%) than for self-directed portfolios (12%), especially for larger portfolios. This difference may be an indication that banks are pushing these products to exploit uninformed investors (Benet, Giannetti, and Pissaris, 2006; Henderson and Pearson, 2011).
B. Time-Series Analysis of Returns
To analyze risk-adjusted return differences, I calculate the alphas of a long-short portfolio, long on the aggregate equally weighted portfolio of advised investors and short on the aggregate equally weighted portfolio of self-directed investors. As Seasholes and Zhu (2010) note, forming portfolios creates a single time series that is free from cross-sectional correlation. In addition, since advisers may have an incentive to devote most of their attention to larger clients, it might be that the effect of advice is more pronounced for large clients. Therefore, I also create portfolios based on various account sizes.
The regression of the monthly common equity return differences uses a three-factor model developed by Fama and French (1993) to correct for different style tilts in the portfolios. I also regress monthly portfolio return differences using a six-factor model that, beyond the three Fama-French factors, features variations in portfolio characteristics (Bauer et al., 2009). I use the following model to calculate differences in alphas for the overall investor portfolio:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (4)
and I estimate the alpha differences in the equity portfolios as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (5)
In these models, [R.sup.ADV.sub.t] and [R.sup.SD.sub.t] are the average equally weighted returns for the advised and self-directed portfolio in month t, respectively, as calculated in Equations (2a), (2b), (3a), and (3b) (see Section IIB). [R.sub.mt] - [R.sub.ft] is the return on the MSCI Netherlands index in month t in excess of the three-month Euribor. [SMB.sub.t] is the return on a zero investment factor mimicking the portfolio for size. [HML.sub.t] is the return on a zero investment factor mimicking the portfolio for value and [BOND.sub.t] is the excess return on the Iboxx 10-year Dutch Government Index. As in Agarwal and Naik (2004), [CALL.sub.t] ([PUT.sub.t]) is a return series generated by a buying two month at the money index call (put) option at the end of each month and selling it again at the end of the following month. The procedure repeats every month, generating a time-series of 52 monthly returns. To avoid multicollinearity problems, both [CALL.sub.t] and [PUT.sub.t] factors are orthogonalized on the [R.sub.mt] factor. The subscript j denotes the parameter estimated from regression j, where six regressions are estimated for both Equations (4) and (5). The computation of standard errors follows the Newey and West (1987) correction and takes into account autocorrelation up to three lags.
The results in Panel A of Table III indicate that, although the average aggregated portfolios of advised investors are associated with lower returns than the average self-directed portfolios in terms of raw returns, none of the return differences are significantly different from zero. Return differences for larger portfolio become smaller, and remain far from significant. Moreover, the risk-adjusted return (alpha) differences indicate that although a negative sign dominates the various alphas, the hypothesis of no return differentials between advised and self-directed portfolios cannot be rejected at conventional significance levels. Panel B of Table III reports that many of the risk exposures across the various specifications are quite similar, while advised investors expose themselves to less market risk over the entire portfolio, consistent with the lower equity exposure of this group. Note that for both groups, the market betas are quite low, with values of approximately 0.8 for the equity portfolios. Investors in this sample apparently prefer low beta stocks. Indicative of this finding is that the two most widely held stocks, in terms of both value and number of portfolios, have market betas of 0.5 and 0.4, respectively, during the sample period.
C. Cross-Sectional Analysis of Returns
The analysis of risk- and style-adjusted performance indicates no differences between the advised and self-directed investor groups. The previous section treated advised and self-directed investors as a homogeneous group, but as Table II reports, large cross-sectional differences between advised and self-directed investors exist in terms of investor characteristics and portfolio compositions, which are known determinants of investor performance (Section IIC). Therefore, to be able to estimate the coefficient for advice taking, I need to control for these return determinants to avoid a potential omitted variables bias. Specifically, I applied the cross-sectional methodology developed by Fama and MacBeth (1973), which Petersen (2009) indicates provides unbiased statistical inferences when cross-sectional correlation is present. (16) For each month, I ran various cross-sectional regressions, according to Equations (6) and (7), and calculated the Fama and MacBeth (1973) estimators as time-series averages of the monthly cross-sectional parameter estimates. To assess the robustness of the results, I performed similar analyses for the first and second subperiods in the sample. Since the overall sample period can be characterized as a bull market, I performed separate analyses for the months when equity markets showed negative returns, providing an indication as to whether the results hold in more adverse market periods. This approach seems appropriate considering the large fraction of inert investors who probably did not change their portfolio behavior dramatically, even during the recent economic crisis. For returns generated by the whole portfolio, I use the following specification for each month:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (6)
and the following regression on common equity returns:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (7)
In these equations, [R.sub.it] denotes the gross or net portfolio or equity return in excess of the three-month Euribor for investor i in month t. [Advice.sub.it] (the main variable of interest) is a dummy variable equal to one for investors with an adviser in the relevant month and zero otherwise. Womanit is a dummy variable equal to one if the portfolio is held by a woman and zero otherwise. [Join.sub.t] is a dummy variable equal to one if the portfolio is held by two people (usually a married couple) and zero otherwise. [Age.sub.it] is the age of the primary account holder in month t, while [Value.sub.it] is the beginning of the month portfolio market value in month t [or equity value of the portfolio for Equation (7)]. [Turnover.sub.it] is the sum of all purchases and sells in month t divided by the beginning of the month portfolio value (or, for Equation (7), the sum of all equity buys and sells divided by the beginning of the month equity portfolio value). Residential [Value.sub.it] and [Income.sub.it] are the average house value and average gross monthly household income, respectively, based on averages of the six-digit zip code of the area in which the investor lives.
Since the portfolios differ remarkably in terms of asset allocation, it is necessary to control for these differences. Therefore, I use the fractions of the total account value allocated to a specific asset class as a percentage of the total monthly portfolio as an additional control when estimating the cross-sectional regressions on the whole investor portfolio. [Equity.sub.it] refers to both individual stock holdings and equity mutual funds, Fixed [Income.sub.it] indicates individual bonds and bond funds, Real [Estate.sub.it] refers to real estate funds, [Structured.sub.it] is structured products, [Mix.sub.it] involves balanced funds, and [Derivative.sub.it] is a dummy variable equal to one if the account holds derivatives (mainly options) in that month.
The most important finding from the regression results in Table IV is that the coefficient for the Advice dummy, that is sometimes positive, but mostly negative, almost never differs statistically from zero at conventional confidence levels. Some coefficients do approach significance though, such as the whole portfolio returns during the entire sample period (Columns 1 and 2, negative) and the equity returns in the first half of the sample period (Column 10, positive). Overall, however, this analysis does not reveal any clear or robust pattern of out or underperformance. As such, the advised investors do not seem to be performing any better or worse than self-directed individual investors.
Furthermore, in contrast with Barber and Odean's (2001) finding, women do not outperform men and trading activity has a positive impact on gross portfolio and equity returns (Bauer et al., 2009). Apparently, trades are motivated by some informational advantage. However, taking trading costs into consideration makes this advantage disappear. Turnover significantly influences net returns adversely, in line with Barber and Odean (2000).
Of the asset allocation variables, two coefficients are particularly noteworthy. First, equity exposure almost always contributes significantly to returns [e.g., positive for the whole sample period (Columns 1 and 2) and negative in adverse market conditions (Column 3)]. Additionally, derivatives add to the returns for the entire sample period (Column 1), but hurt returns when equity markets fall (Column 3). This finding is intuitive. Most of the derivative traders in the sample buy call options, but this is contradictory to Bauer et al. (2009), who indicate that options traders lose the most in bull markets.
D. Cross-Sectional Analysis of Risk
Although portfolio theory suggests otherwise, retail portfolios tend to be under diversified. Goetzmann and Kumar (2008) indicate that most individual investors hold few stocks in their portfolios and that they often select stocks with similar volatilities, thereby exposing themselves to more avoidable risk (Dorn and Huberman, 2010). Table I already reported that risk in advised portfolios is lower than that in self-directed portfolios. In this section, a more rigorous analysis of this finding provides insight into whether advisers actually have an impact on both total and idiosyncratic risk.
Total risk refers to the standard deviation of monthly portfolio returns for investors with at least 24 monthly returns observations. The calculation of idiosyncratic risk relies on the regressions on the returns in the three-factor (equity portfolio) and six-factor (whole portfolio) models (see Section Ill.B). I do not apply these models on the aggregate portfolio, as previously, but instead use the time-series of returns for each individual portfolio. The standard deviation of the return residuals from these regressions is the idiosyncratic risk measure for each individual portfolio.
Table V presents the results. Panel A provides the comparison of the averages between the two groups. Advised portfolios are associated with lower total and diversifiable risk for both the total and equity-only portfolios. For the total risk measure, this finding should not be surprising. Advised portfolios have less equity in their total portfolio and more equity positions in their equity portfolio, both of which reduce volatility. The lower idiosyncratic risk for advised portfolios means better diversification, but it is necessary to take differences in investor characteristics into consideration as well. I apply a single cross-sectional regression of the various risk measures to the time-series averages of the same investor characteristics discussed in Section III.C. Panel B of Table V indicates that for the whole portfolio, Advise is associated with lower total and lower idiosyncratic risk when controlling for observed investor heterogeneity. Residual volatility is 0.53 percentage points lower for advised portfolios, which is considerable, noting the average standard deviation of monthly return residuals of approximately 2%. (17) The equity-only portfolio reveals no significant differences between the two groups for the sample of all households, but those with values exceeding 25,000 [euro] are associated with less risk. (18) These findings imply that although advisers are not associated with higher returns, their added value lies in guiding investors in their asset allocation decisions to lower avoidable risk.
E. Cross-Sectional Analysis of Asset Allocation
The findings in the previous section indicate that advice is associated with less risk. Since nonsystematic risk is a function of diversification, which in turn is a function of the number of securities and their return correlations, it is worthwhile to examine the diversification and asset allocation decisions of the investors in the sample more closely. Many studies indicate widespread under diversification in retail portfolios, but they are limited as they consider only common equity, even though many households use mutual funds as an effective and easy way to diversify. Polkovnichenko (2005) reports that many households simultaneously invest in well-diversified portfolios of mutual funds and undiversified portfolios of individual stock. Goetzmann and Kumar (2008) report that this under diversification is a function of investor sophistication and related to behavioral biases. If advice introduces more sophistication into a portfolio, better diversification should emerge in advised portfolios.
As proxies for diversification, this analysis uses the fraction of equity, the fraction of mutual funds (in both the whole and the equity-only portfolios), the allocation to index funds, the allocation to domestic equity, the number of different asset classes, and the number of common stocks. In addition, I investigate whether advisers tend to push customers into mutual funds managed by their own banks. Although banks sell their own products, advisers may recommend other mutual funds as well. Therefore, the fraction of own bank funds may be an indication of the use of mutual funds for the benefit of the bank rather than the investor. Table VI contains the results. (19)
The advised portfolios are associated with better diversification for almost all proxies (Panel A): more mutual funds, more index funds, less domestic equity, more asset classes, and more common equity positions. In Panel B, controlling for investor characteristics, the results largely remain the same in sign and magnitude. Advised portfolios are associated with a 21% higher allocation to mutual funds and a 26% increase in the equity portion of the portfolio. Mutual funds provide investors with an easy way to diversify, but advisers may also be tempted to push mutual funds that provide maximal benefits to themselves, perhaps through kickback fees (Stoughton et al., 2011). The data cannot confirm the latter interpretation, but in advised portfolios, a large fraction of the mutual fund holdings is allocated to funds managed by the bank that provided the data. However, this trend is even more evident among self-directed portfolios (Columns 4 and 5, Table VI). Therefore, these mutual funds seem to provide both the adviser and its client with benefits, even though better alternatives may be available to the client (e.g., index funds, which are almost absent in the allocation, Column 6). Home bias is much less pronounced in advised portfolios, largely driven by the higher allocation to mutual funds with greater international exposure. Advised portfolios are also associated with a higher number of asset classes. The number of common equity positions is marginally lower for advised portfolios, but not significantly so. For portfolios higher than 25,000 [euro] and 100,000 [euro], the results are generally quite similar in sign and magnitude. Overall, it seems safe to conclude that advised portfolios achieve better diversification and largely drive the lower idiosyncratic risk in Table V.
F. Timing Returns
No evidence thus far suggests better characteristics or risk-adjusted returns for advised portfolios. However, the added value of advisers might appear in the form of changes to asset allocations that enable investors to benefit from future market movements, rather than stock selection.
To assess whether advised portfolios exhibit better timing ability in their asset allocation decisions, I calculate the returns of various portfolios using passive index returns, similar to Bergstresser et al. (2009). I create these portfolios using both fixed allocation weights based on actual allocations in the first month an investor becomes active and changing allocation weights based on actual asset allocation weights at the beginning of each month. These asset allocation weights are the averages across investors, calculated on both an equal and a value-weighted basis. Five asset classes come into consideration: 1) domestic equity, 2) foreign equity, 3) domestic bonds, 4) foreign bonds, and 5) real estate. These asset classes represent 87% and 89% of advised and self-directed portfolios, respectively. Other asset classes, such as structured products and balanced funds, cannot be tied unambiguously to an index and, therefore, are not included. Consistent with the previous results, the findings in Table VII indicate that excess returns of advised investors are considerably lower due to lower equity exposure in favorable equity market conditions. The risk-adjusted returns, based on Sharpe ratios, again reveal few differences between the two groups. In addition, when I compare the returns of the fixed asset allocation with the monthly rebalancing strategy, timing seems to add marginally to the value-weighted return of both groups. However, because volatilities also tend to rise, Sharpe ratios are practically unaffected. Overall, this evidence suggests that tactical asset allocation does not add to the risk-adjusted return for both groups. (20)
IV. Self-Directed Investors Who Switched to Advice
Thus far, the analysis has compared two groups of investors, those who received investment advice and those who did not. In Section I.C, I emphasized that any differences between these two groups are the combined result of both investor heterogeneity and advisory intervention. In this section, I formally test whether advisers influence portfolio decision making. To this end, I use the group of 228 investors in the data set who switched from being self-directed to being advised (hereinafter, I refer to these investors as switchers). The first investors switched in June 2003 and the last in July 2007. Thus, the sample period is 48 months.
To influence portfolio outcomes, an adviser must propose changes to a portfolio. Therefore, I first analyze portfolio turnover around the date of switch. Figure 1 depicts the average portfolio turnover of switchers during the 12-month event window surrounding the switch month. As this graph illustrates, significant changes occur in the month of and just after the switch. Cumulative mean portfolio turnover in Month 0-3 is more than 50%. The majority of this turnover is caused buy-sell imbalance is positively, though moderately, correlated with lagged stock market returns for both groups of investors (but more so for self-directed investors), implying some return chasing. by reallocation within the existing portfolio. (21) Investors also bring in more capital: One month after switching, the average portfolio size of switchers increased by approximately 7,000 [euro] more than the matched control group.
[FIGURE 1 OMITTED]
I then analyze the changes that occur in the portfolio after the switch to an adviser. I compare the same portfolio allocations of switchers analyzed in Table VI just before and after the switch. Following Barber and Odean (2002), who analyze individual portfolio behavior before and after going online, I employ a matched-pair research design. That is, I match each of the 228 switchers to a self-directed investor who does not switch. This matching occurs in the month preceding the switch by means of a propensity score. The propensity score is the probability of switching and is calculated by regressing a switch dummy (one for switchers and zero otherwise) on several key investor (gender, age, residential value, and income) and portfolio (portfolio value and equity allocation) variables. I use the propensity score of the nonswitching self-directed investor who is closest to the propensity score of the switcher in the month preceding the switch as the matched control. (22) Table VIII, comparing switchers with their matched controls, indicates the effectiveness of this matching.
The main analysis is on the changes in differences in asset allocation between switchers and their matched controls from the preswitch month (t = -1) to the postswitch month (t = 1 and t = 3). Table IX presents the results. For insight into the question whether advisers have more influence on large investors because they put more effort into large portfolios than on small investors who make more investment mistakes, I also split the sample into groups of larger and smaller investors according to the median portfolio value in the preswitch month.
As Table IX illustrates, large asset allocation changes occur for switchers. In line with the results in Table VII, advisers tend to recommend lower equity exposure, more mutual funds, less own bank equity funds, less domestic equity, and more asset classes. As an example of the changes that occur because of advisory intervention, Table IX (Column 3 of Panel A) reports that in the preswitch month, switchers allocate 5.2% less to mutual funds within their equity exposure than their matched peers. Two months later (at t = 1), the difference changes by 5.7%, implying that switchers now have 0.5% higher exposure to equity mutual funds. Another two months later (at t = 3), the difference changes by 8.9%, implying a 3.7% higher equity fund allocation. In general, switchers' asset allocations change in the direction of the allocations of all advised investors reported in Table VI. These results demonstrate that in line with Bluethgen et al.'s (2008) findings, advisers have a significant influence on investors' asset allocations and direct their clients to better diversified portfolios.
Panels B and C of Table IX report the results for smaller and larger investors. For small portfolios, advisers recommend less risky portfolios because of their negative impact on equity exposure (Panel B, Column 1). Larger portfolios already contain considerably less equity. Thus, advisers have no need to make further changes. For the other asset allocation decisions, the largest changes occur in the large portfolios (see Panel C, Columns 2, 3, 5, 7, 8, and 9). Although advisers tend to reduce small portfolios' exposure to own bank equity funds (Panel B, Column 4), they increase large portfolios' exposure to the own bank bond funds (Panel C, Column 5). In general, the allocation to own bank funds remains large.
In the final step, I analyze whether the observed changes in portfolio composition due to advisory intervention have any measurable impact on risk and return. I employ the same methodology as that of Barber and Odean (2002) and compare the returns earned by investors who already switched with those who had not yet switched during the same months. Since the first investors switched in June 2003 and the last in July 2007, I calculate a return series for 48 months. I regress the monthly return differences on the same factors as those in Models 4 and 5. (23) Table X presents the results of this analysis. Gross and net monthly portfolio returns are 27 and 25 basis points lower for the investors who already switched. Differences in alphas are also negative, but much smaller and not significant reinforcing the conclusion that advisers do not enhance or reduce risk-adjusted returns. The factor loadings of Columns 1 and 2 in Table X indicate a significant decrease in the exposure to the market factor of 0.15 after investors switched to advice. This finding is consistent with the notion that advisers lower the equity exposure in a portfolio and increase the fixed income allocation. Within the equity-only portfolio, no significant changes in factor loading are observed from pre- to postadvice seeking behavior (see Columns 3 and 4 of Table X).
This paper provides detailed insights into the differences between advised and self-directed investors and their portfolios and provides evidence of the added value of financial advice. Although I find significant differences in the characteristics of advised and self-directed investors, these differences are quite small in general. Differences in portfolio composition are more noteworthy. Advised portfolios contain significantly less equity and more fixed income securities in line with previous findings that retail investors who seek advice are typically more risk averse.
Analyses of aggregate style-adjusted returns, cross-sectional assessments of returns, and returns based on timing skills indicate that the two groups perform similarly. Also, comparing portfolio returns before and after advice seeking indicates no return effects of advisory intervention. Although, generally, professionals are associated with better portfolio performance than retail investors, it is possible that conflicts of interest eliminate that benefit. The large fraction of advised investors holding structured products may be an indication that this is indeed the case.
In terms of diversification, advised portfolios perform much better than self-directed portfolios, thus reducing avoidable risk. Advised portfolios are associated with more mutual funds, less domestic equity, and more asset classes. Additional analyses on investors who switch to advice taking confirm that advisers positively affect diversification. Therefore, the reduction of idiosyncratic risk observed in advised portfolios can (at least in part) be attributed to advisory intervention.
It is widely known that retail investors make suboptimal portfolio decisions. Although advisers are sometimes subject to similarly biased decision making or have incentives to exacerbate their clients' biases, this paper confirms that advisers do add positive value. They improve portfolio diversification.
I thank the anonymous referee and Bill Christie (Editor) for the many valuable suggestions, and acknowledge the contributions of Werner de Bondt, Frans Tempelaar, Auke Plantinga, Henk yon Eije, and other participants of the conference "Investor Behavior and Stock Market Dynamics '" held at the University of Groningen (2007), participants of the EFA Doctoral Tutorial held in Athens (2008), and participants at the EFMA Conference held in Milan (2009)for useful comments regarding previous versions of this paper. I thank the bank for providing the data.
Agarwal, V. and N. Naik, 2004, "Risks and Portfolio Decisions Involving Hedge Funds," Review of Financial Studies 17, 63-98.
Anderson, A., 2008, "Is Online Trading Gambling with Peanuts?" Institute of Financial Research Working Paper, SIFR Research Report Series, Available at SSRN: http://ssrn.corn/abstract=871435. Accessed: 14 December 2011.
Barber, B.M., Y.T. Lee, Y.J. Liu, and T. Odean, 2009, "Just How Much Do Individual Investors Lose by Trading?" Review of Financial Studies 22, 609-632.
Barber, B.M. and T. Odean, 2000, "Trading is Hazardous to Your Wealth: The Common Stock Investment Performance of Individual Investors," Journal of Finance 55, 773-806.
Barber, B.M. and T. Odean, 2001, "Boys Will Be Boys: Gender, Overconfidence, and Common Stock Investment," Quarterly Journal of Economics 116, 261-292.
Barber, B.M. and T. Odean, 2002, "Online Investors: Do the Slow Die First?" Review of Financial Studies 15, 455-487.
Barber, B.M. and T. Odean, 2008, "All That Glitters: The Effect of Attention and News on the Buying Behavior of Individual and Institutional Investors," Review of Financial Studies 21,785-818.
Barras, L., O. Scaillet, and R. Wermers, 2010, "False Discoveries in Mutual Fund Performance: Measuring Luck in Estimated Alphas," Journal of Finance 65, 179-216.
Bauer, R., M. Cosemans, and P. Eichholtz, 2009, "Option Trading and Individual Investor Performance," Journal of Banking and Finance 33, 731-746.
Benet, B.A., A. Gianneti, and S. Pissaris, 2006, "Gains from Structured Product Markets: The Case of Reverse-Exchangeable Securities (RES)," Journal of Banking and Finance 30, 111-132.
Bergstresser, D., J.M.R. Chalmers, and P. Tufano, 2009, "Assessing the Costs and Benefits of Brokers in the Mutual Fund Industry," Review of Financial Studies 22, 4129-4156.
Bhattacharya, U., A. Hackethal, S. Kaesler, B. Loos, and S. Meyer, 2011, "Is Unbiased Financial Advice to Retail Investors Sufficient? Answers from a Large Field Study," Review of Financial Studies 2012.
Binay, M., 2005, "Performance Attribution of US Institutional Investors," Financial Management 34, 127152.
Bluethgen, R., A. Gintschel, A. Hackethal, and A. Muller, 2008, "Financial Advice and Individual Investors' Portfolios," European Business School Working Paper, Available at SSRN: http://ssrn.com/abstract=968197. Accessed 14 December 2011.
Busse, J.A., A. Goyal, and S. Wahal, 2010, "Performance and Persistence in Institutional Investment Management," Journal of Finance 65, 765-790.
Centraal Bureau voor de Statistiek, 2006, Statistics Netherlands, "Kerncij fers Postcodegebieden," Retrieved from: http://www.cbs.nl/nl-NL/menu/themas/dossiers/nederland- regionaal/publicaties/geografischedata/archief/2006/default.htm.
Coval, J.D., D.A. Hirshleifer, and T. Shumway, 2005, "Can Individual Investors Beat the Market?" Harvard NOM Working Paper, Available at SSRN: http://ssrn.com/abstract=364000. Retrieved: 22 October 2010.
De Bondt, W.F.M., 1998, "A Portrait of the Individual Investor," European Economic Review 42, 831-844. Accessed: 19 August 2010.
Dhar R. and N. Zhu, 2006, "Up Close and Personal: Investor Sophistication and the Disposition Effect," Management Science 52, 726-740.
Dietz, P.O., 1968, "Components of a Measurement Model: Rate of Return, Risk and Timing," Journal of Finance 23, 267-275.
DNB, 2006, Dutch Central Bank, "De beleggingsportefeuille van de Nederlandse particulier," Statistisch Bulletin June, 21-26.
DNB, 2008, Dutch Central Bank, "Vroegpensioen, de levensloopregeling en het vermogen van Nederlandse huishoudens," kwartaalbericht March, 34-37.
Dorn, D. and G. Huberman, 2005, "Talk and Action: What Individual Investors Say and What They Do," Review of Finance 9, 437-481.
Dorn, D. and G. Huberman, 2010, "Preferred Risk Habitat of Individual Investors," Journal of Financial Economics 97, 155-173.
Fama, E.E and K. French, 1993, "Common Risk Factors in the Returns on Stocks and Bonds," Journal of Financial Economics 33, 3-56.
Fama, E.E and K. French, 2010, "Luck versus Skill in the Cross-Section of Mutual Fund Returns," Journal of Finance 65, 1915-1947.
Fama, E.E and J.D. MacBeth, 1973, "Risk, Return and Equilibrium: Empirical Tests," Journal of Political Economy 81, 607-636.
Feng, L. and M.S. Seasholes, 2005, "Do Investor Sophistication and Trading Experience Eliminate Behavioral Biases in Financial Markets?" Review of Finance 9, 305-351.
Glaser, M., M. Weber, and T. Langer, 2010, "Overconfidence of Professionals and Lay People: Individual Differences Within and Between Tasks," University of Mannheim Working Paper, Available at SSRN: http://ssrn.com/abstract=712583. Accessed: 1 December 2010.
Goetzmann, W.N. and A. Kumar, 2008, "Equity Portfolio Diversification," Review of Finance 12,433-463.
Grinblatt, M. and M. Keloharju, 2000, "The Investment Behavior and Performance of Various Investor Types: A Study of Finland's Unique Data Set," Journal of Financial Economics 55, 43-67.
Guiso, L. and T. Jappelli, 2006, "Information Acquisition and Portfolio Performance," Centre for Economic Policy Research Discussion Paper, Available at SSRN: http://ssrn.com/abstract=954121. Accessed: 25 August 2010.
Hackethal, A., M. Haliassos, and T. Jappelli, 2011, "Financial Advisors: A Case of Babysitters?" Journal of Banking and Finance 36, 509-524.
Henderson, B.J. and N.D. Pearson, 2011, "The Dark Side of Financial Innovation: A Case Study of the Pricing of a Retail Financial Product," Journal of Financial Economics 100, 227-247.
Hung, A.A., N. Clancy, J. Dominitz, E. Talley, C. Berrebi, and E Suvankulov, 2008, "Investor and Industry Perspectives on Investment Advisers and Broker-Dealers," Rand Report TR-556-SEC.
Hung, A.A. and J.K. Yoong, 2010, "Asking for Help: Survey and Experimental Evidence on Financial Advice and Behavior Change," RAND Center for Labor and Population Working Paper, Available at SSRN: http://ssrn.com/abstract=1532993. Accessed: 11 January 2010.
Inderst, R. and M. Ottaviani, 2009, "Misselling through Agents," American Economic Review 99, 883-908.
Investment Company Institute, 2007, "Why Do Mutual Fund Investors Use Professional Financial Advisors?" Retrieved 7 April 2007 from http://www.ici.org/home/fm-v16n1.pdf.
Ivkovic, Z., S. Sialm, and S.J. Weisbenner, 2008, "Portfolio Concentration and the Performance of Individual Investors," Journal of Financial and Quantitative Analysis 43, 613-656.
Ivkovic, Z. and S.J. Weisbenner, 2005, "Local Does as Local Is: Information Content of the Geography of Individual Investors' Common Stock Investments," Journal of Finance 60, 267-306.
Jensen, M.C., 1967, "The Performance of Mutual Funds in the Period 1945-1964," Journal of Finance 23, 389-416.
Karabulut, Y., 2011, "Financial Advice: An Improvement for Worse?" Goethe University Frankfurt, Retail Banking Competence Center Working Paper, Available at SSRN: http://ssrn.com/abstract=1710634. Accessed: 11 August 2011.
Kaustia, M., 2010, "Disposition Effect," in H.K. Baker and J.R. Nofsinger, Eds. Behavioral Finance, Hoboken, New Jersey, John Wiley & Sons, 171-189.
Kaustia, M., E. Alho, and V. Puttonen, 2008, "How Much Does Expertise Reduce Behavioral Biases? The Case of Anchoring Effects in Stock Return Estimates," Financial Management 37, 391-412.
Kaustia, M., H. Laukkanen, and V. Puttonen, 2009, "Should Good Stocks Have High Prices or High Returns?" Financial Analysts Journal 65, 55-62.
Kaustia, M. and M. Perttula, 2011, "Overconfidence and Debiasing in the Financial Industry," Aalto University, School of Economics Working Paper, Available at: http://finance.aalto.fi/en/people/kaustia.
Korniotis G.M. and Kumar, A., 2011, "Do Older Investors Make Better Investment Decisions?" Review of Economics and Statistics 93, 244-265.
Krausz, M. and J. Paroush, 2002, "Financial Advising in the Presence of Conflict of Interests," Journal of Economics and Business 54, 55-71.
Leuven, E. and B. Sianesi, 2003, "PSMATCH2: Stata Module to Perform Full Mahalanobis and Propensity Score Matching, Common Support Graphing, and Covariate Imbalance Testing,, Scientific-Commons, Version 4.0.4.
List, J.A., 2003, "Does Market Experience Eliminate Market Anomalies?" Quarterly Journal of Economics 118, 41-71.
Lo, A.W. and D.V. Repin, 2002, "The Psychophysiology of Real-Time Financial Risk Processing," Journal of Cognitive Neuroscience 14, 323-339.
Loewenstein, G., 2003, "Emotions in Economic Theory and Economic Behavior," American Economic Review 90, 426-432.
Loonen, A.J.C.C.M., 2006, "Een Kwestie van Vertrouwen," Doctoral Dissertation, University of Amsterdam.
Millward Brown, 2010, Annual Survey of Dutch Private Investors, retrieved 15 April 2010 from: http://www.millwardbrown.nl/?pm=branches_d&type=publicatie&id=106.
Mullainathan, S., M. Noth, and A. Schoar, 2010, "The Market for Financial Advice: An Audit Study," The National Bureau of Economic Research Working Paper, Available at: http://www.cfr-cologne.de.
Newey, W.K. and K. D. West, 1987, "A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix," Econometrica 55, 703-708.
Odean, T., 1998, "Are Investors Reluctant to Realize Their Losses?" Journal of Finance 53, 1775-1798.
Odean, T., 1999, "Do Investors Trade Too Much?," American Economic Review 89, 1279-1298.
Ottaviani, M., 2000, "The Economics of Advice," University College of London Working Paper, Available at www.kellogg.northwestern.edu.
Petersen, M.A., 2009, "Estimating Standard Errors in Finance Panel Data Sets: Comparing Approaches," Review of Financial Studies 22, 435-480.
Polkovnichenko, V., 2005, "Household Portfolio Diversification: A Case for Rank-Dependent Preferences," Review of Financial Studies 18, 1467-1502.
Ross, S.A., 1973, "The Economic Theory of Agency: The Principal's Problem," American Economic Review 63, 134-139.
Schlarbaum, G.G., W.G. Lewellen, and R.C. Lease, 1978a, "Realized Returns on Common Stock Investments: The Experience of Individual Investors," Journal of Business 51, 299-325.
Schlarbaum, G.G., W.G. Lewellen, and R.C. Lease, 1978b, "The Common Stock Portfolio Performance Record of Individual Investors: 1964-70," Journal of Finance 33, 429-441.
Seasholes, M.S. and N. Zhu, 2010, "Individual Investors and Local Bias," Journal of Finance 65, 1987-2010.
Shapira, Z. and I. Venezia, 2001, "Patterns of Behavior of Professionally Managed and Independent Investors," Journal of Banking and Finance 25, 1573-1587.
Shefrin, H. and M. Statman, 1985, "The Disposition to Sell Winners Too Early and Ride Losers Too Long: Theory and Evidence," Journal of Finance 40, 777-790.
Stoughton, N.M., Y. Wu, and J. Zechner, 2011, "Intermediated Investment Management," Journal of Finance 66, 947-980.
Zhao, X., 2003, "The Role of Brokers and Financial Advisors Behind Investments into Load Funds," College of William and Mary, School of Business Administration Working Paper, Available at SSRN: http://ssrn.com/abstract=438700.
(1) Advised investors have an advisory relationship with the bank that provided the data; self-directed (or execution-only) investors do not have such a relationship. This division is overly simple in that advised investors likely make some investment decisions independent of their advisers, and self-directed investors might hire advisers through different channels. However, on average, the decisions of advised investors in this data set should be influenced more by an adviser than the decisions by the group of self-directed investors.
(2) In the United States, for example, 81% of the households investing in mutual funds, outside a retirement plan, rely on a financial adviser (Investment Company Institute, 2007). Similarly, Bluethgen et al. (2008) indicate that roughly 80% of individual investors in Germany rely on financial advice for their investment decisions, and Hung et al. (2008) find that 75% of US investors consult a financial adviser before conducting stock market or mutual fund transactions. In the Netherlands, the domain of the current research, 51% of households with an investment portfolio rely on financial advice (Brown, 2010).
(3) Their method is based on realized returns, however, causing a positive bias in performance measurement due to the disposition effect (Shefrin and Statman, 1985; Kaustia, 2010).
(4) For other papers on retail investor performance, see Barber et al. (2009), Bauer, Cosemans, and Eichholtz (2009), Ivkovic, Sialm, and Weisbenner (2008), Coval, Hirshleifer, and Shumway (2005), Ivkovic and Weisbenner (2005), Barber and Odean (2000, 2001), and Odean (1998, 1999).
(5) Bergstresser et al. (2009) use aggregated holdings of mutual funds. Hackethal et al. (2011) use data from 10,000 accounts over a 34-month period with an average account value of less than 13,000 [euro], which is unlikely to represent the whole portfolio of the investors in their sample. Bluethgen et al. (2008) use data from less than 4,500 accounts.
(6) Other, more advanced econometric methods also provide ways to deal with self-selection bias. For example, the panel structure of the data set supports fixed- or random-effects regressions. However, the fixed-effects estimator needs time-varying data, which are largely absent from the study data set as few investors switch between groups. The randomeffects model requires the stringent assumption of no correlation between unobserved individual effects and explanatory variables, which seems highly unlikely. For example, investment skill would need to be uncorrelated with gender or wealth. Instrumental variable regressions demand variables that correlate well with the choice of hiring an adviser and not with returns. As is the case for many empirical studies, these variables are unavailable. Thus, I am not confident that these methodologies solve the potential self-selection bias in this case, so I use this more qualitative approach. Note, however, that the analysis of switchers in Section IV aims to identify causal effects of advisory intervention.
(7) The fifth percentile of the portfolio value distribution of advised customers was approximately 600 [euro].
(8) Bauer et al. (2009) report that almost 50% of the trades in their sample are option trades. Their data come from a Dutch online broker.
(9) Respondents reported gross assets of 233,000[euro] on average, 20% (47,000[euro]) of which was invested in financial assets.
(10) In the Netherlands, private investors can neutralize these withholdings in their income tax filings.
(11) I winsorize the return distribution at 1% and 99%.
(12) This 70% represents the investors who were advised during the whole sample period as well as the investors that switched from or to receiving advice.
(13) These details do not appear in Table II, but were derived from additional analyses of the underlying data.
(14) The whole portfolio of self-directed investors may not be observable given the average portfolio size of approximately 15,000 [euro], whereas the average portfolio size in the Netherlands is approximately [euro] 70,000 (see Table I). Thus, these figures may be biased.
(15) For further analysis on this issue, see Table VI.
(16) To test whether this technique is appropriate, I follow Petersen's (2009) advice and compare White standard errors with time-clustered or investor-clustered standard errors. Standard errors are indeed affected when I cluster by time, implying that cross-sectional dependence is present. For standard errors clustered by investor, they rise only fractionally and are well within the margins of Factors 3 and 4, which Petersen (2009) indicates as problematic. This implies that the Fama and MacBeth (1973) procedure is justified here.
(17) Obtained from additional analysis of the underlying data set.
(18) I calculate risk using the individual investor return time series, in which observations were cross-sectionally dependent (see footnote 16) that may somewhat inflate the t-statistics in Table VI. Therefore, especially when t-statistics are small, inference is less certain.
(19) A similar analysis, as described in footnote 16, indicates that both serial correlation and cross-correlation are present and that serial correlation has the greatest impact on standard errors. Therefore, I follow Petersen's (2009) advice and report results based on pooled ordinary least squares estimates with standard errors clustered by investor and the inclusion of time dummies in all specifications.
(20) I confirm this conclusion by performing another analysis on the basis of flows (the results are available upon request). This analysis reveals that the aggregate monthly equity buy-sell imbalance (calculated as in Barber and Odean, 2008) is not correlated with leading equity markets returns, implying no forecasting skills. However, aggregate monthly equity
(21) Approximately 40 percentage points of the 50% turnover is based on reallocation. Since I calculate turnover as the sum of buys and sells in a particular month divided by the beginning of the month portfolio value, on average, 20% of the value of a portfolio is reallocated within three months after the switch.
(22) I use the nearest-neighbor algorithm by employing the Stata module psmatch2 from Leuven and Sianesi (2003).
(23) This is a return series from a long portfolio in which switchers already switched and a short portfolio in which switchers did not yet switch.
Marc M. Kramer *
* Marc M. Kramer is a Researcher in the Department of Economics, Econometrics and Finance, Faculty of Economics & Business, University of Groningen, the Netherlands.
Table I. Comparison Investment Portfolio of Average Dutch with Current Sample This table compares the asset allocations and values of the aggregate portfolio of Dutch households with the current sample as of 2006. DNB Data Own Research Sample Equity Allocation 54% 52% Common Equity 37% 30% o.w. Dutch 75% 81 Equity Mutual Funds 17% 22% Fixed Income Allocation 25% 36% Common Bonds 18% 18% o. w. Dutch 56% 87% Bond Mutual Funds 7% 18% Other Allocation 21% 12% Balanced funds 4% 0% Structured Products 6% 6% Other 11% 6% Total Other 21% 12% Average Portfolio Size 70,000[euro] 65,376[euro] Table II. Characteristics, Performance, Trades, and Portfolios (Related to Portfolio Size) of Individual Investors This table presents descriptives of household and portfolio characteristics split across all households and households with beginning-of-the-month portfolio values exceeding 25,000[euro] and 100,000[euro]. Advised is the percentage of households that receive advice at least once during the sample period. Woman is the percentage of accounts held by a woman only. Joint Account is the percentage of portfolios held by two people. Age is the age of the primary account holder. Account Value is the beginning of the month account value. Residential Value is the home value, while Income is the gross monthly household income, both of which are measured at the six-digit zip code level. Gross and Net Portfolio and Equity Returns (in%) are the cross-sectional averages of the time-series average returns of each individual investor calculated using the modified version of Dietz (1968). Portfolio and Equity Volatility are cross- sectional averages of standard deviations of the time-series of returns calculated according to individual portfolios with at least 24 return observations. Turnover is the sum of buys and sells divided by the beginning of the month account value. (Derivative) Trades are the average number of (derivative) buys and sells per month. Equity, Fixed Income, Real Estate, Structured, Mix, and Mutual Funds refer to fractions of the total account value of specific asset classes. Equity refers to both individual stock holdings ("direct holdings") and equity mutual funds ("fund holdings"), Fixed Income to individual bonds and bond funds, Real Estate to real estate funds, Structured to structured products, and Mix to mix funds. Derivative is the percentage of portfolios that held options at least once during the sample period. Common equity positions are the number of common equity positions in each portfolio. All Households Self- Difference Advised Directed ADV- All (ADV) (SD) Panel A. Characteristics Advised (%) 70% Woman (%) 25.6% 26.7% 23.7% Joint Account (%) 39.2% 40.0% 36.0% Age (years) 55.0 56.4 51.7 Account Value ([euro]) 52,468 69,364 15,101 Residential Value ([euro]) 139,809 140,715 137,577 Income ([euro]) 2,099 2,100 20.96 Panel B. Monthly Raw Return and Risk Gross Portfolio Return (%) 0.70 0.62 0.89 Net Portfolio Return (%) 0.62 0.56 0.80 Gross Equity Return (%) 1.78 1.79 1.75 Net Equity Return (%) 1.45 1.43 1.49 Volatility Portfolio 2.54 2.27 3.21 Returns (%) Volatility Equity 5.10 4.92 5.38 Returns (%) Turnover (%) 4.70 4.36 5.48 Trades (#) 0.23 0.27 0.14 Derivative Trades (#) 0.07 0.08 0.04 Panel D. Portfolio Composition Equity (%) 54.9% 47.9% 68.3% Of which direct 46.1% 37.5% 60.4% holdings (% of equity) Of which fund 53.9% 62.5% 39.6% holdings (% of equity) Fixed Income (%) 30.7% 36.1% 20.0% Real Estate (%) 2.3% 3.0% 0.8% Structured (%) 7.5% 8.5% 5.8% Mix (%) 3.0% 3.4% 2.3% Mutual Funds (%) 61.0% 66.1% 48.5% Derivative (% of 4.9% 4.5% 6.0% portfolios) Structured products 23.1% 28.0% 12.3% (% of portfolios) Common Equity 4.40 5.16 3.26 Positions (#) Portfolios (#) 16,053 All Household Households Portfolio at least 25,000[euro] Self- Advised Directed SD All (ADV) Panel A. Characteristics Advised (%) 84% Woman (%) 3.0% *** 25.1% 25.4% Joint Account (%) 4.0% *** 43.0% 43.5% Age (years) 4.7 *** 61.7 61.6 Account Value ([euro]) 54,263 *** 148,431 163,575 Residential Value ([euro]) 3,138 *** 154,879 155,790 Income ([euro]) 3 2,222 2,232 Panel B. Monthly Raw Return and Risk Gross Portfolio Return (%) -0.27 *** 0.70 0.67 Net Portfolio Return (%) -0.24 *** 0.64 0.61 Gross Equity Return (%) 0.05 1.66 1.66 Net Equity Return (%) -0.06 1.51 1.49 Volatility Portfolio -0.94 *** 2.19 2.10 Returns (%) Volatility Equity -0.46 *** 4.33 4.27 Returns (%) Turnover (%) -1.12 *** 5.21 5.25 Trades (#) 0.13 *** 0.54 0.57 Derivative Trades (#) 0.04 *** 0.16 0.17 Panel D. Portfolio Composition Equity (%) -20.4% *** 47.3% 44.3% Of which direct -22.8% *** 46.5% 43.2% holdings (% of equity) Of which fund 22.8% *** 53.5% 56.8% holdings (% of equity) Fixed Income (%) 16.1% *** 38.9% 40.6% Real Estate (%) 2.2% *** 5.3% 6.0% Structured (%) 2.7% *** 6.1% 6.7% Mix (%) 1.2% *** 2.2% 2.2% Mutual Funds (%) 17.7% *** 52.9% 54.5% Derivative (% of -1.5% *** 9.0% 8.4% portfolios) Structured products 15.7% *** 41.3% 45.3% (% of portfolios) Common Equity 1.90 *** 6.83 7.12 Positions (#) Portfolios (#) 5,120 Household Household Portfolio at least Portfolio at 25,000[euro] least 100,000[euro] Difference ADV- Advised (SD) SD All Panel A. Characteristics Advised (%) 93% Woman (%) 25.1% 0.2% 25.1% Joint Account (%) 38.8% 4.7% ** 41.5% Age (years) 62.0 -0.4 63.9 Account Value ([euro]) 65,559 98,016 *** 319,754 Residential Value ([euro]) 149,128 6,662 * 172,840 Income ([euro]) 2,158 73 ** 2,383 Panel B. Monthly Raw Return and Risk Gross Portfolio Return (%) 0.89 -0.22 *** 0.75 Net Portfolio Return (%) 0.84 -0.23 *** 0.69 Gross Equity Return (%) 1.68 -0.02 1.65 Net Equity Return (%) 1.59 -0.10 1.57 Volatility Portfolio 2.81 -0.72 *** 2.16 Returns (%) Volatility Equity 4.67 -0.41 *** 3.92 Returns (%) Turnover (%) 5.08 0.17 5.98 Trades (#) 0.37 0.20 *** 0.99 Derivative Trades (#) 0.13 0.04 0.32 Panel D. Portfolio Composition Equity (%) 60.1% -15.8% *** 49.7% Of which direct 68.2% -25.0 *** 48.3% holdings (% of equity) Of which fund 31.8% 25.0 *** 51.7% holdings (% of equity) Fixed Income (%) 32.5% 8.1% *** 35.7% Real Estate (%) 2.0% 4.0% *** 6.3% Structured (%) 2.7% 4.0% *** 6.8% Mix (%) 2.1% 0.1% 1.1% Mutual Funds (%) 42.9% 11.6% *** 43.9% Derivative (% of 11.9% -3.5% ** 12.8% portfolios) Structured products 17.7% 27.6% *** 60.1% (% of portfolios) Common Equity 5.65 1.47 *** 8.83 Positions (#) Portfolios (#) 1,867 Household Portfolio at least 100,000[euro] Self- Difference Directed ADV (ADV) (SD) SD Panel A. Characteristics Advised (%) Woman (%) 25.1% 23.7% 1.5% Joint Account (%) 42.0% 37.6% 4.4% Age (years) 63.7 67.3 -3.59 ** Account Value ([euro]) 327,917 181,999 145,917 *** Residential Value ([euro]) 172,845 172,764 81 Income ([euro]) 2,382 2,402 -20 Panel B. Monthly Raw Return and Risk Gross Portfolio Return (%) 0.74 0.87 -0.12 Net Portfolio Return (%) 0.68 0.83 -0.15 * Gross Equity Return (%) 1.66 1.58 0.08 Net Equity Return (%) 1.58 1.55 0.03 Volatility Portfolio 2.12 2.96 -0.84 *** Returns (%) Volatility Equity 3.88 4.39 -0.51 *** Returns (%) Turnover (%) 6.03 5.15 0.88 Trades (#) 1.02 0.66 0.36 Derivative Trades (#) 0.32 0.34 -0.02 Panel D. Portfolio Composition Equity (%) 48.5% 61.1% -12.6% *** Of which direct 46.8% 80.4% -33.6% *** holdings (% of equity) Of which fund 53.2% 19.6% 33.6% *** holdings (% of equity) Fixed Income (%) 36.5% 30.6% 5.8% * Real Estate (%) 6.4% 3.7% 2.7% ** Structured (%) 7.2% 1.3% 5.9% *** Mix (%) 1.1% 1.7% -0.7% Mutual Funds (%) 44.7% 29.1% 15.5% *** Derivative (% of 12.5% 14.8% -2.2% portfolios) Structured products 62.9% 21.5% 41.4% *** (% of portfolios) Common Equity 8.91 7.86 1.05 Positions (#) Portfolios (#) *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table III. Investment Performance of Advised vs. Self-Directed Investors This table presents return differences (in%) and factor loadings of advised and self-directed portfolios. Panel A reports the raw and risk-adjusted gross and net returns differences of a portfolio that is long on the aggregate equally weighted advised portfolio and short on the aggregate equally weighted self-directed portfolio. Households are classified as advised (self-directed) if they were advised (self- directed) during the whole period of the sample of 52 months. Risk- adjusted monthly portfolio (equity) returns are calculated from a six- factor (three-factor) accounting for both the three Fama-French (1993) factors (Market, SMB, and HML) and three additional factors. BOND is the excess return on the Iboxx 10-year Dutch Government Index. CALL (PUT) is a return series generated by buying at two months at the money index call (put) option (see the section on methodology). Panel B provides the estimated factor loadings of these three (six) factors based on the net returns of portfolios of advised and self-directed investors. Standard errors are computed in line with the Newey-West (1987) correction, taking into account autocorrelation up to three lags. The results are expressed in percentages for all households and households with portfolio values exceeding 25,000[euro] and 100,000[euro]. t-statistics are in parentheses. Aggregate, Equally Weighted Portfolios of: All Households Household Portfolio at Least 25,000[euro] Whole Portfolio Equity Only Whole Portfolio Gross Net Gross Net Gross Net Panel A. Return Differences Raw return -0.25 -0.23 0.01 0.02 -0.14 -0.15 (-0.79) (-0.72) (0.02) (0.04) (-0.48) (-0.49) Alpha -0.07 -0.05 -0.04 -0.03 0.00 -0.01 (-1.35) (-0.99) (-0.46) (-0.32) (-0.15) (-0.30) Aggregate, Equally Weighted Portfolios of: Household Portfolio at Least Household Portfolio at 25,000[euro] Least 100,000[euro] Equity Only Whole Portfolio Equity Only Gross Net Gross Net Gross Net Panel A. Return Differences Raw return -0.03 -0.10 -0.11 -0.12 -0.02 -0.09 (-0.05) (-0.17) (-0.33) (-0.38) (-0.04) (-0.16) Alpha -0.02 -0.07 0.04 0.03 -0.05 -0.11 (-0.28) (-0.86) (0.65) (0.39) (-0.69) (-1.11) Aggregate, Equally Weighted Portfolios of: All Households Whole Portfolio Equity Only Self- Self- Advised Directed Advised Directed Panel B. Factor Loadings Market 0.31 *** 0.42 *** 0.78 *** 0.75 *** (15.43) (16.42) (9.70) (8.10) SMB 0.12 *** 0.14 *** 0.26 *** 0.31 *** (5.95) (5.21) (6.23) (5.32) HML -0.02 -0.01 -0.10 * -0.06 (-0.58) (-0.37) (-1.79) (-0.80) Bond 0.14 * 0.09 (1.74) (1.23) Call 0.01 *** 0.01 *** (3.60) (4.68) Put -0.01** -0.01*** (-2.62) (-4.03) [R.sup.2] 67% 79% 85% 81% Aggregate, Equally Weighted Portfolios of: Household Portfolio at Least 25,000[euro] Whole Portfolio Equity Only Self- Self- Advised Directed Advised Directed Panel B. Factor Loadings Market 0.32 *** 0.41 *** 0.80 *** 0.82 *** (17.54) (15.23) (10.23) (9.66) SMB 0.11 *** 0.12 *** 0.21 *** 0.29 *** (7.23) (7.56) (5.29) (9.41) HML 0.01 -0.02 -0.11** -0.07 (0.35) (-0.67) (-2.18) (-1.47) Bond 0.18 *** 0.15 ** (2.79) (2.14) Call 0.01 *** 0.01 *** (3.63) (4.75) Put -0.01- -0.01- (-4.29) (-5.31) [R.sup.2] 78% 83% 83% 86% Household Portfolio at Least 100,000[euro] Whole Portfolio Equity Only Self- Self- Advised Directed Advised Directed Panel B. Factor Loadings Market 0.35 *** 0.45 *** 0.79 *** 0.78 (15.14) (9.57) (10.52) (9.17) SMB 0.10 *** 0.09 *** 0.17 *** 0.22 (6.74) (3.04) (4.41) (7.53) HML 0.03 -0.06 -0.13 * -0.14 ** (1.24) (-1.57) (-2.53) (-2.62) Bond 0.17 *** 0.14 (2.74) (2.74) Call 0.01 ** 0.01 ** (1) (0.01) ** Put 0.0 ** (-5.75) (-4.16) [R.sup.2] 83% 81% 81% 82% *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table IV. Cross-Sectional Differences Portfolio Performance This table presents coefficient estimates on various Fama-MacBeth (1973) regressions on investor and portfolio characteristics. The left-hand side of the table uses portfolio returns (in%) as dependent variables, and the right-hand side refers to common equity returns (in%). Both net and gross returns are used. The full sample covers all 52 months from April 2003-August 2007, and Sub 1 (2) refers to the first (second) 26 months of this period. Downward market refers to all months in which the MCSI-Netherlands had a negative excess return and > 25,000 [euro] to portfolios with beginning of the month account values greater than 25,000 [euro]. The dependent variables are various investor and portfolio characteristics. Advice is a dummy variable equal to one if an investor is used. Woman is a dummy equal to one if the account was held by a woman. Joint Account is a dummy variable equal to one if the account was held by two people. Age is the age of the primary account holder. Value (In) is the logarithm of the beginning of the month account value. Turnover (In) is the common logarithm of the sum of buys and sells divided by the beginning of the month account value. Residential Value (In) is the home value and Income (In) is the gross monthly household income, both of which are measured at the six-digit zip code level. Equity, Fixed Income, Real Estate, Structured, and Mix refer to fractions of specific asset classes of the total account value at the beginning of each month. Equity refers to both individual stock holdings and equity mutual funds, Fixed Income to individual bonds and bond funds, Real Estate to real estate funds, Structured to structured products, and Mix to balanced funds. Derivative is a dummy variable canal to one if the account held options or turbos. t-statistics are in parentheses. Return: Whole Portfolio Equity Only sample: Gross Net Full (1) Full Downward Sub 1 (2) Market (4) (3) Advice -0.07 -0.07 0.01 -0.03 (-1.63) (-1.58) (0.11) (-0.58) Woman -0.02 -0.01 0.02 -0.01 (-0.79) (-0.65) (0.71) (-0.38) Joint Account -0.00 0.00 0.01 0.00 (-0.21) (0.06) (0.53) (-0.21) Age 0.00 ** 0.00 * 0.00 0.00 (2.26) (1.90) (-0.57) (1.14) Value (In) 0.09 * 0.12 ** -0.01 0.16 ** (1.75) (2.34) (-0.12) (2.52) Turnover (In) 0.26 *** -0.13 ** -0.20 * -0.19 * (3.98) (-2.11) (-1.88) (-1.78) Residential 0.09 0.09 0.02 0.17 Value (In) (1.46) (1.44) (0.48) (1.39) Income (In) -0.01 0.00 0.02 0.03 (-0.28) 0.00 (0.35) (0.37) Equity 0.75 0.95 ** -1.49 * 1.13 * (1.67) (2.08) (-1.90) (1.86) Fixed Income -0.47 -0.24 0.29 0.19 (-1.08) (-0.57) (0.36) (0.38) Real Estate -0.01 0.23 0.12 0.90 (-0.03) (0.48) (0.15) (1.26) Structured 0.01 0.22 -0.83 0.21 (0.03) (0.55) (-1.15) (0.42) Mix -0.18 0.01 -0.57 0.36 (-0.44) (0.02) (-0.80) (0.73) Derivative 0.19 * 0.15 -0.32 ** 0.01 (1.90) (1.55) (-2.24) (0.07) Intercept -0.06 -0.44 -0.22 -1.09 * (-0.12) (-0.88) (-0.26) (-1.84) RZ 24.8% 24.6% 23.3% 22.7% Return: Whole Portfolio Equity Only sample: Sub 2 > 25,000 (5) [euro] (6) Advice -0.10 0.18 (-1.51) (1.19) Woman -0.01 0.01 (-0.57) (0.25) Joint Account 0.01 -0.07 (0.37) (-1.22) Age 0.00 * 0.00 (1.76) (0.10) Value (In) 0.08 0.22 (0.98) (1.22) Turnover (In) -0.07 -0.15 ** (-1.14) (-2.37) Residential 0.02 -0.27 Value (In) (0.40) (-0.84) Income (In) -0.03 -0.06 (-0.50) (-0.93) Equity 0.77 0.95 (1.11) (2.87) Fixed Income -0.68 -0.35 (-1.00) (-1.06) Real Estate -0.44 0.11 (-0.67) (0.27) Structured 0.23 0.10 (0.36) (0.38) Mix -0.35 -0.19 (-0.52) (-0.41) Derivative 0.29 ** 0.17 (2.13) (1.47) Intercept 0.21 -0.01 (0.26) (-0.02) RZ 26.6% 29.9% Return: Equity Only sample: Net Gross Full Full Downward Sub 1 (7) (8) Market (10) (9) Advice 0.04 0.01 0.00 0.13 (0.63) (0.13) (-0.04) (1.53) Woman -0.00 -0.01 0.06 0.05 (-0.03) (-0.16) (0.57) (0.32) Joint Account 0.04 0.04 -0.12 ** 0.06 (1.20) (0.98) (-2.36) (0.94) Age 0.00 ** 0.00 ** 0.01 ** 0.00 (2.22) (2.15) (2.59) (1.05) Value (In) -0.09 0.02 -0.20 0.11 (-0.85) (0.24) (-1.22) (0.63) Turnover (In) 0.38 *** -0.31 ** -0.42 * -0.38 (3.06) (-2.47) (-1.95) (-1.65) Residential 0.03 0.06 0.05 0.05 Value (In) (0.35) (0.64) (0.49) (0.34) Income (In) 0.17 0.16 0.13 0.49 (0.73) (0.71) (0.74) (1.14) Equity Fixed Income Real Estate Structured Mix Derivative Intercept 1.13 0.61 -1.80 -0.80 (1.09) (0.60) (-1.54) (-0.44) RZ 3.2% 3.2% 2.7% 4.0% Return: Equity Only sample: Net Sub 2 < 25,000 [euro] (11) (12) Advice -0.12 -0.03 (-1.42) (-0.37) Woman -0.07 -0.07 (-0.80) (-0.77) Joint Account 0.01 0.00 (0.35) (0.10) Age 0.00 ** 0.00 (2.22) (0.70) Value (In) -0.06 -0.02 (-0.52) (-0.30) Turnover (In) -0.24 ** -0.29 ** (-2.24) (-2.38) Residential 0.06 0.13 Value (In) (0.71) (1.12) Income (In) -0.17 -0.26 (-1.38) (-1.47) Equity Fixed Income Real Estate Structured Mix Derivative Intercept 2.02 ** 2.11 ** (2.20) (2.33) RZ 2.3% 5.3% *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table V. Cross-Sectional Differences in Risk This table presents averages (Panel A) and coefficient estimates (Panel B) of risk on various cross-sectional differences between investors. Risk is measured as the standard deviation of the net portfolio and equity returns ("Total risk") and the standard deviation of residuals obtained from regressing each individual net portfolio and equity return time series on the three- and six-factor model discussed previously ("Idiosyncratic Risk"). The left-hand side of the table uses all portfolios, while the right-hand side refers to portfolios with values greater than 25,000 [euro]. In Columns 1, 2, 5, and 6, the dependent variable is the risk of the whole portfolio. The other columns refer to common equity risk. Risk is only calculated when portfolios have at least 24 return observations. The dependent variables are various investor characteristics. Advice is a dummy variable equal to one if an investor is used. Woman is a dummy equal to one if the account was held by a woman. Joint Account is a dummy variable equal to one if the account was held by two people. Age is the age of the primary account holder. Value (In) is the logarithm of the beginning of the month account value. Turnover (In) is the common logarithm of the sum of buys and sells divided by the beginning of the month account value. Residential Value (In) is the home value and Income (In) is the gross monthly household income, both of which are measured at the six-digit zip code level. Robust t-statistics are in parentheses. All Households Whole Portfolio Equity Only Total Ideo- Total Ideo- Risk syncratic Risk syncratic (1) Risk (3) Risk (2) (4) Panel B. Regressions Advised 2.27% 1.59% 4.84% 3.55% Self-Directed 3.21% 2.31% 5.29% 4.08% Difference -0.94% *** -0.72% *** -0.45% *** -0.53 *** Advice -0.73 *** -0.53 *** 0.01 -0.02 (-21.18) (-19.71) (0.14) (-0.42) Woman -0.19 *** -0.12 *** -0.06 -0.05 (-5.57) (-4.81) (-0.98) (-0.88) Joint Account 0.03 0.00 -0.08 -0.17 *** (0.95) (0.16) (-1.64) (-3.53) Age -0.00 *** 0.00 ** 0.01 *** 0.01 *** (-2.73) (2.24) (4.41) (5.10) Value (in) -0.61 *** -0.55 *** -1.09 *** -1.19 *** (-25.41) (-29.67) (-26.82) (-32.57) Turnover (In) 3.11 *** 2.08 *** 1.98 *** 1.84 *** (24.31) (21.23) (12.44) (11.61) Residential 0.52 *** 0.29 *** 0.27 * 0.27 * Value (In) (5.51) (3.95) (1.81) (1.82) Income (in) 0.09 -0.02 -0.24 -0.13 (0.64) (-0.16) (-1.08) (-0.59) Intercept 4.12 *** 3.71 *** 9.10 *** 7.91 *** (10.82) (12.60) (15.69) (13.68) RZ 21.5% 21.9% 25.7% 31.1% Household Portfolio at least 25,000 [euro] Whole Portfolio Equity Only Total Ideo- Total Ideo- Risk syncratic Risk syncratic (5) Risk (7) Risk Panel B. (6) (8) Regressions Advised 2.10% 1.37% 4.24% 2.87% Self-Directed 2.81% 1.94% 4.66% 3.33% Difference -0.72 *** -0.57 *** -0.43 *** -0.46 *** Advice -0.74 *** -0.53 *** -0.26 *** -0.29 *** (-10.06) (-10.21) (-3.26) (-3.32) Woman -0.19 *** -0.08 ** -0.06 0.08 (-3.68) (-2.38) (-0.79) (0.88) Joint Account -0.14 *** -0.09 *** -0.02 -0.07 (-2.80) (-2.89) (-0.28) (-0.90) Age -0.00 ** 0.00 0.00 0.01 (-2.06) (1.09) (0.68) (2.66) Value (in) -0.36 *** -0.42 *** -0.86 *** -0.94 *** (-6.51) (-10.92) (-11.73) (-11.60) Turnover (In) 2.43 *** 1.62 *** 1.29 *** 1.49 *** (15.79) (13.70) (6.63) (7.23) Residential 0.38 *** 0.12 0.02 0.05 Value (In) (2.90) (1.41) (0.12) (0.25) Income (in) 0.03 0.04 -0.31 -0.16 (0.16) (0.29) (-1.09) (-0.49) Intercept 3.65 *** 3.37 *** 9.43 *** 7.54 *** (6.58) (9.22) (11.75) (8.51) RZ 16.6% 18.0% 17.0% 17.0% *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table VI. Cross-Sectional Differences in Asset Allocation This table presents averages (Panel A) and pooled ordinary least squares estimates (Panels B, C, and D) of various asset allocation decisions on cross-sectional differences between investors. The dependent variables are calculated (at the beginning of each month) as follows: Column 1: the value of all equity (including equity mutual funds) holdings divided by the portfolio value; Column 2: the value of all mutual funds divided by the total portfolio value; Column 3: the value of all equity mutual funds divided by the value of all equity holdings; Column 4: the value of the equity funds managed by the "own" bank divided by the value of all equity mutual funds; Column 5: the same as Column 4, but for the bond funds of the bank; Column 6: the value of index equity funds divided by the value of all equity funds; Column 7: the value of the equity holdings listed in the Netherlands by the value of all equity; Column 8: the number of asset classes (defined as equity, bonds, real estate, derivatives, and structured products); and Column 9: the number of common equity positions. Advice is a dummy variable equal to one if an investor is advised. Woman is a dummy variable equal to one if the account was held by a woman. Joint Account is a dummy variable equal to one if the account was held by two people. Age is the age of the primary account holder. Value (In) is the logarithm of the beginning of the month account value. Turnover (in) is the logarithm of the sum of buys and sells divided by the beginning of the month account value. Residential Value (In) is the home value and Income (In) is the gross monthly household income. The last two control variables are determined at the six-digit zip code level. Panels C and D present, for larger portfolios, only the coefficients on the advice dummy and not the controls. All specifications include time dummies. t-statistics (based on investor clustered standard errors) are in parentheses. Dependent Equity Mutual Equity Variable Allocation Fund Mutual Fund as% of Allocation Allocation Total as% of as% of Total Portfolio Total Equity (1) Portfolio Allocation (2) (3) Panel A. Averages Advised 0.48 0.66 0.62 Self-Directed 0.68 0.48 0.40 Difference -0.20 *** 0.18 *** 0.23 *** Panel B. All Households Advice -0.16 *** 0.21 *** 0.26 *** (-19.89) (22.97) (25.06) Woman -0.06 *** 0.07 *** 0.07 *** (-6.12) (7.11) (5.56) Joint Account 0.01 -0.01 0.00 (1.11) (-1.47) (-0.29) Age -0.00 *** 0.00 -0.00 *** (-16.48) (1.39) (-6.60) Residential value (ln) 0.21- -0.19 *** -0.12 *** (8.30) (-6.81) (-3.60) Income (In) 0.12 *** -0.08 ** -0.01 (3.23) (-2.01) (-0.26) Turnover (In) 0.05 *** -0.13 *** -0.11 *** (13.70) (-34.19) (-26.75) Value (In) -0.09 *** -0.06 *** -0.04 *** (-18.99) (-10.39) (-5.73) Intercept 0.42 *** 1.41 *** 0.99 *** (4.23) (13.25) (7.98) [R.sup.2] 14.6% 8.4% 9.7% Panel C. Household Portfolio at Least 25,000 [euro] Advice -0.20 *** 0.18 *** 0.28 *** (-11.32) (9.33) (13.01) [R.sup.2] 8.3% 11.4% 9.8% Panel D. Household Portfolio at Least 100,000 [euro] Advice -0.19 *** 0.18 *** 0.34 *** (-4.36) (4.23) (7.86) [R.sup.2] 3.8% 6.9% 12.0% Dependent Own Bank Own Bank Variable Equity Fund Bond Allocation Allocation in% of in% All Equity of All Bond Funds Funds (4) (5) Panel A. Averages Advised 0.58 0.81 Self-Directed 0.74 0.91 Difference -0.16 *** -0.10 *** Panel B. All Households at Least 25,000 [euro] Advice -0.02 -0.03 *** (-1.21) (-2.72) Woman -0.02 -0.02 (-1.23) (-1.45) Joint Account -0.02 0.01 (-1.44) (0.65) Age 0.00 *** 0.00 *** (7.67) (4.76) Residential value (ln) -0.20*** -0.12*** (-5.82) (-4.22) Income (In) -0.15 *** -0.11 ** (-2.94) (-2.35) Turnover (In) -0.11 *** -0.04 ** (-20.22) (-8.92) Value (In) -0.25 *** -0.12 *** (-31.26) (-14.53) Intercept 2.52 *** 1.91 *** (18.59) (14.92) [R.sup.2] 24.5% 8.6% Panel C. Household Portfolio at Least 25,000 [euro] Advice -0.12 *** -0.05 ** (-3.89) (-2.15) [R.sup.2] 22.8% 12.6% Panel D. Household Portfolio at Least 100,000 [euro] Advice -0.04 -0.03 (-0.60) (-0.51) [R.sup.2] 13.3% 12.5% Dependent Equity Index Home Variable Fund Bias: Allocation Dutch as% of Equity All Equity in% All Funds Equity (6) (7) Panel A. Averages Advised 0.01 0.35 Self-Directed 0.00 0.56 Difference 0.00 ** -0.21 ** Panel B. All Households Advice 0.00 *** -0.24 *** (2.79) (-23.80) Woman 0.00 -0.06 *** (-1.06) (-4.90) Joint Account 0.00 0.01 (0.22) (0.84) Age 0.00 0.00 *** (-1.4) (5.82) Residential value (ln) 0.00 0.08 *** (1.05) (2.66) Income (In) 0.02 0.01 (1.45) (0.15) Turnover (In) 0.00 *** 0.09 *** (3.07) (22.41) Value (In) 0.00 *** 0.05 *** (3.10) (7.67) Intercept -0.08 ** 0.03 (-2.22) (0.27) [R.sup.2] 0.8% 8.8% Panel C. Household Portfolio at Least 25,000 [euro] Advice 0.01** -0.26 *** (2.15) (-12.15) [R.sup.2] 1.1% 8.4% Panel D. Household Portfolio at Least 100,000 [euro] Advice 0.00 -0.32 *** (0.23) (-7.42) [R.sup.2] 1.2% 10.9% Dependent Number Number Variable of of Asset Individual Classes Common (8) Equity Positions (9) Panel A. Averages Advised 1.54 5.25 Self-Directed 1.16 3.16 Difference 0.38 *** 2.09 *** Panel B. All Households Advice 0.14 *** 0.15 (14.32) (1.52) Woman -0.04 *** -0.77 *** (-3.06) (-5.02) Joint Account 0.01 -0.07 (0.51) (-0.49) Age -0.00 *** -0.01 *** (-10.85) (-4.38) Residential value (ln) 0.02 0.52 (0.53) (1.25) Income (In) 0.20 *** 0.87 (3.33) (1.47) Turnover (In) 0.16 *** 0.50 (19.04) (5.84) Value (In) 0.54 *** 2.61 (58.04) (30.14) Intercept -1.28 *** -10.40 *** (-7.85) (-6.55) [R.sup.2] 32.0% 28.3% Panel C. Household Portfolio at Least 25,000 [euro] Advice 0.34 *** -0.56 ** (10.28) (-1.96) [R.sup.2] 25.4% 22.1 Panel D. Household Portfolio at Least 100,000 [euro] Advice 0.83 *** 0.02 (10.30) (0.03) [R.sup.2] 15.0% 17.9% *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table VII. Comparison of Timing Returns This table compares timing skills of advised and self-directed investors. Average actual allocation weights in April 2003 ("Fixed Allocation Weights") and average actual allocation weights at the beginning of each month ("Monthly Rebalancing") are applied to passive index returns. Both value and equally weighted allocation weights are used. The following asset classes are taken into consideration: 1) domestic equity, 2) foreign equity, 3) domestic bonds, 4) foreign bonds, and 5) real estate. Excess return refers to the return above the three-month Euribor. Fixed Allocation Weights Value-Weighted Advised Self-Directed Portfolios Portfolios Mean Excess Return 0.84% 0.98% 0.90% per Month Standard Deviation of 1.47% 1.73% 1.56% Monthly Excess Return Sharpe Ratio 0.57 0.57 0.58 Fixed Allocation Weights Equally Weighted Advised Self-Directed Portfolios Portfolios Mean Excess Return 0.90% 1.11% per Month Standard Deviation of 1.56% 1.96% Monthly Excess Return Sharpe Ratio 0.58 0.56 Monthly Rebalancing Value-Weighted Advised Self-Directed Portfolios Portfolios Mean Excess Return 0.88% 1.04% 0.88% 1.12% per Month Standard Deviation of 1.55% 1.88% 1.55% 2.00% Monthly Excess Return Sharpe Ratio 0.57 0.55 0.57 0.56 Monthly Rebalancing Equally Weighted Advised Self-Directed Portfolios Portfolios Mean Excess Return 0.88% 1.12% per Month Standard Deviation of 1.55% 2.00% Monthly Excess Return Sharpe Ratio 0.57 0.56 Table VIII. Descriptives of Investors Who Switched to Advice This table presents descriptive statistics for 228 investors who switched from being self-directed to being advised ("switchers") and a control group of 228 self-directed investors who did not switch ("matched controls"). This control group was selected according to the nearest-neighbor propensity score in the month preceding the switch. Woman is the percentage of accounts held by a woman only. Joint Account is the percentage of portfolios held by two people. Age is the age of the primary account holder. Account value is the beginning of the month account value. Residential Value is the home value and Income is the gross monthly household income, both of which are measured at the six-digit zip code level. Equity refers to the percentage of total account value invested in equity. Switchers Matched Controls Investors (#) 228 228 Woman (%) 25.0% 21.5% Joint Account (%) 46.1% 36.4% Age (years) 57.2 57.3 Account Value ([euro]) 64,433 55,217 Residential Value (*1,000 [euro]) 132.9 135.7 Income ([euro]) 2,077.9 2,021.3 Equity Allocation (%) 51.8% 50.2% Table IX. Differences in Asset Allocation Before and After Taking Advice This table presents asset allocation differences between 228 investors who switched from being self-directed to being advised ("Switchers" or "SW") and a control group of 228 self-directed investors who did not switch ("Matched Controls" or "MC") before and after the switch. This control group was selected according to the nearest-neighbor propensity score in the month before the switch. Portfolio compositions are compared with the preswitch month (t = -1) for various months after the switch (t = 1 and t = 3). In Panel A, the results for all investors are presented, while in Panel B (C), the sample is restricted to investors with below-(above-) median portfolio value in the preswitch month. The following portfolio allocations are presented: Column 1: the value of all equity (including equity mutual funds) holdings divided by the portfolio value; Column 2: the value of all mutual funds divided by the total portfolio value; Column 3: the value of all equity mutual funds divided by the value of all equity holdings; Column 4: the value of the equity funds managed by the "own" bank divided by the value of all equity mutual funds; Column 5: the same as Column 4, but for the bond funds of the bank; Column 6: the value of index equity funds divided by the value of all equity funds; Column 7: the value of the equity holdings listed in the Netherlands by the value of all equity; Column 8: the number of asset classes (defined as equity, bonds, real estate, derivatives, and structured products), and Column 9: the number of common equity positions. Significance of changes in differences between Switchers and matched controls is based on a paired sample t-statistic (two-sided). Equity Mutual Allocation Fund as% of Allocation Total as% of Portfolio Total (1) Portfolio (2) Panel A. All Switchers Switchers (preswitch: t =-1) 51.8% 58.5% Matched Controls (preswitch: 50.2% 62.8% 1 =-I) SW-MC(preswitch:t=-1) 1.7% -4.3% Change in SW-MC(t=Itot=-l) -2.1% ** 3.8% ** Change in SW-MC (t = 3 to t =-I) -2.8% ** 6.0% *** Panel B. Small Switchers Switchers (preswitch: t=-1) 60.3% 60.5% Matched Controls (preswitch: 57.8% 63.7% t =-1) SW-MC (preswitch: t=-1) 2.6% -3.2% Change in SW-MC (t = l - t =-1) -3.2% * 3.7% * Change in SW-MC (t = 3 to t =-1) -5.5% *** 5.3% ** Panel C. Large Switchers Switchers (preswitch:t=-1) 43.5% 56.5% Matched Controls (preswitch: 42.7% 61.9% t =-1) SW-MC (preswitch:t=-1) 0.8% -5.4% Change in SW-MC (t = 1 to t =-1) -0.9% 3.9% * Change in SW-MC (t = 3 to t =-1) 0.2% 6.6% ** Equity Own Bank Mutual Fund Equity Fund Allocation Allocation as% of Total in% of Equity All Equity Allocation Funds (3) (4) Panel A. All Switchers Switchers (preswitch: t =-1) 50.7% 77.5% Matched Controls (preswitch: 55.9% 81.1% 1 =-I) SW-MC(preswitch:t=-1) -5.2% -3.6% Change in SW-MC(t=Itot=-l) 5.7% ** -5.4% ** Change in SW-MC (t = 3 to t =-I) 8.9% *** -5.0% ** Panel B. Small Switchers Switchers (preswitch: t=-1) 58.8% 91.6% Matched Controls (preswitch: 61.2% 90.7% t =-1) SW-MC (preswitch: t=-1) -2.5% 1.0% Change in SW-MC (t = l - t =-1) 4.9% -6.2% * Change in SW-MC (t = 3 to t =-1) 7.5% ** -5.4% * Panel C. Large Switchers Switchers (preswitch:t=-1) 40.1% 53.6% Matched Controls (preswitch: 48.9% 65.0% t =-1) SW-MC (preswitch:t=-1) -8.8% -11.4% Change in SW-MC (t = 1 to t =-1) 6.8% ** -4.0% Change in SW-MC (t = 3 to t =-1) 10.8% *** -4.4% Own Bank Equity Index Bond Fund Allocation Allocation in% as% of of All Bond All Equity Funds Funds (5) (6) Panel A. All Switchers Switchers (preswitch: t =-1) 90.8% 0.0% Matched Controls (preswitch: 91.6% 0.0% 1 =-I) SW-MC(preswitch:t=-1) -0.7% 0.0% Change in SW-MC(t=Itot=-l) 2.6% 0.0% Change in SW-MC (t = 3 to t =-I) 3.8% * 0.0% Panel B. Small Switchers Switchers (preswitch: t=-1) 98.9% 0.0% Matched Controls (preswitch: 92.0% 0.0% t =-1) SW-MC (preswitch: t=-1) 6.9% 0.0% Change in SW-MC (t = l - t =-1) 0.0% 0.0% Change in SW-MC (t = 3 to t =-1) 0.0% 0.0% Panel C. Large Switchers Switchers (preswitch:t=-1) 85.0% 0.0% Matched Controls (preswitch: 91.3% 0.0% t =-1) SW-MC (preswitch:t=-1) -6.3% 0.0% Change in SW-MC (t = 1 to t =-1) 4.5% 0.0% Change in SW-MC (t = 3 to t =-1) 6.6% * 0.0% Home Number Number Bias: of of Dutch Assest Individual Equity Classes Common in% All (8) Equity Equity Positions (7) (9) Panel A. All Switchers Switchers (preswitch: t =-1) 48.1% 1.46 5.05 Matched Controls (preswitch: 43.6% 1.26 4.96 1 =-I) SW-MC(preswitch:t=-1) 4.5% 0.20 0.09 Change in SW-MC(t=Itot=-l) -6.2% *** 0.10 *** 0.00 Change in SW-MC (t = 3 to t =-I) -9.4% *** 0.19 *** -0.13 Panel B. Small Switchers Switchers (preswitch: t=-1) 40.8% 1.12 3.14 Matched Controls (preswitch: 38.5% 1.15 3.93 t =-1) SW-MC (preswitch: t=-1) 2.3% -0.03 -0.79 Change in SW-MC (t = l - t =-1) -5.5% * 0.09 ** 0.38 Change in SW-MC (t = 3 to t =-1) -8.1% *** 0.13 ** 0.40 Panel C. Large Switchers Switchers (preswitch:t=-1) 57.6% 1.79 6.90 Matched Controls (preswitch: 50.2% 1.37 5.97 t =-1) SW-MC (preswitch:t=-1) 7.3% 0.42 0.93 Change in SW-MC (t = 1 to t =-1) -7.1% ** 0.12 ** -0.35 Change in SW-MC (t = 3 to t =-1) -10.9% *** 0.25 *** -0.56 * *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table X. Performance of Investors Who Switched to Advice vs. Investors Who Did Not Yet Switch to Advice (But Do So Before August 2007) This table presents raw and risk-adjusted net returns (Panel A) and factor loadings (Panel B) of the aggregate equally weighted portfolios of (previously) self-directed investors who already switched to advice and the aggregate equally weighted portfolios of self-directed investors who did not yet switch to advice, but did so before August 2007. The first investors (of 228) switched in June 2003 and the last investors switched in July 2007, providing a time series of 48 months. Risk-adjusted monthly portfolio (equity) returns (in%) are calculated from a six-factor (three-factor) model accounting for both the three Fama-French (1993) factors (Market, SMB, and HML) and the three additional factors for portfolio returns. BOND is the excess return on the lboxx 10-year Dutch Government Index. CALL (PUT) is a return series generated by buying at two months at the money index call (put) option (see the section on methodology). Standard errors are computed in line with the Newey and West (1987) correction taking into account autocorrelation up to three lags. The results are expressed in percentages. t-statistics are in parentheses. Whole Portfolio Return Investors Investors Difference Not Yet Already (AA-NYA) Advised Advised (NYA) (AA) Dependent Variable: (1) (2) (3) Panel A. Returns Raw Return 0.72 0.47 -0.25 (-0.83) Alpha 0.08 0.03 -0.05 (0.51) (0.81) (-0.37) Panel B. Factor Loadings Market 0.38 *** 0.23 *** -0.15 *** 0.79 *** (13.21) (5.97) (-6.00) (7.01) SMB 0.11 *** 0.08 * -0.02 0.29 *** (5.60) (1.98) (-0.62) (5.32) HML 0.01 -0.03 -0.04 0.05 (0.15) (-0.45) (-0.88) (0.56) Bond 0.14 *** 0.27 *** 0.13 (2.74) (2.81) (1.42) Call 0.01 *** 0.01 *** 0.00 (4.18) (2.98) (0.46) Put -0.00** -0.01 ** 0.00 (-2.57) (-2.58) (-1.55) Equity Only Return Investors Investors Difference Not Yet Already (AA-NYA) Advised Advised (NYA) (AA) Dependent Variable: (4) (5) (6) Panel A. Returns Raw Return 1.20 1.09 -0.12 (-0.18) Alpha -0.06 -0.18 -0.14 (-0.21) (-0.68) (-0.62) Panel B. Factor Loadings Market 0.80 *** 0.01 (7.17) (0.24) SMB 0.28 *** -0.01 (5.53) (-0.22) HML 0.03 -0.02 (0.39) (-0.29) Bond Call Put *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level.