Several times a year, Advest, a retail financial-services firm based in Hartford, Conn., runs a sophisticated, high-end education and training program dubbed the Institute at Harvard. A mathematics professor from Case Western University teaches the principles of asset allocation and other high-end wealth-management topics to a subset of the firm's advisers. No doubt it's a coveted assignment since the advisers stay overnight at the Charles Hotel and meet for classes at the Harvard Faculty Club in Cambridge, Mass. "Not everybody gets to go to Harvard," says John Wellington, Advest's director of wealth-management and adviser services.
As part of the firm's transition from a transaction-based to a fee-based advice model, Advest's advanced training program targets 150 to 175 advisers of the total 525. "These are the people we feel are trying to transition and can benefit the most from this kind of education and will implement it," says Welllington. "Someday, we'd like to provide it for everybody, but the reality is not everybody wants to learn and we don't have an open checkbook," he explains.
Nevertheless, Advest is spending money on wealth-management technology to give financial advisers the tools they need to implement the risk concepts. This month, Advest is scheduled to rollout the netDecide Wealth Management System - an enterprise-wide system for financial planning - to 525 financial advisers who will have the tools to run a Monte Carlo simulation at the desktop. (The calculations are run server-side and the adviser sees the results in his Web browser.)
Monte Carlo simulation is an analytical technique that predicts expected returns for various asset classes based on possible outcomes. Named for the gambling casinos in Monte Carlo, Monaco, the random behavior in games of chance is similar to how a Monte Carlo simulation selects values at random for uncertain variables (i.e., interest rates, stock prices, inflation). A simulation can consist of hundreds or even thousands of scenarios per second.
Retail financial services never had the computing power to do Monte Carlo simulations before; certainly not at the desktop. Not only is the data crunching intensive, but the concept of predicting possible outcomes requires a profound shift in money-management philosophy toward embracing uncertainty in analyzing an investor's financial plan.
"In addressing the uncertainty, you come up with better decisions for the client," says Matt Schott, senior analyst at TowerGroup, who authored a report on the subject.
In the following example, Schott explains how Monte Carlo works in comparison with traditional investment management.
"Older financial-planning software says equities are going to return 7 percent. It then projects that the investor needs $1.6 million to retire. Monte Carlo says equities, on average, may return 10 percent, but they're going to vary. Then the adviser (working with the investor) selects returns out 30 years for equities, and a different random rate of returns in case two, and case three. Then you do the same for bonds, cash and stocks. So you get a probability of success. If the adviser selects 70 percent equities and 30 percent bonds, it tells him, flat out, if the investor will have enough money to retire," explains Schott.
The investor can also play with cash-flow assumptions. "What if I increase the amount I save only by $100 each month, what does it do. Because you can run these scenarios very quickly, it tells the investor that nothing is risk free. It tells the investor, who wanted to keep their money in CDs, you have to take a little bit of risk. It also tells the investor they have control over the end data," says Schott.
Advest's Wellington says, "Monte Carlo simulation gives you the probability to a successful outcome toward a stated goal and it incorporates cash flows."
Advest is among a small coterie of retail financial-services firms pioneering a risk-based approach to wealth management. Another firm moving in this direction is JPMorgan Chase's Private Banking Group through Morgan Online, which incorporates risk-management technology from RiskMetrics.
Also, Merrill Lynch - which is outsourcing the building of its new adviser workstation to Thomson Financial - is embedding risk components from Financial Profiles, Ibbotson Associates and RiskMetrics, sources say.
"To me, that's a big revolution, when you have a broker that's starting to talk that language and then buy technology that can help everyone in the firm to do it that way," says Shaw Lively, research manager of Wealth Management at Financial Insights. "That's a big change because five years ago I bet that broker was saying, 'Buy this stock,' and everybody was focusing on making as much money as they could," he says.
The evolution of risk-management thinking began on sell-side proprietary-trading floors about 10 to 15 years ago, and started to move into the institutional buy side about five years ago. Now, it's moving into retail financial services, says Eric Stubbs, managing director, private-client services at Bear Stearns.
There are a number of forces driving the trend: the enormous stock-market volatility and the downward spiral in equities; regulatory winds that are leaning toward giving financial-services firms permission to offer advice if it is based on third-party tools and/or modern portfolio theory; advances in the speed and number-crunching capabilities of personal computers, and improvements in quantitative models now available at price points that are affordable to retail firms.
The poor performance of the stock market and investors' loss of confidence is increasing the demand for this type of advice. "With the pain, people have focused on risk management. When the market was going up, people didn't focus on it," says Stubbs. Regulatory and compliance changes that take into account suitability requirements for investing are also propelling the shift "in spirit, though not yet in the letter of the law" adds Stubbs.
Finally, as brokers increasingly look toward a fee-based business model, as opposed to commissions, they need to give advice. "They've taken on a consultative role, you have to have an appreciation of the risk that a client is exposed to," says Stubbs. "Instead of selling the hot stock of the week, you're looking at saying you should have this much in growth stocks, in bonds and in cash. You've got to be able to explain investment risk responsibly," he emphasizes.
However, this is a dramatic departure from the way retail brokerage firms have historically operated. "Traditionally, retail has been marketed purely based around return," says Ethan Berman, chief executive officer of RiskMetrics, a provider of risk analytics and wealth-management technology. "If you had invested $10,000 in 1990, you would have earned $42,000," he illustrates. "Nowhere does it say how much risk it took to get that return," explains Berman.
Changing the mindset of former stock brokers may be difficult. "The irony is the math is the easy part," claims Berman. "The math has been out there for years and the technology is the easy part because they're making machines faster and they're doing more calculations. The hard part is the strength of the data and the training. That's the part that no one wants to talk about. Instead, they want to talk about stochastic modeling, it sounds fancier than it is."
Under the risk-based approach, the focus is shifting from investment return to meeting goals. "We hope that our people will start to manage clients' assets and client's investment decisions towards the highest probability of a successful outcome based on what the client's goals are," says Advest's Wellington.
The difference from traditional brokerage or investment management is that if a 30-year old walks in with two small children and wants to save for their college education with $10,000 in savings, the investment manager will quickly calculate how to make the $10,000 grow to that goal.
Wellington says that Advest is going to analyze all of an investor's risks, not just investment risks, but insurance management, debt management, disability-insurance management and long-term care.
"To do that," he explains, "you need to take into account what a client's goals and objectives are, what the cash flows are that are available to effect those goals. We like to run that through a Monte Carlo simulator which gives you the highest probability of success," says Wellington, noting that the result could be that the investor has a high probability of meeting their college-funding goals, but a low probability of meeting their retirement-savings goals.
Monte Carlo is the Standard
Increasingly, a number of tools are being provided to retail financial advisers, but Monte Carlo simulation is emerging as the standard, experts say. "Interest is growing in the use of Monte Carlo simulations and other tools that address and illustrate the trade-off of risk versus return," notes TowerGroup's Schott.
He says the usage of Monte Carlo simulation "is a very strong trend in the industry driven by what has happened in the stock market." While less than 5 percent of advisers use Monte Carlo today, TowerGroup estimates this number will grow to over 50 percent by 2007. The technology research and consulting firm estimates that U.S. spending by retail financial-services firms, independent advisers and investment Web sites for tools that incorporate Monte Carlo simulation will grow at a compound-annual-growth rate of 23.6 percent, from $99.3 million in 2002 to $286.4 million in 2007.
"We're probably already at the point where very little advice software will be sold without Monte Carlo simulation," says Schott.
A plethora of vendors offer Monte Carlo simulators, according to TowerGroup. Decisioneering offers a generic Monte Carlo engine that can be used by financial-services firms to create tools. Others, such as Efficient Portfolios, Ibbotson Associates, Plan Scan/PPC and Daniel H. Wagner & Associates provide Monte Carlo and other tools focused on asset allocation and investment planning. Advanced Impact, CCH, Incorporated, EISI, Financeware, Money Tree Software, netDecide, and WealthTec also offer financial-planning software with Monte Carlo.
In Advest's case, it selected netDecide to supply asset allocation, financial planning and portfolio analytics. The vendor has a "life-time-planning cash-flow simulator that will encompass Monte Carlo," says Wellington. "netDecide, we hope, is going to allow advisers to leverage what they learn at (Harvard) with all of their clients, because it gives them desktop capabilities that were only available in the home office," he says.
For instance, netDecide provides asset-allocation/confidence (or probability) intervals. (Confidence intervals provide the likelihood of a return falling within a specified interval, according to netDecide.)
Sam Thomas, the professor from Case Western University who instructs Advest's advisers, teaches what standard deviation and confidence intervals mean, as well as mean-variance optimization (An MVO, such as the one developed by netDecide, creates an optimal diversified portfolio that balances risk and return. The parameters that drive the optimization are typically based on past returns, risks and the co-relationships of the underlying assets, according to netDecide.)
As a result of the MVO, Wellington says, "(Advisers) will be able to not just score a client to a particular model, they'll be allowed to use a mean-variance optimizer to pick a specific model for a client, given their criteria," he says.
But Advest may not turn on all the functionality that the software can offer. Although the netDecide software is capable of recommending specific securities, Advest will use it to recommend a specific asset-class structure. "We're still debating how much functionality we'll give them down to the desktop. But we haven't made the final determination whether we're going to turn that functionality on," says Wellington.
One of the major decisions facing retail financial-services firms as they implement a risk-based approach is how much functionality to throw at their advisers and how much to leave up to interpretation.
"With firms that are pioneering in this area, the real challenge beyond technology is interpretation," says Stubbs. "What (information) are you going to tell your investors which is actionable," he says For example, says Stubbs, "Even if you draw a conclusion that someone in their investment account is taking more risk than the S&P 500, you can't say if that's a good thing or a bad thing, or if they're taking too little or too much risk. It's a very difficult thing," he adds.
Interpreting risk statistics is difficult because it's not clear whether the adviser "should be comparing it to a benchmark or to money-market funds, or to the individual's aspirations," says Stubbs.
Bear Stearns developed a Monte Carlo-based asset-allocation model (with RiskMetrics), says Stubbs. It includes asset-level returns and can run at the security level. "We can get a very finely-tuned picture of an individual's current portfolio and alternatives," he says.
However, unlike some regional firms, such as Advest, that are distributing risk tools to the adviser's desktop, Bear Stearns is centralizing the asset-allocation model. "It's a question of expertise and interpretation," says Stubbs, noting this is an area, on the client side, that is highly technical and requires ever-higher levels of experience and expertise. "At this point in the evolution of the technology and the thinking, the better way to do it is to centralize it," he says.
In addition, Bear Stearns is working with outside vendors to characterize, at the securities level, the value at risk (VAR) of individual portfolios. "VAR will tell you how volatile a person's portfolio is today," says Stubbs. But one of the problems is it is not really predictive, though people want to use it for prediction. "It's only as good as your historical information," he says.
One of the most onerous aspects of running Monte Carlo simulations or other risk-based calculations is obtaining the right data. "You have to do enormous amounts of data crunching. You can do a lot of these calculations in real time or, at worst, overnight," says Stubbs. Though people have different ways of running Monte Carlo simulations, in general, "You should use a lot of historical data," says Stubbs.
The distinction is whether firms do it at the asset-class level or at the securities level, he says, warning that, "If a firm has 5,000 to 10,000 clients, each with a different individual portfolio, that's a very data-intensive undertaking."
Wellington agrees, "There is quite a bit of data necessary to produce high-quality analytics. "The less data you have, the lower the quality of the analytics," he says. Advest purchases the data from Ibbotson, which is the firm's asset-allocation consultant and the industry leader, according to Wellington. "They provide us with capital-market assumptions on all the asset classes and they provide us with model portfolios," he says. NetDecide had to incorporate its data into the platform, he notes.
In addition to risk applications, RiskMetrics also supplies data. "We clean and process over 500,000 time series every day," says Berman, who notes that the firm also has institutional clients that just buy that data. "No one who is buying the applications is not buying the data," he says.
Data, however, is not the only challenge. "The much bigger challenge is training and changing the mindset of the person giving the advice. So many of them have never been financial advisers. They were salesmen and transaction oriented," warns Berman.
In spite of these challenges, retail financial-services firms appear to be betting on the risk-based approach.
How successful is the risk revolution?
The jury is still out because it's too early, says Berman, but he sees the trend evolving in the high-net-worth space where there are more assets at stake and where advisers tend to be more sophisticated.
"It's clear to us that this (trend) is happening," says Berman. "Five years from now the way that advice is given is going to be dramatically different, but it's going to take several years, not several months."
Meanwhile, Advest continues to run the Institute at Harvard. Sessions usually consist of 50 advisers in in which a professor teaches the math behind asset allocation and talks about rebalancing strategies, transaction costs, taxation and how the global economic cycles affect asset allocation. "Our objective is to get up to Harvard twice a year. We're running a lot of sessions," says Wellington.
TowerGroup's Schott estimates that only 15 to 20 percent of brokers are "fairly comfortable talking about (risk) concepts and getting these tools." For the rest of them, "It's going to be a slow transition," he predicts.
But that transition is underway. Reflecting that education is an important part of the transition, a new certification is emerging: chartered investment-management adviser (CIMA). In contrast to the chartered financial analyst (CFA) - who analyzes individual balance sheets of companies and bonds - "The CIMA deals more on a risk level, understands different asset classes and how to put those different asset classes together in a portfolio that makes sense," says Schott. Ivy is Editor-at-Large for Advanced Trading and Wall Street & Technology. Ivy is responsible for writing in-depth feature articles, daily blogs and news articles with a focus on automated trading in the capital markets. As an industry expert, Ivy has reported on a myriad ... View Full Bio