Clark Cheng is the chief investment officer at Merrimac Corp., a private investment firm based in Stamford, Connecticut primarily investing in hedge funds. In this interview, he talks about how family offices can be efficient and dynamic compared to other large institutions; his approach for evaluating and selecting great managers; and how an unconstrained investment process gives family offices the opportunity to generate higher alpha.
Prior to joining Merrimac, he was a hedge fund manager HSBC’s Alternative Investment Group. Before that, he performed hedge fund research at Guggenheim Partners and Morgan Stanley between 2002 and 2005. He holds the Chartered Financial Analyst, Financial Risk Manager and Chartered Alternative Investment Analyst designations.
Clark Cheng was named as one of Trusted Insight’s Top 30 Family Office Chief Investment Officers. The following interview has been edited and condensed for clarity.
Trusted Insight: In what ways is Merrimac unique within the family office space?
Clark Cheng: I think we are very effective at allocating resources towards a single purpose. In general, at large institutions, you have to worry about all risks equally, especially those related to reputation and business risk which are less relevant to a business with no clients. In any investment business, as you add clients your incentives and motivations change, which causes the behavior of the investment professionals to change. The number and types of clients, in addition to how the business is paid for their services, will have an impact on how the business is managed and consequent performance. Larger businesses with risk-averse clients and low management fees in addition to no incentive fees will be managed to reduce business risk in order to protect an annuity stream. There is no incentive or motivation to take the risk for the institution or investment professional. The most successful investment organizations are able to understand and overcome these misaligned interests.
In most institutions, you have large checklists and requirements that need to be completed before multiple investment committees can read a lengthy report and then debate the merits of the investment. The problem is that you allocate limited resources inefficiently. You spend the same amount of time on a small investment as you do a large investment. You spend the same amount of resources on a high-risk investment as you do a low-risk investment. Allocating limited resources inefficiently is a problem for many large institutions. Family offices who only have one client can be very efficient and dynamic.
Risk
Risk is measured historically through the four moments of a distribution, return, variance, skewness and kurtosis. Most investors only focus on the first two moments, return and volatility, but skewness is so important in a market dislocation yet difficult to estimate before it is too late. More importantly, exactly risk is a function of multiple factors from investment risk to business risk to behavioral risk. Understanding what drives risks will be important in generating a better return relative to the desired risks. Allocating limited resources inefficiently is a problem for many large institutions. Family offices who only have one client can be very efficient and dynamic.
Relative to other family offices, we have a different approach to investing. The only free lunch in the world is diversification, which is only important during a drawdown. Volatility and correlations are beneficial to compounding capital when it works in your favor. You can take more manager risk if you have true diversification. The biggest risk in a concentrated portfolio is negative skew. An uncorrelated fund with a Sharpe ratio above 1.5 and leverage, illiquidity and pricing risk probably has negative skew and a potential for a large drawdown. This risk can be diversified away in the portfolio construction process through diversification.
Many investors believe they can manage a concentrated portfolio of high conviction managers which is easy in hindsight but difficult to do going forward. I have never seen a bad model/paper portfolio track record. As much as I disliked studying statistics in school, statistics are important in every aspect of creating a portfolio. Assuming a high decision hit rate of 70 percent, the probabilities only work if you make enough decisions in a diversified portfolio. If you run a concentrated portfolio with few decisions, the probabilities are not allowed to work for you and spurious results can drive returns. The goal is to have consistent repeatable returns through a process and spurious results don’t help.
More importantly, you and the organization have to acknowledge mistakes will be made and you need to minimize the impact of those mistakes and learn from them. If the investment process does not allow for people to make mistakes while investing, the investment professional will be motivated to minimize his or her mistakes and consequent career risk. The resulting organization will not have a performance-driven culture. If the investment team is incentivized correctly to openly discuss investment opportunities candidly, good decisions are more likely.
Team
The optimal investment team is five people. Any more or less is not efficient and five allows for a tiebreaker. Based on research, effective teams have higher collective intelligence when they have higher social and cognitive diversity, meaning they have more uncorrelated knowledge and skills to solve complex problems. Diversification across social characteristics include race, ethnicity, gender, and age and cognitive characteristics include education, functional knowledge, expertise, training, experience and abilities.
We are investing in people, not securities, who change over time. We are diversified across a large number of managers with exposures to multiple factors, such as geography, sector, styles, time horizon and risk tolerance. Investing in higher vol managers who are more concentrated or levered without negative skew will create a higher Sharpe/Calmar ratio if the managers are truly uncorrelated. Correlations, however, are not stable and only matter in a drawdown. Correlation benefits can be achieved by building a portfolio with exposure to multiple factors. Individual high Sharpe ratios are only important in a concentrated portfolio of managers because there is more risk per manager, but it is more efficient to build a Sharpe ratio through the portfolio construction process and use diversification to your benefit. Investing in the best managers with a portfolio of intended but diversified exposures is important in managing the downside risk of a portfolio.
Besides selecting the best managers, performance comes from sizing managers and allocating to the right strategies and factors at the right time. Many managers have similar crowded stocks in their portfolios, but sizing and timing the stocks in addition to managing net exposure (directionality) and gross exposure (leverage) determines individual performance. The same applies to managing a portfolio of managers. Allocating to the right strategies, factors, and styles at the right time determine performance. The skill in investing is portfolio and risk management not just picking stocks or managers.
Trusted Insight: In your last interview, you said that “data is limited” when it comes to selecting managers in the alternative business. Can you talk about discuss a time when you were able to distinguish a great manager from a good one?
Clark Cheng: Data is limited, listed references are typically good, and the marketing stories are all well narrated. This business would be very easy if everyone was honest. Historical returns are helpful in questioning how a manager trades around different market environments and drawdowns, but the goal is to predict the future distribution of returns with the data you have.
Given a managers' risk parameters and exposures, you are trying to find managers that have higher positive returns with less risk and negative skew. Looking at historical volatility can be misleading if you don’t understand the negative skew in the future distribution of returns. Leverage, pricing risk and illiquidity can create negative skew in market dislocations and you need to increase your hurdle rate or put a cost on this future risk.
There is also a behavioral element to this potential risk. Some managers are risk seeking and others are risk averse so understanding the manager is important in assessing the probability of a large drawdown in a market dislocation. There is a lot of research that shows how many people become irrational and emotional during drawdowns which is not desirable.
Relationships are extremely important when it comes to reference checks with other investors and previous employers. Most managers are telling you a well-rehearsed story and your job is to discern what is real and what is not. Nurturing relationships with people you trust will help you get the candid references you need to make decisions. Likewise, managers also conduct references on investors to assess who they think would make a good partner or not.
Lastly, the experience can provide you with valuable data to help you make decisions. Understanding how funds are structured and how they operate is important to understanding new managers. For example, new managers always want to have portfolio management experience but not be responsible for losses at their prior firms. You need to understand how decisions are made at their prior firm to know whether new managers ever had decision-making abilities and portfolio management experience. Many new managers have never managed through a market dislocation having started their careers after the last financial crisis which was ten years ago.
Additionally, funds have drawdown limits and non competes. If a new manager shows a volatile month from their prior firm or is able to start a new job before the end of a typical non compete, he was probably fired from his prior firm. When there is a mutual decision for a person to leave a manager, chances are they were fired. Covering a wide breadth of managers for a long period of time helps you build relationships and understand how each fund works.
Trusted Insight: When it comes to hedge fund fee structures, do you have more leeway when working at a family office?
Clark Cheng: As the industry got institutionalized with more pensions and endowments, fees became more important in the industry. Nobody wants to report the high fees paid to managers when net returns are low. The percent of profits being shared between low vol managers and limited partners becomes inequitable. However, this industry is a supply/demand driven market. If the return after fees is low, an investor can redeem at any time. The best managers with more demand and limited capacity can always charge more, especially if their net returns are higher than the average manager with lower fees. As an investor, you only get to take home the net return after fees. Higher alpha managers deserve to be paid more than less skilled managers.
There is progress now as fees have fallen for average managers. Management fees scale down as assets grow, which prevents managers from being incentivized to not take risk and collect a management fee when their assets grow. Performance fees are getting interesting with hurdle rates and scaled fees based on performance. There are also alternative fee structures such as the 1-and/or-30 which is more complicated than people think. In some cases, limited partners negotiate basis points from the management fees, but the daily volatility in performance from the manager may be greater than that annual fee reduction. Don’t lose sight of the big picture.
I think these fee structures are interesting for some of these institutions. The 2-and-20 fee structure does make a difference when the gross returns are too low. However, if you're investing in higher return funds, the fees have less impact.
What we haven't talked about is how to align interests so that the managers and investors are completely aligned. Some of these new fee structures are hitting some of the points. What's fascinating is the first-loss structures. In a market neutral strategy, the manager is collecting 50 percent of the performance fees but they’re also putting up 10 percent of the first loss. In this case, the manager is incentivized to take a lot of risks to generate returns, but also incentivized to manage risk since he suffers the first 10 percent of a drawdown. I don't know if there’s a way to apply this type of incentive structure on a non-market neutral strategy. Maybe there is with benchmarks, but it’s an interesting problem to think about.
For managers and investors that disagree on the risk in a portfolio, first-loss structures or derivatives can solve the problem. For a manager to claim a maximum loss where a portfolio cannot lose capital below a specified amount, the manager should be able to provide a put option to the investor for losses below that amount. This would work well in short vol type strategies where the manager believes they are not short skew. There should be no cost to providing the put because it will never pay out according to the manager. Marketers seem interested in this structure to raise capital, but no portfolio manager has ever agreed, meaning the maximum loss is probably higher than what the marketer is reporting.
Trusted Insight: Family offices don’t have mandates similar to endowments, foundations or pensions. Do you consider yourselves as setting the pace when it comes to new approaches to investing?
Clark Cheng: I think any time the investment process is unconstrained, it allows you to generate higher alpha. This doesn’t mean that higher alpha is achieved, but simply the opportunity for higher alpha is possible. There are multiple ways to generate that alpha in the investment process, from the asset allocation to manager selection to portfolio construction.
We can reduce manager risk through diversification but not sacrifice performance by investing in higher vol managers that are uncorrelated. Diversifying across multiple factors including time and style ensures a robust uncorrelated portfolio. A constrained investment process can stifle alpha. Focusing resources towards riskier investments is more efficient than having a checklist of requirements for all investments regardless of risk or investment size. We are always learning and improving the process to generate the highest performance possible given the risks we desire in the current market environment.
Trusted Insight: One of our goals is to help foster future investment leaders and CIOs. What's the number one lesson learned that might serve an aspiring investment leader?
Clark Cheng: I think there are multiple lessons you learn as you develop and gain experience. Most importantly, return and risk cannot be understood independent of each other. As a result, if something looks too good to be true, you have to figure out what's wrong with it. Usually, it's because you don't understand a risk that the manager is taking. There are very few cases where the managers are that good, such as Medallion. In most cases, there is just some risk factor that you don't understand such as short vol or illiquidity or asset-liability mismatches.
Someone asked a question at a conference recently about what I think about uncorrelated niche strategies. One reason why niche strategies are uncorrelated is that they are illiquid and if things don’t trade, of course, they are not correlated to the markets. Direct lending in hedge fund structures is an example of something that is not correlated to equity or credit markets. Most funds are marked at par for subscriptions and redemptions are paid out from new subscriptions. Similar to direct lending, many strategies such as cat bonds are described as niche strategies but are simply short vol. Collecting stable low vol premiums as returns generate high Sharpe ratios, but are probably short skew. Leverage will enhance the performance even more which seems reasonable with such little vol and perceived risk, but all these characteristics point to a bad ending.
The art of evaluating a manager depends on accurately predicting their future distribution of returns. Gross and net exposure or leverage and directionality combined with concentration will explain much of the performance. Factor exposures towards sectors, geography, market cap and styles i.e. momentum, growth, value, defensive, cyclical are important too. However, illiquidity, leverage and pricing risk create high Sharpe ratios but negative skew which doesn’t show up in historical returns. Short skew will eventually take out a portfolio.
You can view our full library of Q&As here.
View the full list of Top 30 Family Office Chief Investment Officers here.
Prior to joining Merrimac, he was a hedge fund manager HSBC’s Alternative Investment Group. Before that, he performed hedge fund research at Guggenheim Partners and Morgan Stanley between 2002 and 2005. He holds the Chartered Financial Analyst, Financial Risk Manager and Chartered Alternative Investment Analyst designations.
Clark Cheng was named as one of Trusted Insight’s Top 30 Family Office Chief Investment Officers. The following interview has been edited and condensed for clarity.
Trusted Insight: In what ways is Merrimac unique within the family office space?
Clark Cheng: I think we are very effective at allocating resources towards a single purpose. In general, at large institutions, you have to worry about all risks equally, especially those related to reputation and business risk which are less relevant to a business with no clients. In any investment business, as you add clients your incentives and motivations change, which causes the behavior of the investment professionals to change. The number and types of clients, in addition to how the business is paid for their services, will have an impact on how the business is managed and consequent performance. Larger businesses with risk-averse clients and low management fees in addition to no incentive fees will be managed to reduce business risk in order to protect an annuity stream. There is no incentive or motivation to take the risk for the institution or investment professional. The most successful investment organizations are able to understand and overcome these misaligned interests.
In most institutions, you have large checklists and requirements that need to be completed before multiple investment committees can read a lengthy report and then debate the merits of the investment. The problem is that you allocate limited resources inefficiently. You spend the same amount of time on a small investment as you do a large investment. You spend the same amount of resources on a high-risk investment as you do a low-risk investment. Allocating limited resources inefficiently is a problem for many large institutions. Family offices who only have one client can be very efficient and dynamic.
"We are investing in people, not securities, who change over time. We are diversified across a large number of managers with exposures to multiple factors, such as geography, sector, styles, time horizon and risk tolerance."
Risk
Risk is measured historically through the four moments of a distribution, return, variance, skewness and kurtosis. Most investors only focus on the first two moments, return and volatility, but skewness is so important in a market dislocation yet difficult to estimate before it is too late. More importantly, exactly risk is a function of multiple factors from investment risk to business risk to behavioral risk. Understanding what drives risks will be important in generating a better return relative to the desired risks. Allocating limited resources inefficiently is a problem for many large institutions. Family offices who only have one client can be very efficient and dynamic.
Relative to other family offices, we have a different approach to investing. The only free lunch in the world is diversification, which is only important during a drawdown. Volatility and correlations are beneficial to compounding capital when it works in your favor. You can take more manager risk if you have true diversification. The biggest risk in a concentrated portfolio is negative skew. An uncorrelated fund with a Sharpe ratio above 1.5 and leverage, illiquidity and pricing risk probably has negative skew and a potential for a large drawdown. This risk can be diversified away in the portfolio construction process through diversification.
Many investors believe they can manage a concentrated portfolio of high conviction managers which is easy in hindsight but difficult to do going forward. I have never seen a bad model/paper portfolio track record. As much as I disliked studying statistics in school, statistics are important in every aspect of creating a portfolio. Assuming a high decision hit rate of 70 percent, the probabilities only work if you make enough decisions in a diversified portfolio. If you run a concentrated portfolio with few decisions, the probabilities are not allowed to work for you and spurious results can drive returns. The goal is to have consistent repeatable returns through a process and spurious results don’t help.
More importantly, you and the organization have to acknowledge mistakes will be made and you need to minimize the impact of those mistakes and learn from them. If the investment process does not allow for people to make mistakes while investing, the investment professional will be motivated to minimize his or her mistakes and consequent career risk. The resulting organization will not have a performance-driven culture. If the investment team is incentivized correctly to openly discuss investment opportunities candidly, good decisions are more likely.
Team
The optimal investment team is five people. Any more or less is not efficient and five allows for a tiebreaker. Based on research, effective teams have higher collective intelligence when they have higher social and cognitive diversity, meaning they have more uncorrelated knowledge and skills to solve complex problems. Diversification across social characteristics include race, ethnicity, gender, and age and cognitive characteristics include education, functional knowledge, expertise, training, experience and abilities.
We are investing in people, not securities, who change over time. We are diversified across a large number of managers with exposures to multiple factors, such as geography, sector, styles, time horizon and risk tolerance. Investing in higher vol managers who are more concentrated or levered without negative skew will create a higher Sharpe/Calmar ratio if the managers are truly uncorrelated. Correlations, however, are not stable and only matter in a drawdown. Correlation benefits can be achieved by building a portfolio with exposure to multiple factors. Individual high Sharpe ratios are only important in a concentrated portfolio of managers because there is more risk per manager, but it is more efficient to build a Sharpe ratio through the portfolio construction process and use diversification to your benefit. Investing in the best managers with a portfolio of intended but diversified exposures is important in managing the downside risk of a portfolio.
Besides selecting the best managers, performance comes from sizing managers and allocating to the right strategies and factors at the right time. Many managers have similar crowded stocks in their portfolios, but sizing and timing the stocks in addition to managing net exposure (directionality) and gross exposure (leverage) determines individual performance. The same applies to managing a portfolio of managers. Allocating to the right strategies, factors, and styles at the right time determine performance. The skill in investing is portfolio and risk management not just picking stocks or managers.
Trusted Insight: In your last interview, you said that “data is limited” when it comes to selecting managers in the alternative business. Can you talk about discuss a time when you were able to distinguish a great manager from a good one?
Clark Cheng: Data is limited, listed references are typically good, and the marketing stories are all well narrated. This business would be very easy if everyone was honest. Historical returns are helpful in questioning how a manager trades around different market environments and drawdowns, but the goal is to predict the future distribution of returns with the data you have.
Given a managers' risk parameters and exposures, you are trying to find managers that have higher positive returns with less risk and negative skew. Looking at historical volatility can be misleading if you don’t understand the negative skew in the future distribution of returns. Leverage, pricing risk and illiquidity can create negative skew in market dislocations and you need to increase your hurdle rate or put a cost on this future risk.
"Most managers are telling you a well-rehearsed story and your job is to discern what is real and what is not. Nurturing relationships with people you trust will help you get the candid references you need to make decisions."
There is also a behavioral element to this potential risk. Some managers are risk seeking and others are risk averse so understanding the manager is important in assessing the probability of a large drawdown in a market dislocation. There is a lot of research that shows how many people become irrational and emotional during drawdowns which is not desirable.
Relationships are extremely important when it comes to reference checks with other investors and previous employers. Most managers are telling you a well-rehearsed story and your job is to discern what is real and what is not. Nurturing relationships with people you trust will help you get the candid references you need to make decisions. Likewise, managers also conduct references on investors to assess who they think would make a good partner or not.
Lastly, the experience can provide you with valuable data to help you make decisions. Understanding how funds are structured and how they operate is important to understanding new managers. For example, new managers always want to have portfolio management experience but not be responsible for losses at their prior firms. You need to understand how decisions are made at their prior firm to know whether new managers ever had decision-making abilities and portfolio management experience. Many new managers have never managed through a market dislocation having started their careers after the last financial crisis which was ten years ago.
Additionally, funds have drawdown limits and non competes. If a new manager shows a volatile month from their prior firm or is able to start a new job before the end of a typical non compete, he was probably fired from his prior firm. When there is a mutual decision for a person to leave a manager, chances are they were fired. Covering a wide breadth of managers for a long period of time helps you build relationships and understand how each fund works.
Trusted Insight: When it comes to hedge fund fee structures, do you have more leeway when working at a family office?
Clark Cheng: As the industry got institutionalized with more pensions and endowments, fees became more important in the industry. Nobody wants to report the high fees paid to managers when net returns are low. The percent of profits being shared between low vol managers and limited partners becomes inequitable. However, this industry is a supply/demand driven market. If the return after fees is low, an investor can redeem at any time. The best managers with more demand and limited capacity can always charge more, especially if their net returns are higher than the average manager with lower fees. As an investor, you only get to take home the net return after fees. Higher alpha managers deserve to be paid more than less skilled managers.
"I think any time the investment process is unconstrained, it allows you to generate higher alpha. This doesn’t mean that higher alpha is achieved, but simply the opportunity for higher alpha is possible."
There is progress now as fees have fallen for average managers. Management fees scale down as assets grow, which prevents managers from being incentivized to not take risk and collect a management fee when their assets grow. Performance fees are getting interesting with hurdle rates and scaled fees based on performance. There are also alternative fee structures such as the 1-and/or-30 which is more complicated than people think. In some cases, limited partners negotiate basis points from the management fees, but the daily volatility in performance from the manager may be greater than that annual fee reduction. Don’t lose sight of the big picture.
I think these fee structures are interesting for some of these institutions. The 2-and-20 fee structure does make a difference when the gross returns are too low. However, if you're investing in higher return funds, the fees have less impact.
What we haven't talked about is how to align interests so that the managers and investors are completely aligned. Some of these new fee structures are hitting some of the points. What's fascinating is the first-loss structures. In a market neutral strategy, the manager is collecting 50 percent of the performance fees but they’re also putting up 10 percent of the first loss. In this case, the manager is incentivized to take a lot of risks to generate returns, but also incentivized to manage risk since he suffers the first 10 percent of a drawdown. I don't know if there’s a way to apply this type of incentive structure on a non-market neutral strategy. Maybe there is with benchmarks, but it’s an interesting problem to think about.
For managers and investors that disagree on the risk in a portfolio, first-loss structures or derivatives can solve the problem. For a manager to claim a maximum loss where a portfolio cannot lose capital below a specified amount, the manager should be able to provide a put option to the investor for losses below that amount. This would work well in short vol type strategies where the manager believes they are not short skew. There should be no cost to providing the put because it will never pay out according to the manager. Marketers seem interested in this structure to raise capital, but no portfolio manager has ever agreed, meaning the maximum loss is probably higher than what the marketer is reporting.
Trusted Insight: Family offices don’t have mandates similar to endowments, foundations or pensions. Do you consider yourselves as setting the pace when it comes to new approaches to investing?
Clark Cheng: I think any time the investment process is unconstrained, it allows you to generate higher alpha. This doesn’t mean that higher alpha is achieved, but simply the opportunity for higher alpha is possible. There are multiple ways to generate that alpha in the investment process, from the asset allocation to manager selection to portfolio construction.
We can reduce manager risk through diversification but not sacrifice performance by investing in higher vol managers that are uncorrelated. Diversifying across multiple factors including time and style ensures a robust uncorrelated portfolio. A constrained investment process can stifle alpha. Focusing resources towards riskier investments is more efficient than having a checklist of requirements for all investments regardless of risk or investment size. We are always learning and improving the process to generate the highest performance possible given the risks we desire in the current market environment.
Trusted Insight: One of our goals is to help foster future investment leaders and CIOs. What's the number one lesson learned that might serve an aspiring investment leader?
Clark Cheng: I think there are multiple lessons you learn as you develop and gain experience. Most importantly, return and risk cannot be understood independent of each other. As a result, if something looks too good to be true, you have to figure out what's wrong with it. Usually, it's because you don't understand a risk that the manager is taking. There are very few cases where the managers are that good, such as Medallion. In most cases, there is just some risk factor that you don't understand such as short vol or illiquidity or asset-liability mismatches.
Someone asked a question at a conference recently about what I think about uncorrelated niche strategies. One reason why niche strategies are uncorrelated is that they are illiquid and if things don’t trade, of course, they are not correlated to the markets. Direct lending in hedge fund structures is an example of something that is not correlated to equity or credit markets. Most funds are marked at par for subscriptions and redemptions are paid out from new subscriptions. Similar to direct lending, many strategies such as cat bonds are described as niche strategies but are simply short vol. Collecting stable low vol premiums as returns generate high Sharpe ratios, but are probably short skew. Leverage will enhance the performance even more which seems reasonable with such little vol and perceived risk, but all these characteristics point to a bad ending.
The art of evaluating a manager depends on accurately predicting their future distribution of returns. Gross and net exposure or leverage and directionality combined with concentration will explain much of the performance. Factor exposures towards sectors, geography, market cap and styles i.e. momentum, growth, value, defensive, cyclical are important too. However, illiquidity, leverage and pricing risk create high Sharpe ratios but negative skew which doesn’t show up in historical returns. Short skew will eventually take out a portfolio.
You can view our full library of Q&As here.
View the full list of Top 30 Family Office Chief Investment Officers here.