The Real Cost of Software

A quick analysis

Software costs money. This cost is pretty visible, you get an invoice from your software vendor depending on the model; one off license costs, annual support agreements, annual subscriptions, upgrades etc. It’s even fairly simple to track some associated costs of your new software, like the server hardware required to run the software and the server software licensing costs. Beyond this however, it gets a little cloudy when trying to piece together all the associated costs of departmental software applications. No matter how small you think an implementation will be you will always incur costs when maintaining a platform for software.

Okay, let’s look at an example. Before we start it’s important to understand the main assumptions made so the numbers below make sense. The example is based on the following assumptions;

• The application is a departmental application requiring a live production system, a failover system and access to a test system for doing user acceptance testing with new versions etc.

• The application will be deployed on physical server hardware in a single data centre site.

• A major release upgrade will be taken and deployed every 18 months.

There are additional assumptions made on hardware costs, hardware maintenance and internal management costs. Details of these costs are given below the example.



As you can see, the above figures do not include the purchase cost of the software in the first place. These costs represent the additional costs incurred when maintaining software applications in-house. They are real costs and are very easily blurred into the day to day costs of running a business. This need not be the case – these are not costs of running a business, these are costs that have been chosen during the decision to run an application internally versus hosting with an external provider. Many of these costs can be eliminated and replaced with lower external hosting costs simply by switching to a hosted platform rather than an in-house system.
On this example above, switching to a StatPro hosted service would save over 50% on the annual costs of maintaining an in-house system.

Notes on costs

1. Based on the annual lease cost of three physical 2 CPU quad core servers with local memory and disk storage.
2. Based on 20 hours required to procure, prepare and provision each server with a labour cost of £28.23 per hour.
3. Based on the cost of server administration management time for tasks including change management, problem and incident management, monitoring, performance and availability management, asset management, security management, patch and upgrade management, backup and recovery, restores, storage management, disaster planning, compliance management/reporting, vendor and contracts management and financial/budget management. Based upon 3 server workloads consuming 8% of a server administrator’s overall workload (average of 40 workloads managed per administrator staff) over the year with the burdened labour cost per server administrator being £54,578.
4. Based on the annual cost of backing up 30GB of data to an off-site facility at a cost of £8 per GB per month. Includes 10% data growth per year.
5. Based on the annual cost of software assurance licenses for MS Windows, MS SQL Server, MS Office and 5 user access licenses.
6. Based on each server requiring 516 Watts of operating power and 645 Watts of cooling power at £0.0646 per kWatt hour.
7. Based on one server rack in a data centre facility consuming 23 square foot at an average annual cost of £205.05 per square foot.
8. Based on an average of 14 days a year consulting from the application vendor during application upgrades and maintenance releases at a cost of £1,350 per day.
9. Based on an average of 6 days a year of internal IT project management with a burdened labour cost per project manager of £500 per day.
10. Based on an average of 13 days a year of internal business unit project management with a burdened labour cost per project manager of £500 per day.
11. Based on an average of 5 days a year of internal IT application support requirements with a burdened labour cost per application support resource of £500 per day.
12. Based on an average of 3 days a year of internal IT training time for new/replacement application support resources with a burdened labour cost of £500 per day.
    

Asset Managers need to reduce their costs – SaaS will help

Why StatPro is evolving its business model

Time changes everything. What was once state of the art becomes a cumbersome nuisance at some point. The explosion of activity that occurred in the 90’s when hundreds of small software companies built applications to sell to asset managers allowed a big increase in efficiency. Processes that had taken weeks were reduced to hours or minutes. New processes that were impossible dreams became a reality. This was all possible because of windows and Visual Basic and the process was not unique to asset managers but affected every business sector. By comparison the Dot Com boom did not reach the business community but was aimed at the retail market and when it busted prematurely, development slowed.

So today the IT landscape of the Asset management industry is one where any company of any size has 10 to 20 important systems upon which they depend for a variety of essential services. The majority of these systems are all provided as software that is installed on the company’s own servers in their own IT centre. They will probably have many other applications of less importance that they run as well. The result is that they have large IT teams, complex processes and relatively high costs.

Some companies have tried to outsource the whole thing to a third party such as a custodian or IT services provider. However, there are many horror stories that warn this is not a simple solution. In fact the rule of thumb is that if your IT department works well and efficiently, you can outsource it, but if it is inefficient you can’t. The outsourcer cannot be expected to unravel your mess.

What has worked however is selective outsourcing. This typically means getting the supplier of your system to host their own application. This makes sense as the supplier ought to know how to support their own product and if they do this for all their clients, they can get economies of scale which a generalist outsourcer cannot hope to achieve.

The other factor is that Web 2.0 has unleashed a new revolution in IT making it possible for suppliers to provide “Software as a Service” (SaaS) solutions over the Internet and with relatively low cost. Incumbent suppliers are reluctant to adopt this new technology as they fear cannibalisation of their existing business and because they calculate that such is the complexity of the processes that surround their embedded systems, clients will be loathed to move anyway. That leaves the door open for start-ups, but this generation have a much bigger battle than the class of ’95 as the standards and functionality they have to match are significantly higher and the gains in efficiency are less about time and volume and more about money. Clients want the same for less rather than more for the same.

This is the thinking behind StatPro’s own strategy. We believe that eventually all our clients will want to access their services over the web just in the way that we all use email rather than a fax (or telex, or telegram). Indeed, email is fast being superseded by instant messaging. However, we also recognise that clients cannot simply throw everything out and start again, but rather need to evolve in a sensible direction according to a plan that has been tried and tested. The last thing anyone wants to do is jeopardise their business for the sake of saving a relatively small amount of money.

This means that we have focused on deploying new services and products that offer quick savings for clients whilst moving them towards the strategic objective of low-cost web-based applications.

StatPro has built a solid reputation as a supplier of Data for valuations and Analytics systems that cover Performance, Attribution, Risk, Compliance, GIPS and Reporting. Right now, most of StatPro’s clients deploy our software on their own servers. They feed the software with index data and market data from third parties and finally with their own data. The result is that the cost of the software we supply (on a subscription basis) amounts to less than 10% of their total annual cost of ownership in most cases. IT costs are about 10%, data supply will be about 30% and 50% employee costs although some of those employee costs relate to managing IT and data.

We are in a position to offer our clients the IT platform and the third party data and we believe that by doing so we can reduce the total cost of ownership to 70% or less of what it is despite charging additional fees for IT and data. This is because we can offer the IT platform together with the index and market data at a very low rate.

Apart from saving our clients money, we will also improve their service as we will have direct access to their system in the event there is a problem. We will also save our clients further money because we will not have to charge them to upgrade their software. The clients will receive upgrades when we are ready, not when their IT departments are ready. Overall, the complexity of process and the management time required to follow it all will be slashed so there will be other hidden savings and benefits. If a client buys more products from us to replace legacy software we will offer further discounts and savings.

It is also important to note that it will be easy for the clients to take this new service as it is a simple matter of us putting their existing system on to our platform. From there we will upgrade them progressively and, from their point of view, painlessly.

The second stage in our strategy is what we call SaaS2 which is a SaaS product built from scratch. This product takes full advantage of all the latest technology as well as our accumulated knowledge about our products. Once the platform is launched however, it will be relatively easy for us to migrate clients to it from the SaaS1 platform if they wish to use it. This new platform will be released in beta form during 2009 and go live in 2010.

Many asset managers must be thinking about SaaS as a way to reduce their costs, but the obstacle they will encounter is the complexity of getting there and doing so for a wide number of applications. We believe that our approach offers a strategic shift with clear and immediate savings whilst laying the groundwork for the final move to a pure SaaS platform which in turn will allow greater savings.

Justin Wheatley
Chief Executive StatPro Group plc

Mark to Model or Mark to Myth?

Why valuations of credit instruments based on Mark to Model might be a really good idea.

Asset managers and Custodians have a real problem valuing illiquid assets in their portfolios even though many of these assets are perfectly sound and the asset manager has no intention of selling them. Assets should be valued at mark to market prices as required by the Basle rules. However mark to market prices should not be used in isolation in my view, but rather compared to model prices to test their validity. Models should be improved to take into account the greater amount of market data available.

No less a person than Warren Buffet decried the traders of Wall Street for “Marking to Myth” their assets as they sold untold trillions of them to the ill-informed. The traders were using models to price these assets but miraculously the models always seemed to offer high valuations. This has given the process of Marking to Model a bad rap. What Warren Buffet was really criticising was the bubble in valuations caused by the systematic under-pricing of risk. In the same way that analysts during the Dot Com bubble extrapolated forecast profits into the multi-billions for companies that had not yet $1 of revenue and then sold over-priced junk stock to unwary investors, so too did Wall Street’s finest with credit instruments.

The credit crunch is now 19 months old and the fundamental reason why we had one in the first place is due to the mis-pricing of risk and therefore of assets. With cheap and plentiful supplies of money, competition drove down the prices at which banks would do deals and drove up the prices that investors were willing to pay. A lethal combination of faulty models and bonus-driven salesmen did the rest. When at last people started calling into question the values of these accumulated assets (not just ones based on falling property prices but also on optimistic future profits and continuing economic fair weather) the giant game of financial musical chairs began as everyone searched desperately for any remaining buyers in the market to off-load their “investments”.

With the new Basle rules about marking to market this has had a domino effect on a prodigious scale. Assets that otherwise would have been held to term have had to be dumped as prices have been marked down to the most recent market prices. As corporations and banks have been unable to get any credit, the lack of liquidity has turned to potential (and actual) insolvency and that in turn has triggered a mass of downgrades of credit ratings. This once again has caused many investors to sell off downgraded assets as their investment mandates forbid them to hold any bonds with less than AA ratings.

It stands to reason that on the way up prices that were marked to market were affected by over supply of buyers and on the way down the mark to market prices are affected by a massive under supply of buyers. However, just because there are no buyers for your asset doesn’t mean that the asset is worthless if (and this is a BIG if) you are not a forced seller. The core issue is whether you need to sell your asset to raise cash. If you are not in a rush, over time the market price will recover. Many fixed income assets are bought and held to term (with no leverage). Provided the issuer does not default on interest or capital, the asset has the value of its cash flows and principle. So if you have such assets and a long term investment horizon and no liquidity issues, how should you value these assets? What happens if you are put into the perverse position where you have to mark the value of your assets down to a level where your terms of reference force you to sell them? Mark to market is good for transparency, but where there is no liquidity does this approach still work?

Look at value another way, if you wanted to sell your house that you bought for £1.0 million 18 months ago, you would probably go to an agent and get them to estimate what the going rate was for similar houses. Let’s say they gave a range of £650K to £750K you would then “model” your house price on this and put it on the market for £790K hoping to get £750K. If someone offers you £50K for the house you would be unlikely to sell it to them, however their offer would represent the only “market” price for the specific asset: your house. For better or worse, you would stay put and wait for better market conditions and in the meantime you would probably make a mental note that your house was probably worth £650K despite it only being a model valuation rather than £50K (the last real offer).

In other words there is a problem with mark to market when there is no liquidity. By definition a liquid market is where consenting parties agree on a price at which they will deal with each other freely. It is not surprising that rather graphic terms are used for buyers in the current illiquid markets such as sharks, rapists, bottom feeders and vultures as they are only interested in buying from the truly desperate. In the housing market the forced sellers are the people who have defaulted on their mortgages, are divorcing or have inherited a property and want to get cash quickly. These are the asset sales that at the margin define the market price. Until this overhang of supply works its way through the system there will be no motivation for buyers to pay more.

I am not arguing that one should ignore the market but rather that blind faith in it does not make sense and that the law of unintended consequences can play havoc as a result. Rules that enforce one approach can have perverse results. An example is with Index Funds. When a new stock enters an Index, they all buy it, propelling the price of the stock higher. When a stock exits the index, they all dump it, causing the price to fall further. What is needed is a circuit breaker or alternative way of thinking to validate the orthodox approach. Indeed the accountants that set the rules for the International Accounting Standards Board are also concerned about “Fair Value” over a market price. To them Fair Value is the amount at which an asset could be exchanged in a current arm’s length transaction between willing parties where each acted knowledgeably, prudently and without compulsion. If that cannot be obtained from quoted prices, then they look to proxies and finally to model prices.

When Warren Buffet buys an asset, he uses his own model for determining whether the asset is good value or not. Sometimes he will pay more than others for a given asset. He works out the price he will pay and then he sticks to it. This is a good example of making the market come to the model price. Buffet’s model is based on a fundamental approach to valuations that stays consistent irrespective of “Mr Market” (as he likes to call it) and he maintains the discipline of sticking to his model for valuation even if he has to wait several years to buy.

The problem faced by asset managers and the custodians of their assets is what approach to take vis-à-vis valuing so many illiquid assets where there are seldom any trades? There are some 4 million or so bonds in issue not counting CDOs, CDSs and other derivative assets and 95% have no regular market made in them. Even where a broker price can be obtained these are often skewed and reflect the trading book of the broker. It makes sense then for custodians and asset mangers to use multiple sources to gauge more accurate valuations of their assets including model prices. Market prices in any case are based on the models used by the investment banks so it is good to have an alternative model to validate them.

However, not all model prices are equal and custodians and asset managers should look for more sophisticated models. For example, historically, credit ratings provided the only means of gauging the risk of default, but now the CDS market for a given issuer is a much more up-to-the-minute guide on the likelihood of default. Further one can build up the price of an asset by looking at the various spreads over LIBOR including the sector spread by using highly liquid indices like iBoxx and iTraxx. Thus the price of an asset can be modelled from LIBOR, its credit spread, its CDS spread and its sector spread and whatever other factor based on its term sheet. Such a price will be very robust and a sound basis to challenge the “market” price from an investment bank.

In summary, a valuation is not the same as a transaction and at times of stress dogged application of rules can cause more harm than good. There is no harm in challenging “market” prices if they seem skewed and the best way to do that is to have alternative opinions in the form of independent model prices. New methods and new data are available to help improve models and these should be used. In the end all prices start off from a model.

Justin Wheatley is Chief Executive of StatPro Group a leading supplier of portfolio data, pricing and analytics.

Why we need to create the role of the Super Employee

How the current employment law protects CEOs from losses


Credit crunch fatigue is probably setting in for many people, but I am sure that even they have some venom left for those individuals who presided over the whole debacle and yet walked away with millions of dollars and pounds. How was it possible that we “allowed” this to happen? Why is it that there is not a regulation that stops all this?

Well the short answer is that there are lots of regulations, but most of these are focused on stopping an organisation from doing things it should not. The officers of these organisations can be held personally responsible if it can be proved that they deliberately endangered the business entrusted to them or behaved fraudulently. The problem is that 99% of these bosses were just incompetent and when they are hauled up in court by angry shareholders, the judge is going to say that whilst the fellow was clearly a fool that is not grounds for suing him and as an employee, he has rights under statute that mean the company have to pay him off. Why is this?

Well employment legislation is (quite rightly) heavily weighted toward the employee. The employee is seen as subject to the power of the employer and so in a vulnerable position. If an employer imposes a clause in a contract with the employee which the judge finds unreasonable, it can be struck out on the grounds that the employer must have bullied or fooled the employee into accepting it. The judge effectively acts as a retrospective lawyer for the employee. The reverse is not true of course.

These days we have a situation where individuals who have not created a giant public enterprise are entrusted to run it for a short period of 3 to 5 years. They may well have fought hard to achieve this position of tremendous responsibility and they will certainly have a detailed employment contract put in place setting out terms they get and they probably feel they have a relatively short time to maximise their gains and establish their reputation before they move on. They will not think like the entrepreneur who started the original business. They will not see the institution which they work for as all important but rather think that they are important because they run the institution.

People have long recognised this problem and thought that loading an executive with options would motivate him or her properly by aligning interests with other shareholders. However this method of incentive has been proven to not have succeeded if the objective was to increase shareholder value. Indeed, I suggest that shareholders, remuneration committees and others can make up new schemes as much as they like but they will not solve this problem. The hired hand does not think like the person that built the business. He or she is either good at their job or not. The individual either behaves responsibly and takes long terms matters into consideration or not. In fact, some schemes provide perverse incentives to do the wrong thing. For example an acquisition might not make complete sense, but if a £10 million bonus is on offer for doing it and £0 for not doing it, it would take a person of extreme integrity to call it off.

Some people might argue that levels of pay should be capped or taxed out of existence. I say that this makes no sense. If someone succeeds, they should be rewarded, the issue is people being rewarded for failing and that is where employment law steps in.

Companies can try to impose a claw back of bonuses and pay in the event of failure, but, apart from the problem that they will probably not get many applicants for the job, their lawyers would probably advise them that it may not be enforceable to claw back pay that has already been paid, rendering the contract less effective. This is because under current employment law it would be seen as unreasonable to take back money paid in return for services already rendered even if it was by mistake.

If however, the law created a new definition of employee, let’s call it a Super Employee, this issue could be dealt with much more easily. The Super Employee would be seen as an equal to the employer not subordinate. The Super Employee would indemnify the Employer to a level that the two parties would freely negotiate with each other in the same way that a supplier may indemnify a client. I would suggest that this amount would be limited to say 80% or 90% of the money received under the contract (over several years potentially). Importantly the Super Employee would be considered to be in breach of the contract if he or she was proven to be incompetent. So in the same way that a supplier provided you with faulty equipment you can demand your money back, so too will the shareholders be able to reclaim their money if the Super Employee does not provide the services they said they would.

Whilst it is clear that the CEO of a large enterprise falls into this category, the concept could be used with highly paid employees as well. All these multi-million dollar bonuses for bankers were thoroughly deserved no doubt, but what if some people actually lost their bank lots of money having earned millions in bonuses in previous years? I think that a company might put in place a policy which said that if an employee wants to earn more than say £250,000 they have to agree Super Employee status. That way the shocking news that John Thain signed off $4 billion of bonuses for Merrill Lynch employees in a year where Merrill Lynch lost $28 billion would just not happen. Right now it is entirely rational behaviour for the individuals concerned as even if the institution loses the money the individual gains a reward irrespective of success. The worst sanction they can get is to be given a large sum of money to leave. Indeed, only in the nuclear situation where the business is bankrupt will the employees potentially lose out, but let us bear in mind those long suffering shareholders who are wiped out by bankruptcy.

The concept of Super Employee should also spread to fund managers, because of course the wiped out shareholders are not the individuals managing the money on behalf of millions of small investors in pensions, insurance funds and unit trusts. The fund managers make the decisions to invest or not. They get the big bonus if things go well, it is your pension that gets it in the neck if they fail and they just move on. Because they want to be paid large bonuses they put pressure on the CEOs to make bigger returns, which generally requires the CEOs taking on more risk. Because the CEO is not around for long, he or she happily complies with this. This suits the investment banks that make their real money from juicy transaction fees. The result is a focus on personal gain through visible action (such as an acquisition) rather than steady improvement.

All these institutions feed off each other and when something goes wrong with one it affects the others. However, the individuals who run them are not affected financially and the incentives put in place to “motivate” them do not encourage the right behaviour in all cases. There will always be failures, but the law inadvertently over-protects the people with the power to make a difference. The solution is to change the balance of legal power to a more balanced position where enforcing a rebate is relatively easy provided loss can be proven.

Once that is done, hopefully the danger of perverse incentives encouraging CEOs (and others such as traders) to take foolish risks will diminish. I don’t think that it will stop risk taking but rather align interests strongly. After all there can still be great rewards for success. Nor will it stop CEOs trying to wriggle out of their commitments as it will still be hard to prove that it was the CEO’s fault, but the key element is that if they are to blame it would be possible to get the money back. This is currently impossible unless a crime has been committed. Most of all, it would leave the problem where it deserves to be left – with the shareholders. It would be their decision to sue or not and their decision to give a CEO an easy contract or not, their decision to hold the CEO accountable if they wish.

Summary
The crisis involves the destruction of institutions by the individuals who run them
The individuals who run them have employment law on their side
The individuals can earn large sums by taking risks with the institution whilst experiencing no risk themselves
All the institutions are interlinked so destruction of one can lead to the destruction of the others
The public interest is therefore at stake and yet there is no accountability
Super Employees will lose the protection of normal employment status making it much easier to claw back remuneration using existing laws
Super Employees will only take well calculated risks if they too can lose all they have made

Justin Wheatley

How sharp is the Sharpe ratio?

Any discussion on risk-adjusted performance measures must start with the grandfather of all risk measures the Sharpe Ratio or Reward to Variability which divides the excess return of a portfolio in excess of the risk free rate by its standard deviation or variability.

Most risk measures are best described graphically, a measure of return in the vertical axis and a measure of risk in the horizontal axis as shown below.

Ideally if investors are risk averse they should be looking for high return and low variability of return, in other words in the top left-hand quadrant of the graph. The Sharpe ratio simply measures the gradient of the line from the risk free rate (the natural starting point for any investor) to the combined return and risk of each portfolio, the steeper the gradient, the higher the Sharpe ratio the better the combined performance of risk and return.

Funds are ranked in order of preference with the Sharpe ratio but it is difficult to judge the extent of relative performance. M2; first proposed by Leah Modigliani and her grandfather Professor Franco Modigliani (1997) offers an alternative risk-adjusted return using the Sharpe ratio of the portfolio but calculated at the risk of the benchmark thus allowing direct comparison.

Investment statistics can either be grouped as Sharpe type combining risk and return in a ratio, risk adjusted returns such as M2 or descriptive statistics which are neither good nor bad but provide information about the pattern of returns of the portfolio manager. The first moment of a return series is the mean, the second moment is the variance, the third moment is skewness and the fourth moment kurtosis. Kurtosis measures the weight of returns in the tails or the peakedness of a return distribution. Investors should prefer high average returns, lower variance or standard deviation, positive skewness and lower kurtosis. The adjusted Sharpe ratio suggested by Pezier (2006) explicitly rewards positive skewness and low kurtosis (below 3, the kurtosis of a normal distribution) in its calculation and thus potentially removes one of the possible criticisms of the Sharpe ratio.

The regression statistics b (or systematic risk), r (correlation) and R2 are descriptive statistics. Jensen’s alpha is often misquoted as the portfolio manager’s excess return above the benchmark, more accurately it the excess return adjusted for systematic risk.

Treynor ratio or Reward to Volatility is similar to Sharpe ratio, the numerator (or vertical axis graphically speaking) is identical but in the denominator (horizontal axis) instead of total risk we have systematic risk or volatility as calculated by beta. Although well known the Treynor ratio is less useful precisely because it ignores specific risk.

The appraisal ratio first suggested by Treynor & Black (1973) is similar in concept to the Sharpe ratio but using Jensen’s alpha, excess return adjusted for systematic risk in the numerator, divided by specific risk not total risk in the denominator.This measures the systematic risk adjusted reward for each unit of specific risk taken.

In the same way that absolute return and absolute risk are combined in Sharpe ratio excess return and tracking error (the standard deviation of excess return) are combined in the information ratio, although given the need of an appropriate benchmark less useful for hedge funds.

The Sharpe, appraisal, Treynor and information ratios are familiar measures used by the industry for decades. More recently hedge funds have encouraged the use of further risk measures designed to accommodate the risk concerns of different types of investors. These measures can be categorised as based on normal measures of risk, regression, higher or lower partial moments, drawdown or value at risk (VaR).


Predominately hedge fund management styles are designed to be asymmetric in their return patterns. If successful this leads to variability of returns on the upside but not on the downside. Investors are less concerned with variability on the upside but of course are extremely concerned about variability on the downside. This leads to an extended family of risk-adjusted measures reflecting the downside risk tolerances of investors seeking absolute not relative returns.

Standard deviation and the symmetrical normal distribution are the foundations of Modern Portfolio Theory. Post-modern Portfolio Theory recognises that investors prefer upside risk rather than downside risk and utilises semi-standard deviation.

Downside risk measures the variability of underperformance below a minimum target rate. The minimum target rate could be the risk free rate, the benchmark or any other fixed threshold required by the client. All positive returns are included as zero in the calculation of downside risk or semi-standard deviation.Downside potential is simply the average of returns below target, upside potential the average of returns above target.

In their article “A Universal Performance Measure” (2002) Shadwick & Keating suggest a gain-loss ratio, Omega (W) that captures the information in the higher moments of a return distribution implicitly adjusting for both skewness and kurtosis; dividing upside potential by downside potential.

A natural extension of the Sharpe and Omega is suggested by Sortino (1991) which uses downside risk in the denominator. Total risk has simply been replaced by downside risk, portfolio managers will not be penalised for upside variability but will be penalised for variability below the minimum target return.

The upside potential ratio suggested by Sortino, Van de Meer & Platinga (1999) can also be used to rank portfolio performance and combines upside potential with downside risk. Even “Prospect Theory” the fact that investors dislike losses far greater than they like gains can be built into a Sharpe like measure in the form of the Prospect ratio.

If value at risk is your preferred measure of risk then, of course, there is a Sharpe type measure that replaces standard deviation with VaR in the denominator; called reward to VaR. VaR does not provide any information about the shape of the tail or the expected size of loss beyond the confidence level. In this sense it is a very unsatisfactory risk measure; of more interest is conditional VaR otherwise know as expected shortfall, mean expected loss, tail VaR or tail loss which takes into account the shape of the tail. Historical simulation methods which make no assumptions of normality are particularly suitable for calculating conditional VaR. The conditional Sharpe ratio replaces VaR with conditional VaR.

Perhaps the simplest measure of risk in a return series from an absolute return investor’s perspective, wishing to avoid losses, is any continuous losing return period or drawdown. The average drawdown is the average continuous negative return over an investment period, three years being a typical period of measurement for comparison purposes.

The maximum drawdown not to be confused with the largest individual drawdown is the maximum potential loss over a specific time period, typically three years. Maximum drawdown represents the maximum loss an investor can suffer in the fund buying at the highest point and selling at lowest.The Calmar ratio is a Sharpe type measure that uses maximum drawdown rather than standard deviation to reflect the investor’s risk. In the context of hedge fund performance it is easy to understand why investor’s might prefer the maximum possible loss from peak to valley as an appropriate measure of risk.

The Sterling ratio replaces the maximum drawdown in the Calmar ratio with the average largest drawdowns.

Similar measures including the Pain ratio and the Ulcer Performance ratio incorporate the duration and depth of drawdowns since the previous high water mark. The range of combined risk and return measures available for hedge fund investors is almost limitless.

With so many similar ratios the natural question to ask is “which is the best measure to use?” In fact Eling & Schuhmacher (2006) have published an article “Does the Choice of Performance Measure Influence the Evaluation of Hedge Funds” which concludes that most of these measures are all highly correlated and do not lead to significantly different rankings. Both the question and their article to some degree miss the point, risk like beauty is in the eye of the beholder, the investor most decide ex-ante which measures of return and risk best reflect their preferences and choose the combined ratio which reflects those preferences. One, and only one, of the above ratios are most likely to reflect the preferences of the investor. Care should also be taken to ensure hedge funds are not hiding volatility by using smoothed valuations. Consistent valuation criteria must by applied each month, although Global Investment Performance Standards (GIPS) do not require that specific risk measures are used they do require documented policies and procedures for valuations consistently applied and are therefore valuable and a source of comfort for any potential investor.