The Accidental Trader User Manual
The Accidental Trader User Manual
http://atg.creditspectrum.net
INTRODUCTION. ....................................................................................................................................... 3 MAIN ELEMENTS OF THE GAME......................................................................................................... 4 PORTFOLIO.................................................................................................................................................. 4 EXPOSURE................................................................................................................................................... 4 Ratings. .................................................................................................................................................. 5 Trading. ................................................................................................................................................. 5 Buying and Selling Protection. .............................................................................................................. 5 CAPITAL...................................................................................................................................................... 7 Long Term Capital................................................................................................................................. 7 Short Term Capital. ............................................................................................................................... 7 Renting Capital. ..................................................................................................................................... 7 USER INTERFACE. .................................................................................................................................... 8 MAIN SCREEN. ...........................................................................................................................................10 Game Information.................................................................................................................................12 Portfolio Parameters. ...........................................................................................................................13 Expense Meter. .....................................................................................................................................14 Portfolio. ...............................................................................................................................................14 Control Buttons.....................................................................................................................................15 CAPITAL MANAGEMENT. ...........................................................................................................................16 Interest Rates. .......................................................................................................................................16 Renting From The Market.....................................................................................................................17 Renting From Other Teams. .................................................................................................................17 SCALE. .......................................................................................................................................................18 PORTFOLIO PROGRESS. ..............................................................................................................................19 Average Portfolio Rating X(t). ..............................................................................................................19 Probability Of Default. .........................................................................................................................19 VaR. ......................................................................................................................................................19 Total Cost Of Capital............................................................................................................................20 SKYPE. .......................................................................................................................................................21 WORLD EVENTS. ........................................................................................................................................21 LAST ACTIONS. ..........................................................................................................................................21 READY BUTTON AND TRAFFIC LIGHT. .......................................................................................................22 BASIC GAME STRATEGIES. ..................................................................................................................23 HOW THE SCORING WORKS. ......................................................................................................................23 MANAGING PROBABILITY OF DEFAULT.....................................................................................................25 THE COST OF CAPITAL. .............................................................................................................................25 TRADING EXPOSURES. ...............................................................................................................................26 BUYING PROTECTION RISK ELIMINATION. ..............................................................................................26 SELLING PROTECTION TAKING MORE RISK TO REDUCE THE COST OF CAPITAL. ...................................28 RENTING CAPITAL FROM OTHER TEAMS. ..................................................................................................28
Introduction
The Accidental Trader Game is an educational tool designed to simulate the real-life experience of a risk officer or trader managing the credit risks of a large portfolio of corporate bond or loan exposures. Delegates have an opportunity to manage a large portfolio undergoing multi-period credit challenges by adopting various credit risk management strategies to keep the portfolio probability of default [Pd] at a pre-determined value, which the game administrator has set using a corresponding Value at Risk (VaR) level. In addition, the teams portfolio default probability must not exceed a given, critical level no matter what else is going on. If this level is breached, the game considers that the manager, or team of managers, has lost the ability to control the portfolio. That outcome corresponds in practice to economic loss for the firm and probable job loss for the manager. The game ends after a certain number of time periods, usually twelve, or when each team but the last - the winning team - is out of business. If more than one team survives, the team with the best Pd record is declared the winner [statistic factor], while other teams are ranked by their longevity in the game [stochastic factor]. At the beginning of the game, each team is provided with an initial portfolio of risky exposures. All exposures are assigned a credit rating reflecting its ex-ante probability of default. Associated with each risky exposure, is a set of financial characteristics commonly used as input variables in credit risk modeling and management, such as the recovery rate, loan-equivalent exposure and credit-default swap [CDS] spread. Those characteristics change with the passage of time and trigger various credit events such as upgrades, downgrades and defaults, which are reported at the beginning of each time step or round. So are key portfolio-level parametric averages: portfolio mean loss, portfolio loss standard deviation, and VaR to help managers in forming risk judgments and making tactical portfolio adjustments. At the beginning of each time step, teams borrow short-term funds from the market or one another in order to keep the portfolio probability of default at the target level. The game is not designed to create a portfolio whose value will increase secularly over time, the way a trader might view the credit problem. This is strictly about credit risk management, not credit risk speculation. Although the up side is always much more attractive in the stock market for instance, credit risk professionals are much more focused on the down side, hence the essence of the game. Teams can improve their performance through various risk-management tactics. They can swap out of risky exposures for anti-correlated equivalents to reduce portfolio sensitivity to extreme events, or buy or sell CDS contracts on their portfolio exposures. By doing so, teams can realize gains or losses as well as increase or decrease their exposure to the market.
Exposure
Even though a single exposure does not reflect much of the characteristics of a large portfolio, it is still useful to know information about individual risks. Usually, a risk manager will be interested in monitoring the most risky components of the portfolio. Therefore, the game gives participants the ability to review individual exposures and take actions, like trading and buying or selling protection. This is how a manager can bring the whole portfolio to a desired risk level. The game provides the following parameters, displayed on the exposure screen, portfolio table or both: Name. Every exposure is assigned a unique identifier and a name. The exposures name doesnt carry any useful information and serves only to identify single loans in a more realistic context. Thus, the fact that an exposure is called Microsoft does not mean that it should have a high rating and therefore, a low probability of default. Rating. A Moodys letter-grade rating is assigned to each exposure. The rating is based on the probability of default. The lower the probability of default for a certain exposure, the higher the credit rating. Ratings can change from one time step to another as a result of natural exposure evolution, as well as from random world events. Participant can observe current and previous ratings at any given time step. Probability of default. One of the most important exposure characteristics. It is used to compute the likelihood that an exposure will default, its credit rating and its CDS spread. It also changes naturally as time passes based on the behavior of an arbitrary driving function y(t) and sometimes randomly jumps up or down as a result of world events generated by the system [a Levi-jump process]. Historical data describing probability of default x(t) and the driving function y(t) are displayed on the exposure screen as well. Loan Equivalent Exposure (LEE). This variable shows the dollar amount invested in each exposure. For the sake of simplicity, LEE is assumed to be constant
across all portfolio exposures. The LEE amount influences the loss/profit profile in case of credit events, as well as CDS premia received or paid. Loss Given Default (LGD). The percentage of LEE which will be lost if the corresponding exposure defaults. LGD is used to compute expected losses, VaR, and the consequence to the holder of the exposure, or CDS seller, when default occurs. CDS Spread. Annual premium received or paid for protection on any given exposure. The more risky the exposure, the higher its CDS spread. The spread changes with time depending on the exposures system-computed probability of default to reflect market conditions. For the sake of simplicity, we assume that the bid/ask spread is zero.
As a result of a credit event or a natural evolution every exposure can default. Defaulted exposures stay in the original holders portfolio but are no longer used to compute risk characteristics. It is also not possible to trade an exposure after it has defaulted, buy and sell protection on it or do any other operation available with respect to the non-defaulted exposures. The loss in the event of default is added to the Total Realized Losses account and displayed in real time on the teams main screen.
Ratings
Letter-grade ratings are directly connected to the exposures probability of default, which itself varies naturally with time. The rating table can be accessed from the main screen (more on this in the user interface section below). Ratings are presented for informational purposes only. All computations use actual default probabilities. However, letter-grade ratings are easier to monitor and interpret for most market participants.
Trading
Every exposure in the portfolio can be traded. Every month a certain number of possible trading candidates will be available to all teams. These choices consist of exposures anti-correlated to the portfolio as a whole. There is a small transaction cost involved in trading, but in general it does benefit the risk manager to replace a highly risky exposure with an anti-correlated one because, as a result of this transaction, VaR will decrease. As a result, the amount of required risk capital will be lower. Consequently, less shortterm capital will be needed and lower interest payments will be made to those who provide such capital, thus decreasing the total cost of risk management. If a manager has bought or sold protection on the traded exposure, the said CDS contract will be automatically terminated upon trade execution.
fully protect an exposure, a notional amount equal to LGD should be purchased. On the other hand, any amount of CDS notional can be purchased. If the exposure defaults, the protection buyer will receive the corresponding LEE amount and the CDS contract will be terminated. Protection can also be sold on the open market. However, the amount of protection which can be sold is limited by market interest. As a result of selling protection via a CDS contract, the seller will collect monthly premiums but will have higher expected losses, which will increase the selling teams capital requirements. If a credit event occurs and the target exposure defaults, the seller will have to pay the amount corresponding to the notional amount of the CDS contract. At that moment, the seller collects the last premium payment and the CDS contract is automatically terminated. Participants can terminate any CDS contract by entering into an offsetting contract with identical terms. In that case, a termination value will be computed as of the current time step, resulting in a profit or a loss on the position.
Renting Capital
At the beginning of each session, participants will decide how much Long Term capital to secure. After that, only Short Term capital will be rented. There are two ways to rent it. First, any amount of STC can be rented directly from the market at a given, floating interest rate. Second, it can be rented from other participants. In that case, the interest rate will be lower than the market rate. However, the renting team will have no effective control over how much capital is available for rent. The system will determine the amount of excess capital, if any, available to the lending team and then will transfer this excess amount to the renting teams account. To enter into rental agreements, teams can communicate with each other using their Skype account. Each team needs a Skype account and a pair of earphones.
User Interface
How to Join the Game
In order to join the game, you will need to have a password, enabling you to access the game server. The password will usually be provided by the game administrator. After logging into the server, teams will be able to register their name with the system. Below are detailed instructions on the registration step. Each name needs to select a unique team-identifier.
Server Login
Now, you will need a log-in name and a password to login to the server. This information will be provided to you by the game administrator. This process will let you log in to the server not to the particular game. Usually, all participants will have the same name and password at this step. After logging into the server, each team will have an ability to choose which game to play, register itself, continue playing a previous game, etc.
The maximum total capital available at the outset is equal to your portfolios current VaR. This amount should be divided between Long Term- and Short Term capital. Remember that the amount of Long Term Capital you choose remains constant during the session. As a general rule of thumb, it is reasonable to dedicate about 80% of total capital to Long Term capital and to rent the rest as the game proceeds. However, depending on a strategy a team chooses, this could be an under- or overestimate. After your capital structure has been set, press the Confirm and Close button. The system will wait until all teams have made their decision about the initial capital structure before proceeding to the next time step. Therefore, setting up the initial capital structure is the only required task during the first month of any session.
10
Main Screen
Now, lets look at the main screen which will be used by playing teams to obtain current information and make their decisions. Before we proceed to the description of information and controls however, we need to say a few words about the sub-section of the main screen called World Events.
World Events
At the beginning of each time step, all teams will see this type of screen:
It informs players about what happened in the games world during the past month. The top part of the screen shows upgrades and downgrades. Upgrades are indicated by green, down-arrow symbols. The direction of the arrow means that the probability of default went down, hence the green (improvement) color. On the other hand, when an exposures default probability went up, a red (deterioration) arrow is displayed. Participants will also be able to see the name of each upgraded/downgraded exposure and how its letter-grade credit rating has changed. The screen section beneath upgrades and downgrades shows defaults that occurred during previous time steps. Defaults usually dont happen often; hence, it is very likely for no default to have occurred during any given month. However, there can be time steps where a few defaults will occur in a single month. Once a player presses the Confirm and Close button, this informational screen disappears and the main control screen becomes fully functional.
11
On the right of the table, a graph labeled Expense Meter is displayed. It shows the comparative performance of all teams playing a given game. As is implied by the graphs label, the lower a team is positioned on the graph, the better it is currently performing. Players can see all historical changes from the beginning of the session (time 0) until the current time step. The large section underneath the graph and the parametric table shows the teams current portfolio. Participants will see all exposures in the portfolio as well as their main characteristics (see above). To obtain more detailed information about particular exposures, you need to click on its name. For a description of the exposure screen, please see the corresponding section below. Finally, at the bottom of the main screen, a series of control buttons are displayed. They all have different informational or operational purposes and are described in details below.
12
Game Information
The top part of the main screen shows general game information. Here, you can find contact information in case of technical issues, the name of connected team and of the current game, the current session and the current time step.
A game will normally consist of three sessions. In the example above, Session Number 1 of a three-session game is now in progress. The same logic applies to the Month indicator, i.e. there are twelve months in the session and we are now playing Month 6.
Portfolio Parameters
This is the most informative section of the user screen since it displays the main portfolio parameters. The most important ones are shown in bold font.
For instance, the amount of Short Term Capital rented during the current month is displayed below the label Short Term. As mentioned above, STC and LTC taken together determine VaR, which in turn determines the portfolios current probability of default. Thus, in the picture screen, the total amount of capital available is $855.00 while current VaR is $855.26. This computes to a current probability of default of 1%, which is the set point for the game. Also, a team can monitor its prior probability of default as well as the critical portfolio-default probability beyond which it is expelled from the game. In other words, if a teams default probability rises above this level at any time during the game, it will be expelled until the end of the session and will be allocated the highest expense level.
13
The next line shows Value at Risk (VaR), mean portfolio loss, and portfolio loss standard deviation. These statistical measures are used to evaluate the credit risk in the portfolio and derive its probability of default. The final row in the table displays the Total Capital Expense one of the parameters used to determine a teams score. Expense can be positive or negative. A negative expense means that profits were realized as a result of risk management. Portfolio expenses include trading costs, any CDS premia paid and received, interest on rented capital, and interest received from renting a teams excess available capital to other teams.
Expense Meter
This graph shows the relative performance of all teams playing a game. Depending on actions taken by every team and due to a series of random worldevents, the Expense Meter for each team will be associated with different values as time progresses. The Expense value is cumulative, therefore only the last value is used when computing the final score. The graphs purpose is to give team members an idea of how they are doing during the session. Based on this information, a team can decide which actions, if any, to take in order to increase its risk management efficiency.
Portfolio
The portfolio is represented via the exposure table. Each exposure is displayed on a single row in the table. Exposures can be sorted by Name, Probability of Default [x], LGD, Loss, and associated CDS position. To sort the table, simply click on the corresponding label. When the same label is clicked a second time, the sorting order toggles between ascending and descending.
14
Sorting the table is convenient in cases where, for example, participants want to see the exposures with the highest default probabilities or else their current CDS position on said exposures. As the default probability [x] changes over time, the corresponding arrow will be shown next to the x value. These arrows work the same way as those on the downgrade and upgrade screen mentioned earlier, i.e. when the default probability goes down, the arrow is green; when it goes up, it is red.
Control Buttons
The set of control buttons at the bottom of the user screen allows the team to perform various actions and to obtain useful feedback-information. The Capital Management button brings up the screen where Short Term Capital can be rented from the market, as well as traded with other teams. The Scale button provides the relationship scale between probability of default and letter-grade credit rating. Portfolio Progress shows the prior evolution of various portfolio characteristics.
The Skype button has Skype IDs of all playing teams and helps to establish communication between them. The Refresh Portfolio button is included to help players with older versions of web browsers. If you believe that your portfolio table does not reflect the real picture (for example just traded exposure is not in the portfolio), please press this button and the latest information will be retrieved from the database. The World Events button provides the log of all events that occurred during the current session. The log file contains all upgrades and downgrades, as well as all defaults with corresponding time step. The My Last Actions button displays a history of all prior actions taking by a playing team. These actions are recorded with their corresponding time step and other meaningful information. Sometimes, it can be helpful for players to see which actions have been taken as a guide to a better risk management strategy. The Ready button, when pressed, lets the system know that a team has made all its current decisions and is ready to proceed to the next time step. After a team presses this button, changes to any parameter can only be made at the next time step. In connection with this button, it is crucial not to forget to rent Short Term Capital at every time step. Also, please note that the system will prompt each team if the amount of its requested Short Term Capital is zero. At that moment, the team can decide to change it or else proceed if this value is a conscious part of its current risk management strategy, i.e. to rent zero Short Term Capital during the next time step. This maneuver should only be done carefully.
15
Capital Management
Using this screen participant can mange Short Term capital. The small graph on the right hand side shows interest rates. The green line indicates Long Term Capital interest rate, which is constant over the course of the game. The blue line represents Short Term Capital interest rate, which changes from one time step to another. Any player can communicate with other players using the Skype buttons placed in front of team names in the table below. As mentioned before, there are two ways to obtain Short Term Capital, i.e. from the market and from other teams. We will now look at both ways a little closer.
Interest Rates
As you can see on the graph the Short Term Capital interest rate changes over time. Its always higher than the long term rate and intimately tied to the teams portfolio default probability. Normally, when a teams default probability rises, this rate will increase and conversely, when the default probability decreases, the rate decreases. Long Term capital is a much cheaper source of capital. However, it lacks the flexibility provided by Short Term Capital. Therefore, it is a good idea to wisely allocate an optimal amount of Long Term Capital at the beginning of every session. Participants should be careful when making a decision about this allocation. For instance, allocating too much Long Term Capital can lock in a teams total capital amount at a relatively high level, and hence prevent the portfolios probability of default from being maintained at the desired level. This means that the portfolio default probability will be too low even with zero Short Term Capital allocation. Conversely, allocating too little may lead to a higher total cost of capital, and therefore, to abnormally high total capital expenses.
16
17
Scale
The Scale button acts as a source of additional information on how letter-grade ratings derived from the probability of default. For example, in the table at right, you can see that a Baa2 rating, on average, corresponds to a 1.25% annualized portfolio default probability. Actual default probabilities in the game are unlikely to be exactly the same as those in the table. In intermediate cases, the game uses the geometric average bracketing two lettergrade ratings to figure out which rating should be assigned to a particular exposure. Normally, a participant will not use this table explicitly. Nevertheless, it helps to establish an intuitive feeling between letter-grades ratings and default probabilities.
18
Portfolio Progress
The Portfolio Progress button brings up a screen showing a time series of a teams benchmarks. There are five graphs available on this screen. We now briefly discuss each one below.
Probability of Default
This graph shows the evolution of the portfolios default probability given total available capital in each round. As you will recall, every teams goal is to keep its portfolio default probability equal the set point (usually 1%) by renting capital, trading exposures, etc. Therefore, this graph shows how efficient a team is able to manage credit risk. The more volatile this graph, the worse the teams performance.
19
VaR
VaR shows the portfolios Value At Risk metric as it evolves over time. This may be useful when making tactical decisions. One of the main tasks of a risk manager is to keep enough cash to cover VaR at all times. For example, on the graph at right, VaR at time step 9 was $750. If the playing team had a total capital in the amount of $750, it would be considered a very efficient team at managing credit risk.
Expense Meter
To enable a quick comparison and an assessment of the game situation, a global Expense Meter is shown on the main screen. It shows the capital expense path only and a normalized numerical value of expenses. Using this graph, players can evaluate their performance from one time step to the next, and potentially correct their strategies. Thus, the expense meter can be regarded as the main game benchmark.
20
Skype
The Skype button acts as a small communication center, providing the list of all playing teams and their Skype IDs. Players will need to use Skype to make verbal capital trading agreements with each other. The Status field shows whether a team is available to talk and, if pressed, automatically connects Skype to the corresponding team.
World Events
This is an informational screen displaying a log of all of all exposure upgrades, downgrades and defaults. Every entry shows the corresponding time step. All parts of this log have are shown to participants before every round. Therefore, the purpose of this screen is to provide players with additional historical information about World Events.
World Events occur randomly. Nevertheless, there is logic behind this randomness. Thus, its very unlikely that an exposure rated AAA would ever default. However, in some rare cases an exposure rated AA will be made to default. This can happen in the real world and the game carries this possibility as well.
Last Actions
The Last Actions screen keeps track of actions taken by players in the course of the session. Every action has a corresponding time step. This information can be used to
21
refresh your memory on which actions were taken previously and to see how they affected the current situation. Among other things, participants can see how much capital they rented from the market, what CDS positions they entered to, which exposures where traded, etc.
A green light means that players can make their decisions. In this mode, all buttons and screens are available to participants. After the Ready button is pressed, the light turns red, indicating that the system is waiting for all teams to be ready to start processing information. In this mode, all buttons are disabled and teams may not take any action. When all teams are ready and all requests have been processed by the system, teams are presented with the information window previously shown, the one that shows upgrades, downgrades and defaulted exposures. At that time, the traffic light turns yellow. This informs players that, when the Confirm button on the information window is pressed, they will be able to start making decisions (the green mode) that will affect the subsequent round.
22
Initially, both components are equally weighted. However, the game administrator can change the weights. This allows one of the above factors to influence the teams score to a greater degree than the other. The portfolios default probability at each time step should be close to the set point, normally 1%. The further the default probability departs from this level, the worse the teams performance and therefore the lower the score. There is no advantage to having the default probability below the set point. Moreover, there is no credit given to teams evidencing default probabilities lower than the set point, compared to those with default probabilities higher than the set point. For instance, a team with a portfolio default probability of 2% and another one with a probability of 0% would be scored the same because, in both cases, the probability is one percent away from the set point of 1%. The game contains another critical default probability level called the Doomsday level. Teams will be expelled from the current session when its portfolio default probability reaches the Doomsday value. Expelled teams will be ranked lower then any other team in the game no matter what score they currently register. If more than one team is expelled from the session, the team which survived the longer will receive the higher rank. The Total Capital Expense is the second contributing factor to a teams final score. The lower the price paid for managing its credit risk, the higher the value of the associated component in the final score. Therefore, it is not a good strategy from the expense standpoint to buy protection on every exposure in the portfolio, in an obvious and nave attempt to forestall all consequences of default, because the premia paid for this protection would increase the Total Capital Expense by a large amount. In general, a happy medium should be found by managing the portfolios default probability and paying as little as possible for doing so. This is the secret of winning the game.
23
Each session is independent of any other session with respect to team scoring. After all sessions are completed, the system will compute the final game score and rank the participating teams based on the combined results of all sessions. Teams that were never expelled will be ranked based solely on their scores. Thereafter, teams that were expelled at least once will be ranked based on how many times they were expelled, how long they survived until expulsion and the score they had achieved at that time.
24
25
Trading Exposures
Trading exposures provides a participating team with a very efficient and inexpensive tool for managing portfolio credit risk. Every month, each team will be offered a set of exposures that are deemed anti-correlated to its own portfolio exposures. Any exposure in the teams portfolio, except defaulted ones of course, can be traded in this manner. Transaction costs will be added to the Total Expense account. They are very low comparing to the amounts of money invested in the exposures, interest payments and CDS premia. Once exposure from the given set is chosen as a trading target, it replaces the original exposure in the teams portfolio. If CDS protection was purchased previously on the original portfolio exposure that is traded, the contract is terminated and the CDS termination value is added to the teams Total Capital Expense. The same procedure applies when protection was sold on the target exposure. As mentioned above, trading will be very helpful in cases when the portfolio contains some extremely risky exposures that make the task of estimating the VaR on the next time step practically impossible. Clearly, a team can always set up its own heuristic trading rules, like Dont keep any exposure with a default probability higher then 15% in the portfolio. Trading becomes a must have feature if such rules are to be enforced. However, the games main goal is not to trade exposures trying to stay away from all possible defaults. Rather, judicious a and spare use of trading is what is intended, hence the name of accidental trader
26
assumed to be equal to the number of rounds remaining until a given session terminates, i.e. no CDS contract is allowed to survive the current session. This procedure will result in a profit or a loss that will then be reflected on the Total Capital Expense Account.
27
28
communicating with other teams is crucial to find the balance between the needs and the abilities of borrowing and lending teams.
29