Part 5: Tactical portfolio

5.4
Consistent with Part 4 of this report, we have assessed NZDMO against five key criteria for ensuring a robust risk management framework with respect to the tactical portfolio. Given the technical nature of the review, we have chosen to assess financial risks and operational risk in separate sections.

Background

Tactical activities

5.5
The PMP describes tactical management activities as being the discretionary management of the net debt portfolio within established limits around the strategic portfolio. Within those limits (designed to mitigate market and credit risk), NZDMO’s Portfolio Managers have discretion as to the use of instruments and timing of transactions to effect movements in the portfolio.

5.6
NZDMO’s tactical activities have evolved over time. One of the first uses of the tactical portfolio was as a way for NZDMO to achieve the Government’s policy objective of a net foreign debt position of zero, without having to actually repurchase the gross debt. Since then, the NZDMO’s tactical activities have expanded to include funding the RBNZ’s foreign reserves, as well as intermediating foreign exchange (FX) activity for Crown entities and government departments. In recent years, government surpluses have added to the investment of funds within NZDMO’s tactical activities.

5.7
Operationally, NZDMO has established sub-portfolios within its tactical portfolio to reflect the nature of the assets and liabilities managed and to allow attribution of value added for management reporting purposes.

Tactical portfolio performance

5.8
NZDMO measures the performance of its tactical activity through both an absolute value-added figure and a risk-adjusted performance measure (RAPM) to reflect the amount of risk taken on to generate that value added.

5.9
The performance of the tactical portfolio under both these measures has shown substantial and sustained improvements during the last six years. This has been driven largely by the liquidity desk, as Figures 8 and 9 show. The majority of the value added created by NZDMO in the 2006 financial year related to cash management of $4,000 million of assets. Incremental value of $30 million was achieved primarily as a result of the unusually high treasury bill spread of more than 60 basis points (treasury bills being the key funding instrument for the cash assets).

Figure 8
Value added by tactical portfolios (2000/01 to 2005/06)

Source: NZDMO

Figure 9
Tactical portfolios risk-adjusted performance (2004/05 and 2005/06)

Source: NZDMO

5.10
The key driver of NZDMO’s value-added performance is its investment of surplus funds raised from the issue of government securities (New Zealand Government Stock and treasury bills) into higher yielding securities issued by institutions.

NZDMO’s tactical policy framework

5.11
We expected that:

• NZDMO’s PMP would describe the nature and rationale for its tactical activities;
• there would be a clear benefit to the Crown from NZDMO undertaking what it has identified as tactical activities; and
• clear limits would exist as to the extent of the tactical activities that can be undertaken in terms of the financial risks identified.

5.12
We found that:

• there is a clear benefit to the Crown from the tactical activities that NZDMO is currently undertaking;
• clear limits exist restricting the extent of the tactical activities undertaken (these have been approved by the Minister of Finance);
• the nature of NZDMO’s tactical activities was not explicitly stated in the PMP; and
• the rationale given in the PMP for undertaking tactical trading is now less relevant to NZDMO’s current tactical activities (which are typically not of a trading nature).

Tactical activity

5.13
The Minister of Finance approved NZDMO to conduct tactical trading in 1993. The stated arguments for undertaking tactical trading were that:

• temporary pricing imperfections sometimes occur, making it possible to generate profit from tactical decision-making;
• tactical trading builds knowledge of how various markets operate under a variety of circumstances. This improves NZDMO’s understanding in managing the overall portfolio. It is important, for instance, to maintain high-quality information flows about markets or sectors where intermediation transactions occur but are infrequent (intermediation transactions are where a substantial proportion of the value of tactical management is realised); and
• tactical trading enables NZDMO to build and maintain skills in analysis, decision-making under uncertainty, negotiations, and deal closure. The immediate benefit is a reduced risk of mistakes when transacting and the projection of a more professional image.

5.14
In terms of managing its tactical portfolios, the bulk of NZDMO’s activities are involved in the matching of assets and liabilities to off set interest rate and FX risks while minimising credit risk, and securing a margin through doing so. At the time of our audit, NZDMO was reluctant to take outright risk positions with a view to profiting from market movements.

5.15
In the context of undertaking this activity, the arguments advanced in the PMP for undertaking tactical management or trading appear less relevant. A lack of significant position-taking removes much of the opportunity to profit from temporary market pricing imperfections, other than in credit markets.

5.16
The tactical activities that NZDMO undertakes are both necessary and valid and have the potential to add value to the Crown within a managed risk framework. However, they are more in keeping with prudent asset and liability management and FX intermediation activity than traditional tactical trading.

Recommendation 8
We recommend that NZDMO update its Portfolio Management Policy to reflect that the tactical activity currently being undertaken is asset and liability management activity and foreign exchange intermediation, rather than primarily outright tactical trading.

Benefit to the Crown

5.17
The matching of financial assets and liabilities and the use of derivatives to manage the cash flow mismatches in a portfolio is standard asset and liability management practice. There is a clear benefit to the Crown from this. It allows for the minimisation of market risks while also focusing NZDMO Portfolio Managers on adding value.

5.18
A fairly recent and strongly growing part of NZDMO’s tactical business is FX intermediation for other parts of the Crown. NZDMO is able to provide these entities with forward exchange contract (FEC) pricing that is superior to that which they would otherwise be able to get from banks. This FX business currently involves turnover of about $50,000 million each year.

5.19
These arrangements provide a clear benefit for the Crown. In each case, the entity involved is able to access FX rates at prices that are better than those that they would receive from banks on their own. We note that entities retain the ability to trade with banks instead of NZDMO if there is an advantage to them in doing so. This acts to ensure the competitiveness of NZDMO’s pricing.

5.20
In all the above cases, after entering into a trade with a counterparty, the Portfolio Managers have discretion as to whether they immediately extinguish the risk (and the opportunity for profit or loss) on the deal by entering into a reverse trade with the market, or whether they leave the position open and seek to close it out later with the market at a more advantageous rate. This is subject to the risk of their trading positions remaining within their delegated limits.

Limits framework

5.21
Management of NZDMO’s tactical desks is performed within a limits framework to establish the maximum extent of risk that Portfolio Managers can take on from their discretionary activities. The limits framework has evolved over time and currently consists of both VaR limits to protect against potential market losses, as well as stop loss limits to protect against exposure to further losses once actual losses reach a certain point. Credit exposure limits are also applied to limit the risk of financial loss from a counterparty credit rating downgrade. These limits are set to reflect the Crown’s risk appetite and were approved by the Minister of Finance.

Deal execution processes and controls

5.22
Appropriate formal and informal management controls and practices are in place around the activities of the Portfolio Managers within the tactical portfolio. Formal controls include:

• granting of written delegated authority by the Head of Portfolio Management to transact certain types of instruments;
• approval by the Treasurer or Head of Portfolio Management to transact interest rate and currency swaps;
• strict constraints applying to NZDMO staff in relation to accepting entertainment and gifts;
• a requirement for the immediate entry of transactions into NZDMO’s treasury management system, and timeliness and sign-off of written confirmations; and
• a prohibition on historical rate rollovers of foreign exchange transactions.

5.23
Informal controls include:

• Portfolio Managers undertake transactions dependent upon their dealing competency as reviewed by the Head of Portfolio Management;
• Portfolio Managers are expected to estimate the profit and loss derived from their dealing activities, and are challenged on the rationale for any exposures not hedged;
• the Head of Portfolio Management undertakes minimal transactions, which lends itself to effective oversight of the actions of the Portfolio Managers;
• weekly Portfolio Manager meetings are held where topical matters and approaches are outlined act as a valuable control and oversight mechanism;
• distribution and management review of portfolio position reports are produced daily at 4.30 p.m., which includes reporting of profit and loss and adherence to approved VaR limits; and
• a procedures manual that comprises key operational information (for example, funding margins for advances) is used.

5.24
In terms of processes:

• responsibilities for issuing government bonds and treasury bills to the financial market are detailed in the agency agreement between the RBNZ and the Treasury – associated procedures are well-honed, given the number of tenders entered into;
• funding advances to entities are dictated by the formal lending arrangements in place; and
• for foreign exchange, NZDMO has made available comprehensive guidelines for the management of Crown and departmental foreign exchange exposure. This document sets out a standard operating process that helps ensure compliance with NZDMO dealing and settlement practices.

Financial risk identification and analytical techniques

5.25
We expected that:

• all trading positions giving rise to the financial risks identified would be recorded within the risk modelling and reporting framework;
• appropriate tools to model exposure to market risk would be in place and limits based on the modelled outcomes would be applied;
• credit risk exposures would be accurately calculated and managed within an appropriate limits structure; and
• risks around exposure concentrations (by instrument, counterparty type, country) would be managed.

5.26
We found that:

• trading positions within the tactical portfolios are recorded for risk modelling and reporting purposes;
• appropriate tools are in place to model market risk using a VaR methodology and stress testing. Limits to tactical risk taking are set using the VaR model outputs. While the VaR model’s output is not consistent with the stated 95% confidence level, the risk of a loss greater than the VaR limit is remote because of low levels of risk taking;
• current counterparty credit exposures are adequately modelled and exposures are maintained within an appropriate credit limit structure. Breaches of exposure limits are infrequent, are escalated and reported appropriately as they occur, and are resolved satisfactorily;
• NZDMO is aware of a number of errors in its potential credit exposure calculation methodology. A new credit risk methodology to correct current shortcomings is being created, though its implementation is dependent on development resources becoming available; and
• NZDMO manages some aspects of portfolio exposure concentration risk (counterparty, credit rating, and instrument limits). The residual concentration risk is relatively immaterial when compared to other financial risks.

Recording risk positions

5.27
NZDMO includes only positions within its tactical portfolio when calculating and reporting market risk against its overall portfolio limit. The risks arising from the mismatch of funding and lending (made through the quasi-tactical advances desks), are excluded, despite NZDMO actively managing these. This is because the quasi-tactical portfolio forms part of the strategic portfolio described in Part 4.

5.28
For management information purposes, NZDMO currently calculates VaR and value-added for this advances activity.

Recommendation 9
We recommend that NZDMO manage risks in its quasi-tactical advances desks within NZDMO’s risk limits for overall tactical activity (asset and liability mismatch management).

Market risk measurement and management

5.29
NZDMO measures its exposure to potential losses of value in its tactical portfolio from adverse market movements in two ways – VaR and stress testing.

VaR methodology

5.30
NZDMO uses a variance-covariance approach to measuring its VaR usage, and calculates VaR on a daily basis with a 95% confidence level for 1-day, 1-month, and 1-year periods. The VaR model uses an exponential weighting methodology (EWMA) to give greater emphasis to the most recently observed market price volatilities in the 120 calendar day set of market movement observations modelled. The database set of market movements from which the VaR model draws its volatilities and correlations is updated on a weekly basis.

5.31
The covariance approach to measuring VaR is well-established, and is widely used by a large cross-section of treasury operations. Furthermore, the employment of appropriate swap and sovereign debt curves for valuing positions is better practice. It ensures that the risk of fluctuations between government and swap curves are measured accurately. This is considered very important in the context of NZDMO’s business objectives and portfolio composition.

5.32
There are different schools of thought on the benefits of using EWMA. The determining factors for using it are usually the speed by which market fundamentals change, and the responsiveness required of the VaR number to these changes. In the former case, EWMA is often used for rapidly changing markets in which historically observed properties are no longer relevant. In the latter case, traders and Portfolio Managers usually prefer a VaR that responds quickly to changing market conditions. In contrast, a stable VaR would be preferred for determining capital adequacy or risk appetite for market risk.

5.33
In this context, EWMA is not considered inappropriate for NZDMO’s purposes. However, given the results observed from backtesting over a sustained period, EWMA may be responsible for dampening the volatility of the market movements used to generate the distribution of returns. This is because it dilutes the value of longer-dated historical data. As a result, only the most recent observations really count, and longer-dated historical stress events are strongly suppressed. In turn, this may have the effect of underestimating volatilities, correlations, and VaR.

5.34
Another issue is the length of data history to be used. More conservative and stable models use longer data histories and higher confidence levels. As a rule of thumb, histories of 500-750 trading days are typically observed in the market. This history is rolled forward on a daily basis. The use of EWMA is not compatible with enlarging the 120 calendar day dataset used by NZDMO, because of the methodology’s weighting of longer-dated historical movements.

5.35
Daily, monthly, and annual VaR limits for NZDMO’s tactical activity have been agreed by the Minister of Finance. The VaR limits have been set in terms of the degree of risk acceptable to NZDMO. The limits allow for increases in market volatility and short-term bulk intermediary transactions. Stop loss limits are also used to complement use of VaR and to protect NZDMO from further losses on its portfolio once they reach a certain point.

VaR Limits

5.36
Daily, monthly, and annual VaR limits for NZDMO’s tactical activity have been agreed by the Minister of Finance. NZDMO’s Treasurer has delegated authority to the Head of Portfolio Management to manage the tactical portfolios so that the calculated VaR does not exceed 50% of the authorised limit. From our observations, the VaR limit usage is typically 5-10% of the limit. This represents the low outright position-taking by NZDMO, with its tactical activities typically being asset and liability mismatch management and FX intermediation. As a result, there have been no VaR limit breaches or instances where actual losses have exceeded the VaR limit. There have also been no instances where incurred losses have reached the stop loss limits.

Backtesting

5.37
Validation of the calculated VaR numbers is performed through backtesting to gain comfort that losses do not exceed the calculated VaR more than 5% of the time. The model’s performance has improved substantially in recent years – from red to amber on the backtest scale of the Bank for International Settlements (BIS) – as a result of improvements to the model, as well as moving to a BIS-compliant clean profit and loss backtesting approach. Results have stabilised in recent rounds of backtesting, but the model consistently underestimates the observed confidence level of market risk losses (that is, modelled losses exceed VaR more than 5% of the time).

5.38
The sustained amber performance of the VaR model observed in backtesting points to an underlying systemic issue. This refers to an ongoing error that is inherent in the methodology or application of the model. The most likely sources of such errors tend to be:

• lack of data integrity;
• a shortcoming in the assumptions of the model; and
• errors in the implementation or function of the model.

5.39
There has been considerable work done by NZDMO, and there are ample indications, that suggest the data used to determine model parameters has a high level of integrity. A more viable source of error is indicated by the backtesting results. In particular, variance tests and non-normality analysis show that the VaR distribution is considerably “non-normal”. The covariance method enforces and relies on a normal distribution in estimating VaR. It is feasible that this causes an underestimation of VaR.

5.40
Accuracy can be confirmed only by a comprehensive validation of the VaR model. This highlights a process and control issue in NZDMO regarding the VaR model. Better practice would require an independent validation of the model engine and outputs. Furthermore, this function would need to be undertaken on a periodic basis (at least annually). Some elements of such validation are already being covered by backtesting. However, validation also includes data maintenance, implementation, and functional accuracy, and the security and control framework of the operational models. A best practice validation framework entails adequate independence and a formal approval process.

5.41
We do not consider the systemic error revealed by backtesting to materially aff ect NZDMO’s business and risk objectives, as a result of the current low level of VaR usage. The fact that VaR limit use is typically 5-10% of the overall NZDMO limit means that the likelihood of the NZDMO incurring a loss from its tactical activities greater than the VaR limit is remote. However, this issue may become material in future if NZDMO undertakes a greater level of risk-taking activities. In terms of better practice, the VaR model performance should be addressed.

5.42
At the time of our audit, NZDMO management informed us that they were also trialling VaR calculation using a historical simulation approach for VaR. This is a positive development. NZDMO reports that its backtesting validation of this model (revalued using 500-1000 days of historical prices) through backtesting has produced promising results.

5.43
We consider that using such an alternate approach for VaR may address the systemic errors. Historical simulation does not assume or enforce statistical relationships in the underlying market dynamics. Rather, market relationships are preserved by using the precise historical market data. This is one reason why the industry is moving toward using historical simulation for VaR. The downside is the increased eff ort, time, and cost relative to the covariance approach. This may mean a greater drain on resources, systems (for data), and operational run times. As NZDMO is some way toward addressing such factors, implementation of such a model would appear to be a viable proposition.

Recommendation 10
We recommend that NZDMO continue to seek improvements to its market risk modelling by considering implementation of the historical simulation methodology, and by producing periodic reports showing graphical time series of daily Value at Risk versus actual profit and loss.

5.44
Implementation of historical simulation methodology would be consistent with better practice. It removes the assumptions inherent in the NZDMO’s variance-covariance model of normally distributed market volatilities.

Stress testing

5.45
Stress testing is performed to overlay the daily VaR utilisation results. These tests provide an extra level of comfort that outlier events or unlikely losses are not being underestimated. Three particular stress tests are run that incorporate parallel and steepening/flattening yield curve shifts as well as stressing spread movements between swap and interest rate curves. These are typical scenarios used in the market. Several market participants use a broader array of shifts including combinations of curve shifts. This is more appropriate for portfolios that allow latitude for discretionary trading and are likely to contain considerable curve risk. This would not appear to be the case for NZDMO’s current tactical portfolios.

5.46
NZDMO is also creating stress tests of FX spot rates. The FX tests had not been implemented at the time of our audit, but will operate in a modular style in which a series of FX spot rate shifts may be entered in the model.

5.47
It may be instructive for NZDMO to introduce a fourth test representing an inverse curve scenario. This would be a stressed shift (50-100 basis points) at the long end of the curve, with the short end remaining unchanged. Two reasons for such a test are that:

• the current “steepening/flattening” test pivots at two years and does not stress the part of the curve that is close to the portfolio’s duration (at about three years); and
• the domestic curve is currently inverse.

Credit risk

5.48
The PMP does not allow NZDMO to transact or maintain an exposure to any counterparty (as either a security issuer or counterparty in a transaction) with a credit rating of less than “A-”, unless required to do so by government policy. Because of the resulting high credit quality of its counterparties, NZDMO assumes the probability of default by a counterparty to be zero (noting that any positions with a counterparty whose credit rating fell below “A-” would be closed out as soon as possible). Credit risk management is therefore concerned with the likelihood of incurring a financial loss as a result of a counterparty’s credit rating being downgraded.

5.49
NZDMO calculates credit risk using credit spreads and the probability of downgrade. The method incorporates credit spread losses that are expected to be incurred as a result of credit downgrades and break costs if NZDMO has to close out the position because the credit rating falls below “A-”.

5.50
Given the high quality of NZDMO’s credit counterparties and the composition of its portfolios, we consider the current methodology adequate for the portfolio and particularly in its treatment of probability of downgrade and default probability.

5.51
Maximum allowable mark-to-market credit exposure limits are set at the counterparty level with the approval of the Minister of Finance. These are based on the institution’s long-term debt credit rating, as assigned by the credit rating agencies Standard & Poor’s or Moody’s, and its type (sovereign, financial institution, or corporate). Where necessary, credit exposure limits for non-rated institutions are based on appropriation or Ministerial direction. Limits on sub-entities are approved individually, and the aggregated limit for the group is the highest credit limit of the entities that make up the group. Credit exposure reports are produced daily that show current exposure against limits by entity and group. Actual usage is generally low in comparison to approved limits.

5.52
If NZDMO has not had an exposure to an institution for two years, then that institution is normally removed from the list of approved institutions and must go through formal approval procedures again before it can be added to the approved institutions list.

5.53
If an institution is downgraded below “A-/A3”, NZDMO’s credit exposure for all outstanding transactions with that institution must be eliminated as soon as practicable, even if the net credit exposure to the institution is zero.

5.54
Credit exposures for an institution must not exceed the policy limit that is allocated to that institution, be it an institution or group limit. The Head of Risk Policy and Technology (RPT) is responsible for monitoring and reporting breaches of credit exposure limits to the NZDMO Treasurer. If a breach occurs, the Head of Portfolio Management is responsible for proposing a plan to the Treasurer for resolving the breach. The Treasurer is responsible for approving the planned resolution, or may authorise the maintenance of an exposure above limits for up to two months (after this, the permission of the Minister of Finance is required). Any breaches, how they occurred, and the steps taken to resolve them are reported in the next monthly report.

5.55
Breaches of credit limits seldom occur. There were two breaches in the 2006 year, and these were resolved within two to five days. Explanations of the causes and resolutions of these breaches were included in monthly reporting.

5.56
A further enhancement to existing practice would be to model potential exposure scenarios where the value of the instruments that give rise to the exposure has increased. This would demonstrate the effectiveness of exposure control procedures such as collateral calling and payment netting arrangements.

5.57
NZDMO models potential exposures under a market scenario where the value of the investments that produce the exposure have increased at the 99% confidence level.1 While this information is calculated daily and reported in the monthly report, it is used for management information purposes only, and no maximum potential exposure limits are applied.

5.58
NZDMO is aware that there are a number of errors in the calculation methodology. Potential exposures are not calculated for contracts that are presently out of the money. The current information produced is therefore not reliable. Additionally, the methodology does not apply the benefits of netting arrangements or the effects of collateral calls. As a result, the credit exposures calculated are likely to be overstated (that is, conservative).

5.59
NZDMO recognises the shortcomings of its current methodology and that market practice has improved significantly since its methodology was introduced in 1996. At the time of our audit, work was being done to calculate credit risk using simulation techniques. It is intended that this new system will, when implemented, provide a portfolio-wide potential loss measure by simulating potential market parameters into the future and revaluing the current portfolio exposures under these scenarios to arrive at potential exposure and loss measures. These include the effects of collateral and netting arrangements.

Recommendation 11
We recommend that NZDMO implement an up-to-date credit risk methodology, particularly with regard to potential exposure management to address the inherent conservatism in the current methodology. Potential credit exposure for each counterparty should be calculated for all outstanding transactions and incorporated into a credit limits framework.

Concentration risk

5.60
In a credit exposure context, concentration risk refers to losses from deterioration of the credit quality of a class of exposures (for example, exposures to an industry or geographical area).

5.61
Quarterly, NZDMO includes, in its monthly report, an analysis of its exposures by counterparty type, credit rating, and country level as a percentage of its overall portfolio exposure. However, no specific limits exist at these levels.

5.62
Better practice entails the management of the risk of portfolio credit concentrations through setting explicit portfolio limits by relevant concentration category. Such limits might theoretically be set as high as 100% of the overall portfolio exposure for exposure categories where credit concentration is not a concern.

Recommendation 12
We recommend that NZDMO manage concentration risk in its portfolio exposures by implementing a limits framework of maximum exposures (as a percentage of total portfolio exposure) by counterparty type, credit rating, and country.

Portfolio management control environment

Delegated authority framework

5.63
We found that the delegation chains were up to date and covered all of NZDMO’s activities. We found the control environment under which NZDMO operates to be effective both in terms of policy and procedures. In practice, NZDMO’s management takes a keen interest in all operational aspects of the function, and have instilled a risk-averse culture.

5.64
When transacting for a new government entity, NZDMO gains comfort around that entity’s authority to enter into the transaction through enquiry with the relevant Treasury Vote Analyst for that entity. Procedures for undertaking advances to entities are formalised through official debt arrangements. At the time of our audit there had been no cases of transactions being disowned or mismanaged by entities. Therefore, the risk of the Crown incurring financial loss through unauthorised activity is low. Suggested improvements to NZDMO’s control environment are detailed below.

Derivative instrument approval

5.65
NZDMO’s authority to transact in options was approved in principle by the Minister of Finance in 1994. They are listed as “approved in principle” instruments for transaction purposes in Schedule B: Approved Instruments of the Portfolio Management Policy (the Schedule). Final approval is based on the NZDMO Treasurer formally advising the Minister that the risk and controls surrounding transacting in options are sufficient. But this final approval has not been sought and options have not been transacted. The options that have been considered for use are exchange traded options on approved futures contracts, interest rate caps and floors, swaptions, and currency options.

5.66
Any derivative product for which the NZDMO Treasurer does not have authority to transact should not be stated in the PMP as a permitted product.

Recommendation 13
We recommend that NZDMO immediately remove options from the approved instrument schedule of its Portfolio Management Policy, and periodically review the lists of approved financial instruments (that is, Schedules B, C, and D of the Portfolio Management Policy) and update as required.

Product control

5.67
There are strong controls in NZDMO to ensure that only currencies and instruments approved in the PMP are transacted. The deal entry interface into NZDMO’s Treasury Management System (TMS) is limited to approved currencies and products with any required systems changes for new products or currencies overseen by Risk Policy and Technology staff . This limits the ability of Portfolio Managers entering trades in unauthorised products or currencies. If a deal is entered into outside of the TMS (for example, verbally), or a trade in an unauthorised currency or instrument is mis-booked as an authorised one, it would be detected by the Accounting and Transactional Services staff when the confirmation from the counterparty is received and no booked deal is found in the TMS or the confirmation details are found to not match those in the TMS.

5.68
While Forward Rate Agreements (FRAs) are listed as an approved instrument for transaction purposes in Schedule B: Approved Instruments, Portfolio Management Policy, FRAs have not recently been transacted by the Portfolio Managers, and the TMS no longer supports this derivative product.

5.69
Better practice suggests that any derivative product that is not supported by the in-house TMS, in terms of reporting and/or valuation and/or settlement and/or credit assessment, should not be transacted. The product should be removed from the list of approved instruments to help ensure compliance. Failure to do so could lead to undesirable consequences, such as monetary loss.

5.70
The addition of new financial instruments to the PMP is infrequent, the most recent being Mortgage Backed Securities in June 2004. The number of currencies in which NZDMO transacts has grown in recent years, to accommodate the foreign exchange hedging needs of NZDMO clients. This has seen the addition of Mexican pesos, Polish zloty, Thai baht, and Russian roubles to the list of authorised currencies.

5.71
The application of derivative instruments can change over time for legitimate business reasons – for instance, as a result of changes in customer requirements or in the approach to hedging or trading. As a result, it is important to ensure that the parameters surrounding the use of derivatives remain suitable for the organisation in terms of risk.

Recommendation 14

We recommend that NZDMO update its list of approved instruments to reflect only those instruments that NZDMO has the functional capability to process and that have been recently traded (for example, within the last two years) 5.72 Implementing recommendation 14 would ensure that NZDMO has the information technology capability to correctly identify the risk and record the profit or loss on instruments. It would also act as a check to ensure that NZDMO’s skills in transacting in the instrument are current and that the business case remains valid.

Consistency of delegated authority

5.73
Of the derivatives approved for use by the Minister of Finance, only asset swaps and interest rate swaps have a formal restriction on the transaction term for which the Portfolio Managers (in this case, the Head of Portfolio Management) are able to transact up to 10 years’ maturity.

5.74
There is a formal term limit on the authority to transact some financial transactions but not others, which make up a substantial portion, of approved financial instruments (for example, security investments, futures, and currency swaps). This limit is inconsistently applied.

Recommendation 15
We recommend that NZDMO review the term limitation formal authority across all the financial products that the NZDMO Treasurer can transact, and seek to apply a term limit that is based upon Portfolio Manager seniority and potential economic outcome.

Risk and performance reporting

5.75
NZDMO’s reporting of its tactical activities against its policy framework serves two key purposes:

• ensuring compliance with the risk policies within the PMP; and
• reporting NZDMO’s performance in managing the tactical portfolio.

5.76
We expected that:

• effective daily reporting procedures would be in place to provide management with information on the tactical portfolio’s financial performance, risk position, and compliance with the PMP limits framework;
• sufficient information would be reported to senior Treasury management and the Advisory Board to enable them to gain an accurate appreciation of NZDMO’s performance in managing the tactical portfolio, the structure and risks of the portfolio, and its compliance with the PMP limits framework; and
• NZDMO’s management of the tactical portfolio would be measured and reported using an appropriate performance measurement framework.

5.77
We found that:

• NZDMO’s daily reporting is performed in a timely manner and provides management with high level information on the portfolio performance, position, and compliance with the risk policies;
• monthly reports to senior Treasury management and the Advisory Board, while containing appropriate summaries of risk limit usage and compliance with the PMP, did not provide information on portfolio composition; and
• the risk-adjusted performance measure (RAPM) that NZDMO uses to report the performance of its management of the tactical portfolio has limited information content for measuring NZDMO’s risk management activities.

Daily management reporting

5.78
A range of end-of-day reports are normally available to NZDMO’s Risk Policy and Technology staff and the front Office from 4.00p.m. each afternoon. The reports include reporting against the risks identified in the PMP framework, as well as daily profit and loss measures.

5.79
This reporting process is the way any limit breaches are detected and escalated for resolution by front Office staff . The reports are also the mechanism through which the Head of Portfolio Management and the NZDMO Treasurer maintain oversight that the changes to the portfolio each day are consistent with their expectations.

5.80
We found that the reports contained high level analysis of the portfolios, showing current net position value and VaR by currency. However, they did not readily provide diff erent views of the portfolios, such as analysis of exposures by maturity or risk sensitivity. While we understand that such reports are able to be run as required, we consider that they should be produced for management as part of the daily reporting to provide further insight into the drivers of the reported VaR measure.

Recommendation 16
We recommend that NZDMO consider expanding the suite of daily management reports to include portfolio composition views. These could include an analysis of portfolio assets and liabilities by maturity, stated in dollars and/or risk sensitivities.

5.81
Expanding the suite of daily management reports would provide readily accessible information to Portfolio Managers, risk managers, and executives on the risk construction of the portfolio. This would provide greater insight into the drivers of the high level VaR measure than is currently provided.

Monthly external reporting

5.82
Monthly reports are prepared for the Secretary to the Treasury and senior management within the Treasury. These monthly reports are also included in papers that go to NZDMO’s Advisory Board ahead of its quarterly meetings. The reports contain information regarding performance metrics (RAPM), value added (refer to comments below on the meaning of this in NZDMO’s context), and VaR, as well as commentary on the NZDMO’s activities and its compliance with its PMP.

5.83
The monthly reports contain very little detail regarding balance sheet composition and portfolio characteristics.

Recommendation 17
We recommend that NZDMO include an overview of the current balance sheet composition of the tactical portfolio in monthly reports to senior management and NZDMO’s Advisory Board to better convey the extent of matching of assets and liabilities (and the associated residual risks) within this portfolio.

Risk-adjusted performance measure

5.84
NZDMO measures the performance of its management of the tactical portfolios using a risk-adjusted performance measure (RAPM). This measure is designed to report the value added from the NZDMO’s funding and risk management of its tactical activities compared to the amount of notional capital that is required to underpin these activities.

5.85
It makes sense that there should be correlation between the performance of the risk management activities and the amount of risk being taken.

5.86
One of the contributors to the relatively high reported returns by RAPM is that the value-added calculation includes one-off gains that are achieved from issuing New Zealand Government Stock and treasury bills (in recent times, 100 basis points below the New Zealand swap curve) and swapping the proceeds into floating rate funds. For daily management reporting purposes (not external financial reporting), NZDMO values both its bonds and swaps off a single valuation curve that is equivalent to the swap curve less 12.5 basis points. As a result, a one-off gain is recognised when debt is revalued (because of a reduction in its value) using a higher swap-based interest rate than it was issued for.

5.87
However, swapping of a low fixed rate coupon on debt for a floating rate coupon at swap rates less a margin does not (excluding interest rate movements) give rise to a profit in the traditional sense. The gain that is being included in the RAPM measure is representative of the Crown’s ability to issue debt at a lower cost than swap rates.

5.88
In our view, it is inappropriate to describe this gain as value added in the same way as other components of the value added by NZDMO in its management of the tactical portfolio. This is because it is an inherent advantage of NZDMO’s funding activities rather than a result of its explicit risk management activities. It is also because there are a number of other wider macroeconomic and financial market drivers, outside of NZDMO’s control, that also influence the swap spread.

5.89
In NZDMO’s view, it is appropriate to describe the gain primarily as value-added because NZDMO explicitly manages the risk in terms of the spread between its sub-100 basis point funding cost and the comparative investing curve. We accept that there are aspects of the spread that are explicitly managed by NZDMO, for example:

• maintaining the liquidity premium for the New Zealand Government bond market;
• managing key relationships to maintain a strong government credit rating; and
• managing the credit risk associated with assets invested against the funding instruments.

5.90
Under NZ IFRS, no one-off gain would be recorded from NZDMO’s funding advantage, as the bond would be valued off the New Zealand Government curve and the swap off the swap curve would be adjusted for the margin received. However, for management reporting purposes (under the RAPM model), NZDMO will record a one-off benefit from the higher yields it is able to invest in compared to its funding cost.

5.91
RAPM can increase rapidly if notional risk capital is reduced while simultaneously the margin between funding costs and investment returns is maintained or increased. NZDMO has acknowledged that aspects of the RAPM model could be improved.

5.92
We have discussed with NZDMO one method that could be employed to measure risk adjusted performance, which is basically an Economic Value Added (EVA) approach.

Recommendation 18

We recommend that NZDMO refine its value-added performance measure (RAPM) to better report the diff erent components of market risk taking returns and the one-off gains achieved by borrowing funds at the New Zealand Government rate and reinvesting them in marketable securities (which attract a higher rate).NZDMO should also mark-to-market all assets, liabilities, and derivatives at the appropriate yield curves that include the credit spread for each issuer.

5.93
Implementing this recommendation would remove the effect of reporting one-off gains of the Crown’s funding advantage. Instead, the margin gained would be then recognised over the life of the transaction. This would ensure that the performance measure provides NZDMO’s treasury management analysis of returns associated underlying market risk positions only, and would be consistent with reporting under NZ IFRS.

5.94
From this performance analysis, the cost of capital could be deducted providing a net income after capital calculation. To calculate the cost of capital, NZDMO should determine the notional risk capital required for the tactical portfolio, which would then be multiplied by the cost of capital (which should be an appropriate New Zealand Government bond rate). NZDMO should still be able to monitor and report mark-to-market gains or losses on government and swap curve-related instruments as a separate item within its management reporting if this information is considered to be useful.

5.95
We note that NZDMO’s move to financial reporting under NZ IFRS has provided it with an appropriate valuation methodology that could be extended to its management reporting and performance analysis.

Recommendation 19
We recommend that NZDMO use, where appropriate, the valuation methodologies developed for external reporting under New Zealand equivalents to International Financial Reporting Standards for its management reporting.

---

1: This is effectively the converse of VaR. The potential exposure methodology measures a profit scenario at the 99% confidence level, while VaR measures a loss scenario at the 95% confidence level.

page top