70897 PAKISTAN: ISSUES RELATED TO THE GOVERNMENT SECURITIES MARKET AND GOVERNMENT DEBT MANAGEMENT Policy Note The government of Pakistan borrows in the domestic market through a range of instruments, and this market is a critical source of funding for both shorter-term cash management and longer-term deficit-financing. The government has taken actions over the past 18 months that have enhanced the effectiveness of the market as a source of funding, as well as its efficiency. These include the movement toward more predictable, volume-based, market- determined pricing of government securities. Taking account of the dynamics of demand will be important as the government continues to develop its medium-term debt management strategy. Doing so will help identify potential constraints that may impede the implementation of the chosen strategy. Specific actions that the government is recommended to take include (a) reducing the number of tenors issued, (b) consolidating the debt stock so as to improve liquidity in individual bonds, (c) reducing time delays in auction processing, and (d) developing and investor-relations function, with readily accessible information on the operation of the government securities market. World Bank Washington, June 25, 2010 I. INTRODUCTION 1. The government of Pakistan borrows in the domestic market through a range of instruments. In the wholesale segment, the government issues Market Treasury Bills (MTBs), Pakistan Investment Bonds (PIBs), and some shariah-compliant sukuks, the issuance of which began in mid-September 2008. Non-market retail debt, represented by non-tradable National Savings Scheme (NSS) instruments issued by the Central Directorate of National Savings (CDNS), makes up the largest share of the portfolio; recently, in January 2010, tradable retail bonds, called National Savings Bonds (NSBs), were also introduced. In addition, the government has recourse to direct borrowing from the State Bank of Pakistan (SBP) in the form of 6-month Market-Related Treasury Bills (MRTBs) that are created on demand. 2. The domestic market is a critical source of funding for both shorter-term cash management and longer-term deficit-financing. The government has taken actions over the past 18 months that have enhanced the effectiveness of the market as a source of funding, as well as its efficiency. These include the movement toward more predictable, volume-based, market- determined pricing of MTBs and PIBs. 3. Identifying and analyzing potential funding sources is an essential part of debt management. Taking account of the dynamics of demand will be important as the government continues to develop its medium-term debt management strategy. Doing so will help identify potential constraints that may impede the implementation of the chosen strategy. 1 4. This policy note presents observations on current dynamics of the government securities market, as well as recommendations for improving the functioning of the market, which would contribute to more effective debt management. The focus of the note is on the primary market for wholesale securities, although other issues are examined. The note draws on consultations with participants in the financial market, both bank and non-bank institutions: local commercial banks (Habib Bank, JS Bank, MCB Bank, National Bank of Pakistan, and United Bank), foreign commercial banks operating in Pakistan (Citibank, Royal Bank of Scotland, and Standard Chartered Bank); Islamic Banks (Bank Islami, Dawood Bank, and Meezan Bank), insurance companies (State Life Insurance, Adamjee Insurance, and EFU Insurance), pension funds (Employees Old-Age Benefit Institution), mutual funds (Arif Habib Securities and National Investment Trust), and exchanges (Karachi Stock Exchange and National Commodities Exchange). 1 Refer to the annex for an overview of the steps involved in preparing a debt management strategy. 2 II. ANALYSIS OF KEY ISSUES A. Issuance Calendar, Volumes, and Frequency 5. The government’s credibility and commitment to market-based funding costs in the primary market has improved over the past year. This view is held uniformly by market participants. One factor is the publication of, and adherence to, an auction calendar, which enables the market to plan for issuance and better absorb the government’s funding needs. 6. Almost all wholesale issuance occurs according to a calendar. While MTBs have been issued fortnightly under long-standing practice, an auction calendar of PIBs was introduced for the first time in the 2008/09 fiscal year. For MTBs, the calendar covers 3 months. It specifies auction dates, the tenors to be offered at each auction, and the total target volume; tenor-specific target volumes are not announced. For PIBs, the calendar covers 6 months. It specifies auction dates, the tenors to be offered at each auction, the total target volume, and (in contrast to the practice for MTBs) the target volume for each tenor. Sukuks are auctioned from time to time, but not according to a precise calendar. 7. Importantly, the government discontinued its prior practice of applying low or arbitrary cut-off yields in auctions. Bid acceptance by and large respects the volumes announced in the calendars. This has been another factor that has enhanced the credibility of the government as a price-taker, instead of being a price-maker, in the primary market. 8. Market participants report mixed views regarding the frequency of auctions, but most are comfortable with current practice. Fortnightly auctions of MTBs are considered appropriate by the majority of institutions, although some view that as too frequent and propose monthly issuance instead. PIBs are issued approximately every two months, which is satisfactory to the majority of institutions. A few institutions, however, consider that more frequent auctions of PIBs would be beneficial, such as every month, while one institution, on the contrary, reports that only one PIB auction each quarter would be appropriate. 9. Determining the optimal frequency of auctions is not straightforward and has implications for motivating the secondary market and managing the government’s exposure to interest-rate volatility. Conducting more frequent, but smaller-volume (as opposed to less frequent, but larger-volume) auctions could improve price discovery (insofar as secondary market yields are regarded as less reliable) and reduce inventory risk for primary dealers. This approach would also assist the government in diversifying its funding costs through the year by sampling interest rates across time. The appropriate balance needs to be struck, though. Too frequent auctions dampen incentives for outright secondary trading and increase the likelihood that some primary dealers may decline to participate in auctions from time to time, thereby reducing price tension. Recommendations • Continue the practice of publishing and adhering to an auction calendar for MTBs and PIBs. 3 • Consider introducing an auction calendar for sukuks. • Continue the practice of issuing to the pre-announced target volume. B. Tenors and Fragmentation of the Debt Stock 10. The government issues securities at a wide range of tenors, at 10 points on the conventional yield curve. MTBs are issued with tenors of 3, 6, and 12 months. PIBs are issued with tenors of 3, 5, 7, 10, 15, 20, and 30 years. For both MTBs and PIBs, all tenors are offered at every auction. The sukuks issued to-date had tenors of 3 years. 11. On the one hand, the wide range of tenors enables the government to accommodate the maturity preferences of different market segments. For example, commercial banks usually have a strong demand for short-term securities for their liquidity management needs. This is buttressed in Pakistan by the Statutory Liquidity Requirement (SLR). Their demand may also extend to medium-term maturities if their retail deposit bases are stable or they are faced with persistent excess liquidity, although this has been less true in Pakistan. Given their long liability tails, life insurance companies and pension funds have a natural appetite for long-term securities. In Pakistan, two of these institutions have dominated holdings of longer-term securities, with State Life Insurance being almost a monopsony at tenors beyond 10 years. Contractual Savings Institutions and Government Securities The development of contractual savings institutions, with their emphasis on the long term, is helpful for the development of the long end of the market. Pension funds and life insurance companies have the potential to stimulate the development of the government securities market because they benefit from economies of scale and have access to highly professional asset management services. To achieve their full potential, however, a number of pre-conditions must be met. First, they should not be treated as captive sources of government funding and should be free from compulsory investment requirements. Second, they should have a pluralistic structure and not be dominated by a small number of public sector entities. Third, they should reach a mature stage in their development. During the phase of rapid accumulation of long-term capital resources, pension funds and life insurance companies tend to engage in “buy and hold� strategies and rebalance their portfolios through changes in the asset allocation of net new inflows. During this phase, they can be important participants in the primary market, but they need to reach a mature stage in their development before they become active participants in the secondary market. They reach this mature stage when their new net inflows become a small fraction of their total assets and rebalancing of their portfolios requires active trading of their asset stocks. Fourth, even when they reach this stage of maturity, they need to adopt sophisticated asset allocation strategies and emphasize the benefits of diversified portfolios and professional asset management. This implies access to a competitive and professional asset management industry, the use of specialized asset management mandates, and monitoring of investment performance against appropriate benchmarks. Even under such circumstances, institutional investors may engage in “buy and hold� strategies with regard to some of their assets. In general, however, mature institutional investors engage in active trading for rebalancing their asset portfolios and even pursue active yield-enhancement strategies, such as securities lending and options writing, when these are permitted. 4 12. On the other hand, issuing at too many tenors contributes to fragmentation of the debt stock, which is the case in Pakistan. As at end-May 2010, the stock of PIBs totaled PKR 505.3 billion. This was spread across 41 distinct bonds, with up to six separate bonds maturing in some years. Individual outstandings range from under PKR 0.1 billion up to PKR 39.4 billion, with an average of PKR 12.3 billion (refer to figure 1). Figure 1: Redemption Profile of PIBs, as at end-May 2010 80 70 60 50 PKR billions 40 30 20 10 0 2011 2013 2015 2017 2019 2021 2023 2025 2027 2029 2031 2033 2035 2037 2039 13. A sound issuance practice would be to promote benchmark issues in key maturities that facilitate the growth of secondary markets. This has not been achieved in Pakistan. The commercial banks consulted are uniform in the view that the number of tenors offered could be reduced by eliminating issuance at the 7-year point. Instead, issuance should be concentrated in the 3-, 5-, and 10-year securities, which would still provide a sufficient range to accommodate different investor preferences. Similarly, there is a strong case for consolidating issuing at the ultra-long end (beyond 10 years) by collapsing issuance into one tenor, such as 20 years. According to State Life Insurance, however, the only material investor in this segment, collapsing issuance would complicate its liability-driven portfolio management. The counter argument is that its portfolio management is, in fact, passive and that similar aggregate interest- rate risk management could be achieved through combinations of long- and medium-term securities. Importantly, reducing the number of tenors offered would not have an undue impact on the government’s ability to manage its interest-rate and refinancing risks in respect of domestic debt. 14. Routine re-opening of seasoned issues to build them to adequate size was introduced in May 2006. Each year, a new PIB is launched at each maturity point up to 10 years, which is re- opened at the subsequent auctions in that year. PIBs beyond 10 years are re-opened over the course of two years. Re-opening was a positive innovation and has helped to increase the supply of on-the-run issues. To ensure compliance with shariah principles, it is in practice infeasible to re-open sukuks. 5 15. In the absence of liability management operations to transform the structure of the portfolio, such as buy-backs of rump issues or switch operations, consolidating the debt stock in fewer (and more liquid) bonds will take time. A useful step would be to extend the re- opening arrangements. For example, the 10-year bond could be re-opened in five years’ time as the on-the-run 5-year bond, and again in two years’ time after that as the on-the-run 3-year bond. The success of this approach depends on two main considerations. First, market participants must not have an aversion to PIBs whose coupons may be significantly off-market. The market participants consulted report that off-market coupons are not a concern. Second, the government must be able to manage the large cash flow that would be represented by the redemption of a very deep bond. This may be more problematic in the case of Pakistan, although the government would not face this problem for several years, during which time it could develop approaches for managing such a situation, such as pre-redemption buy-back windows or within-year pre-funding to accommodate upcoming redemptions. 16. A related issue is the scope to rebalance the relative shares of MTBs and PIBs. As at end-May 2010, the stock of MTBs totaled PKR 1,290.3 billion, against PKR 505.3 billion for PIBs. Although Pakistan has been able to issue successfully long maturities, the relative weighting is strongly skewed toward the short end of the yield curve. In light of the historic vulnerability of the macroeconomic environment and the resulting interest-rate volatility, investors are reluctant to shift from MTBs to PIBs. One consequence is a heightening of the government’s interest-rate risk. The commercial banks consulted support a gradual re-allocation from MTBs to PIBs, consistent with ensuring that the supply of MTBs remains sufficient for them to satisfy SLR. Recommendations • Discontinue issuance of the 7-year tenor. • Collapse issuance of the 15-, 20-, and 30-year tenors into one point. • Discontinue launching new securities each year. Instead, to the extent possible, re-open existing securities as the new on-the-run issues. • Commence a gradual re-allocation of issuance away from MTBs toward PIBs. C. Auction Disclosure and Market Information 17. Pakistan presents an interesting, and unusual, example of auction disclosure. At the close of an auction, bids are opened in the presence of representatives of all the primary dealers. Not only is the full bidding schedule disclosed (which is sound practice), but so is the identity of the bidder of each bid. This practice is not followed in other markets, as it can disclose sensitive information about each bank’s position to its competitors. Market players with more sophisticated trading strategies especially value confidentiality. 18. The case for changing the current disclosure practice is not clear-cut. Some primary dealers favor a move toward confidentiality of bidders, whereas others support the status quo. 6 The different in preference is not attributable to distinctions between local versus foreign banks, or between relative sophistication of trading activity. Instead, it is generally one of market size. Smaller primary dealers see greater benefit in knowing the bidding patterns of their competitors than a cost in their own bidding patterns being revealed, whereas larger primary dealers prefer confidentiality. The existing arrangements are well-understood and do not appear to reduce participation in auctions materially. Given at the current state of market development, the benefit of transparency outweighs any cost introduced by (arguably minor) disincentives arising from the current practice of full disclosure. 19. A more pressing issue is the delay in releasing auction results to the wider market. It usually takes several hours for the results to be released, and this often happens only after the market has closed for the day. The delays have grown longer over the past 18 months, as exclusive responsibility for auction decisions was transferred from SBP to the Budget Wing in the Finance Division. Market participants are uniform in the need for more rapid release of results. It is not clear that the Budget Wing is the entity best placed within the Finance Division for approving auctions. Market-facing financial operations are not one of its core activities, which may contribute to a lack of urgency on auction days. Similarly, the process of internal approvals is not conducive to rapid decision-making. A reasonable alternative would be to transfer responsibility for auction approvals, as well as other coordination related to government securities issuance, to the Debt Policy Coordination Office (DPCO). 20. The prospectus arrangement for government securities should be improved. At the highest level, government securities are governed by the Public Debt Act 1944, and the terms and conditions for the issuance of government securities are established in the PIB Rules and the MTB Rules. These are available through the Pakistan Gazette. As auction agent for the government, SBP prepares more operational terms and conditions, including regulations governing the primary dealer system. These are well-understood by market participants and are publicly available from SBP in print format, as well as from its website, although in a highly disaggregated format. SBP has a project in place to prepare a government securities market handbook, which would consolidate all information and relevant terms and conditions in one document that would be easier for current and potential market participants and other investors to use. This is a useful initiative. Recommendations • Streamline the process within the Finance Division for approving the acceptance of bids in auctions, so that results can be reported to the market within one hour of the close of an auction. To that end, transfer responsibility for auction approvals, as well as other coordination related to government securities issuance, to DPCO. • Complete preparation of the government securities market handbook. 7 D. Secondary Market 21. Secondary market trading has been facilitated by the Bloomberg electronic platform launched in January 2010. The primary dealers and three additional banks post indicative two- way prices on MTBs and PIBs on the platform. To conduct a trade, a participant would normally request firm quotations from one or several banks. Trades can be conducted on the platform (approximately two thirds of all trades) or off the platform. If conducted off the platform, however, trades must still be reported on the platform before the end of the day. SBP publishes the total trades in each MTB and PIB on a daily basis on its website. 22. The Bloomberg platform has not led to an increase in the total volume of trading, but it has delivered improvements. Price discovery and post-trade transparency have both been enhanced. Likewise, spreads in some maturities have tightened. In the 10-year tenor, the most active on-the-run PIB, spreads were previously around 10-15 basis points. The maximum spread is now 5 basis points, and often at 2-3 basis points. 23. The secondary market remains thin, however. Outright sales (call and clean) are limited, and trading in PIBs in particular is especially thin. Over the 11 months of July 2009 through May 2010, average total monthly repo volumes of MTBs and PIBs were 6.9 times through of non-repo transactions (refer to figure 2). Among repo transactions, MTBs accounted for 98 percent. Similarly, monthly turnover ratios for both repo and non-repo transactions of MTBs and PIBs are low (refer to figure 3). Figure 2: Trading Volumes of MTBs and PIBs (Nominal Amounts) 1,400 1,200 1,000 PKR billions 800 Non-repo 600 MTB Repo 400 PIB Repo 200 0 Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May 2008 2008 2008 2009 2009 2009 2009 2009 2009 2010 2010 2010 8 Figure 3: Trading of MTBs and PIBs (Relative to Outstandings) 1.2 1.0 0.8 Turnover ratio 0.6 Repo 0.4 Non-repo 0.2 0.0 Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May 2008 2008 2008 2009 2009 2009 2009 2009 2009 2010 2010 2010 Recommendations • Continue promoting the Bloomberg platform for government securities trading, while ensuring that it does not preclude further evolution of the market. E. Non-Resident Participation 24. Non-resident participation in the government securities market is negligible. Several factors are behind this. Macroeconomic instability in recent years, plus the continuing fragility of the economic outlook, is a disincentive. In addition, insofar as non-residents have an appetite for Pakistani sovereign risk, they would prefer to avoid a PKR currency exposure and, instead, would choose to invest in the eurobonds that Pakistan issued earlier. Continuing improvements in the economic outlook are needed to mitigate these factors. 25. Another factor is that non-resident investors may not fully understand the operation of the market. Local banks, in particular foreign banks operating in the local market, have not actively promoted government securities to non-residents, nor does the government have a strategy or systematic mechanisms for managing relations with non-resident investors. The government securities market handbook being prepared by SBP should be helpful in improving the accessibility of information about the market. In addition, the absence of Pakistan from local-currency bond indices may mean missing out on some “automatic� demand that inclusion in indices would induce. There is no short-term solution for this issue, as it depends on the inclusion criteria set by the maintainers of various indices. 26. Until recently, the taxation treatment of income from government securities was problematic for non-resident investors (and, indeed, some resident investors). This is being mitigated through the measures announced with the budget for 2010/11, which would impose a withholding tax of 10 percent as a full and final tax liability. 9 Recommendations • Consider, as a medium-term measure, developing an investor-relations function to coordinate all market-facing communication about government securities, promote these securities to resident and non-resident investors, and ensure the timely dissemination of information about the government’s issuance program. F. Islamic Securities 27. There is strong demand for government-issued shariah-compliant securities. Although Islamic banking is still relatively small in Pakistan, amounting to approximately 5 percent of the total banking market, it is growing. The sector already has excess liquidity. 28. Demand for sukuks has been strong. At the most recent auction, in September 2009, for example, the government offered PKR 10 billion, while bids totaling PKR 30 billion were received. A total of PKR 14.2 billion was issued, the maximum possible given the available hypothecated underlying asset. Sukuk outstandings as at end-May 2010 were PKR 42.2 billion, or roughly 8 percent the size of the stock of PIBs. The profit rates on the sukuks issued to-date have been lower than the yields on comparable 3-year PIBs. 29. The Islamic banks indicate that they could easily absorb an additional PKR 30 billion to PKR 40 billion of issuance of sukuks. Their preference is strongly for short-term paper, however, ideally a shariah-compliant MTB-equivalent. This suggests that there is opportunity for the government to increase its total issuance of conventional and Islamic securities cost- effectively. The conventional and Islamic segments are not completely separate, however. Currently, Islamic banks invest much of their excess liquidity in shariah-compliant deposits at conventional banks, which in turn invest their excess liquidity in MTBs. Although the offset effect would be unlikely to be one-to-one, increasing sukuk issuance would be expected to reduce the capacity, at the margin, of conventional banks to absorb government securities. The primary challenge for the government in increasing sukuk issuance is to identify suitable assets to hypothecate. 30. There is no primary dealer system for sukuks. All banks, Islamic or commercial are eligible to participate in auctions. There is practically no secondary trading, although one Islamic bank reported that it would be prepared to quote indicative two-way prices on the Bloomberg platform. Recommendations • Continue to identify suitable assets to hypothecate for future shariah-compliant issuance. • Increase the volume of annual shariah-compliant issuance at a gradual pace, so as to minimize the potential of increasing market fragmentation. 10 ANNEX: OVERVIEW OF THE STEPS IN PREPARING A DEBT MANAGEMENT STRATEGY Formulating a medium-term debt management strategy involves eight steps. This annex summarizes the key elements of each step. Step 1: Identify the objectives for public debt management and the scope of the strategy The purpose of this step is to clarify what objectives the strategy should seek to achieve. This will also help clarify the tasks and responsibilities for which specific debt management entities are accountable. • Identify the primary objective(s), as well as any subsidiary objective(s), for public debt management. Usual objectives include (a) satisfy the government’s financing need; (b) minimize cost; (c) maintain risk at a prudent level; and (d) develop the domestic debt market. • Ensure that the objective(s), where not set down in law, is (are) properly documented. • Define the scope of the strategy. In that regard, considerations include (a) central government, general government, or the wider public sector; and (b) which contingent liabilities. Step 2: Identify the current debt management strategy and cost and risk of existing debt The purpose of this step is to determine clearly the starting position for the analysis. This will help identify whether the strategy should seek to change the characteristics of the existing debt portfolio in a specific way, for example, to reduce a specific risk. • Explicitly identify the current strategy. This provides a benchmark against which alternatives can be evaluated. The current strategy may be a de facto, as opposed to formal, strategy. • Identify outstanding debt and its composition. This is necessary for determining the debt- servicing profile of outstanding debt. • Calculate basic cost and risk indicators for the portfolio. This should also identify sources of vulnerability to the existing debt stock. Step 3: Identify and analyze potential funding sources, including cost and risk characteristics The purpose of this step is to determine the range of possible strategies that may be feasible and desirable to implement. This will also help identify any potential constraints that may impede the implementation of a chosen strategy. • Identify potential sources of funding, their financial characteristics (including cost and risk parameters), and potential amounts available. To that end, (a) list existing and 11 potential instruments, domestic and external, and describe their financial characteristics; (b) evaluate the potential quantum of borrowing available through each instrument; (c) identify any constraints that may impede the availability of funding; and (d) discuss and rank the instruments, based on their cost/risk characteristics and within the context of the vulnerabilities in the debt portfolio previously identified. Step 4: Identify baseline projections and risk in key policy areas (fiscal, monetary, external, and market) The purpose of this step is to determine the baseline scenario for the analysis of the performance of alternative strategies and identify specific risk scenarios to be evaluated. This requires interaction with the fiscal policy, monetary policy, and financial-market authorities, and (where relevant) market participants. • Identify the baseline medium-term projections for key fiscal and monetary policy variables. The projections should be consistent with any debt sustainability analysis also undertaken. • Identify whether there are any external constraints relevant for formulation of the strategy. In particular, (a) discuss any anticipated change in the exchange-rate or capital account regime; and (b) discuss any required financing of international reserves. • Identify the baseline medium-term projections for market rates. • Clarify assumptions about likely pricing of non-market instruments. • Determine specific risk scenarios. These may draw on scenarios identified in the debt sustainability analysis, as well as other specific changes in terms of market conditions and demand (such as a shock of global liquidity). Step 5: Review longer-term structural factors The purpose of this step is to take a longer-term perspective and identify factors that could influence how the debt composition should ideally change over the longer-term. This requires interaction with the fiscal and monetary policy authorities. • Set out long-run structural features of the economy that the strategy should try to take into account. These may include, for example, (a) access to concessional financing; (b) trends in the real effective exchange rate; (c) inflation trends; and (d) commodity- price vulnerability. 12 Step 6: Assess and rank alternative debt management strategies on the basis of the cost/risk trade-off The purpose of this step is to analyze a number of alternative debt management strategies, assess their performance, and identify a small number of candidate strategies, including a preferred strategy. • Make a systematic assessment of a range of specific strategies. In particular, (a) assess how costs could change under the various risk scenarios; (b) assess how well each strategy helps mitigate the identified portfolio vulnerabilities; (c) assess how well each strategy meets the debt management objective(s), both primary and subsidiary; and (d) assess whether each strategy would be feasible to implement given assumptions about potential sources of financing. Step 7: Review the implications of the candidate strategies with the fiscal and monetary policy authorities, and for market development The purpose of this step is to determine clearly that the candidate strategies are consistent with sound fiscal and monetary policies and with maintaining debt sustainability, and are also in line with plans for market development. • Review the candidate strategies to the fiscal and monetary policy authorities. This involves (a) discussing any points of interaction; and (b) confirming that there is a positive impact on debt sustainability. • Review the potential debt market implications of the candidate strategies. Step 8: Propose and approve the strategy The purpose of this step is to propose the preferred strategy to the appropriate policymaker and secure formal agreement. • Document the preferred strategy. It is essential to (a) outline why the preferred strategy is superior to the alternative candidate strategies; and (b) describe the key costs and risks associated with the preferred strategy and its relationship with the debt management objective(s). • Present the strategy proposal to the responsible policymaker and secure agreement. • Disseminate the approved strategy. 13 Recommended Format of the Debt Management Strategy Document The minimum content of the debt management strategy document includes the following: • Objectives and scope: This section describes the objective(s) for debt management, the scope of the strategy, and the types of risks being managed under the strategy. • Existing debt portfolio: This section provides the historical context for the debt portfolio, describing changes in its size (including relative to GDP and revenue) and composition through time. Changes in relevant market variables should be included, along with commentary on significant events in the evolution of the debt. • Environment for debt management going forward: This section describes the environment for debt management in future, including fiscal and debt projections, assumptions about exchange rates and interest rates, and constraints on portfolio choice, including those related to market development and the implementation of monetary policy. • Strategy to be followed: This section sets out the preferred strategy and describes the analysis undertaken to support it. The assumptions used in, and the limitations of, the analysis should be made clear. At a minimum, this section should (a) describe the desired debt composition and the core arguments for that composition; (b) discuss the key risk factors that influenced the choice of the strategy; (c) describe progress to be made toward the desired composition over a three- to five-year planning horizon, with ranges specified for the key risk indicators of the portfolio and associated financing program; and (d) outline any specific measures or projects that are planned managing non-quantifiable risks or in support of debt market development. • Review and follow-up: This section outlines the periodic review process that will apply to check whether key assumptions continue to hold and that the strategy remains appropriate. The process that would be followed if circumstances were to change significantly outside that regular review cycle should also be identified. 14