Australian Government Debt Future Options for CGS Market Development – Presentation to the ABE Forum

4 June 2013

Speech delivered by AOFM’s CEO, Rob Nicholl, to the ABE Forum

Sydney 4 June 2013

Thank you to the ABE for inviting me back. I feel I am fast running out of different perspectives from which to cover the issuance program for the year ahead and an outlook of development of the market as we look to the coming years (to the extent of course that I can feel comfortable giving one).

Last year I spent a considerable part of the presentation covering the perception of what would happen if the CGS market were to shrink rather quickly following a return to fiscal surplus and whether liquidity would suffer appreciably in the event this was to occur. We had prior to that time heard comment about the need for the Government to continue issuing CGS at a rate that would leave investors feeling confident that their continued participation was warranted on the basis of relative value, liquidity and the general resilience of the market. In this regard their hopes regarding supply will have been met, at least for another year or two.

I also covered to some extent the issue of offshore demand for our paper because this was very topical at the time, and in fact the offshore bid has played a very important role in the changes that the CGS market has undergone over the past four years or so. We know from our discussions with central bank investors in particular that the factors among the most important influences on their decision to allocate into $AUD, and into CGS are:

  • safe haven or highly rated currencies and their respective sovereign bond markets offer the greatest probability of preserving capital – this is a critical consideration for the bulk of their reserves;
  • both the currency and the CGS market offer very good liquidity; and
  • the $AUD and CGS markets have provided an attractive diversification alternative, although this tends to be more of a secondary consideration against capital preservation and liquidity.

These considerations remain important and we are aware that there are certain things we do as the issuer that help to promote a market that offers the liquidity they are seeking. Not surprisingly we continue to focus a lot of our discussion with investors on such issues.

This year I thought it would be useful to spend some time highlighting what we see as the main issues surrounding further extensions of the CGS yield curve. The possibility for this is either imminent or some way off depending on how you view the role of the CGS market and what it may look like in terms of its size over the forthcoming years. And as I know you will expect some comment from me on issuance for the coming year, I will happily provide that, as well using some time toward the end to make some comments about RMBS.

Market Development – extending the yield curve

Over the past two years in particular there has been growing speculation about whether the AOFM should be seeking to further lengthen the yield curve. While the market for CGS in terms of outright size was small by global standards up until around 2011, there simply wasn’t the supply of bonds available to build many deep and liquid lines. With outstandings of less than $60 billion during the years leading into the GFC, this barely allowed for issuance to maintain support for the 3 and 10 years futures baskets. Thus, a yield curve of around 12 years on average was about all that could have been expected.

Now we have a market that is about 5-times larger than it was prior to 2008 and this has allowed for an extension of the yield curve to around 15 years, while at the same time establishing maturities for almost all of the intervening years. At the same time, most lines have been built to well in excess of $10 billion (which is typically considered sufficient to start generating very good liquidity).

Extending the curve to around 15 years was done for portfolio management reasons and it allowed us to establish sufficient issuance points to begin increasing overall duration of the portfolio. While we do not target duration as such, we noted a fall in duration as total issuance increased substantially with a shorter curve. Over the last two years we have appreciably increased the average tenor of our issuance from just less than 6 years to it being just less than 9 years – as of last week. The weighted average tenor of the stock of nominal debt outstanding has gone from 4.6 years to 5.2 years over the same period. We have observed our investors following us on this initiative as our tenders continue to be bid with strong coverage ratios and price outcomes. We also received strong bids for the 2027 and 2029 syndicated launches.

Prior to launching the 2027 bond line in 2011 we were unsure as to how longer dated CGS would be received by the market. This was particularly so given that over 70 per cent of our paper is held offshore by central banks, which on average tend to focus their buying on the short to mid-part of our yield curve. However we now know through our discussions with them that there are central banks and sovereign wealth funds that hold CGS across most if not the entire curve. Without doubt that bid, along with the offshore pension funds, fund managers and ongoing domestic interest, has been instrumental in supporting our curve extension. Australia’s triple-A credit status and low debt exposure relative to other advanced economies are no doubt important factors in securing an ongoing offshore bid.

We are also aware that there was broad support for us extending the yield curve because this helps pricing for semi-government and corporate paper. It stands to reason then that consideration of a further yield curve extension is a sensible question given, at least in principle, the opportunities for further increasing duration and the benefits of facilitating further development of the Australian fixed income market. Each of these, however, has separate considerations; those relating to pure portfolio management in our view being more difficult to resolve. It is these that I will outline shortly. On the other objectives, we do not see it as the role of the AOFM to be taking large leaps towards broader policy objectives without appropriate guidance from government that such objectives in their own right outweigh the potential costs and risks of doing this.

Let me begin this outline by providing you some background as to how we view the possible costs and risks associated with developing the portfolio. But before I do that I would make clear that any sovereign debt management office will necessarily be weighing the possibilities of establishing appropriate longer benchmark bond lines. These are questions we consider on an ongoing basis as part of our issuance and portfolio management responsibilities. But when it comes down to it, longer curve liquid bond markets will necessarily require a greater amount of outstanding stock and demand for the paper than will be the case for shorter yield curve markets – all other things equal.

We have a number of modelling approaches and guiding principles that assist us in making decisions about the shape of the portfolio and the planned issuance in any year. In essence we are seeking each year to develop a financing strategy based on a mix of instrument types and an associated maturity composition of the debt to be issued that will best support the Government’s funding requirements for the period. At the same time we are seeking to develop a portfolio based on a balance between the costs of achieving that funding task and the risks associated with alternative portfolio structures, whether the risks are volatility in financing costs or refinancing risk.

Determining an appropriate balance between these broad possible impacts requires considerable judgement. If the Government’s financing requirements were known with certainty long in advance, together with how the yield curve and inflation would evolve overtime, selecting an optimal funding strategy each year would be a straightforward task. Obviously this is not the case with uncertainty about the future an inherent part of all financial market activity. Macroeconomic conditions, policy decisions and debt servicing costs are all variables that can shift appreciably through time.

Cost is perhaps the easier metric to define and for this we use ‘accrual’ debt servicing costs. We do this because it is closely related to the measure of ‘Public Debt Interest’ which is the basis on which debt servicing costs are recorded in the Budget papers. We do not use a mark-to-market measure because the debt is unlikely to be bought back prior to maturity.

Risk on the other hand while relatively easy to define, is a far more difficult metric to measure. We define it in terms of the variability of the cost I have just described, the logic underpinning this being that variability in debt servicing costs will spill over into Budget volatility – all other things equal. It also equates to uncertainty about future funding costs. However, neither of these is easily quantified.

Putting these together gives the following points to note:

  • First – as our time horizon for consideration is extended, so too is the variability of funding cost outcomes for a given financing strategy;
  • Second – if yields were to rise appreciably above the historic lows we have been experiencing, average portfolio costs will decline while existing debt issued over the past decade at even higher rates is refinanced at these still relatively low yields, but will eventually begin to rise; and
  • Third – if the AOFM were only interested in funding cost performance of the portfolio in the short-term, we would be issuing with a bias toward shorter rather than longer dated paper, given the slope of the yield curve.

From our modelling framework we are clear that as the duration of the portfolio is lengthened (regardless of the outright level of yield over the prevailing period) risk, or cost variability, is expected to be lower. At the same time costs will be higher, assuming that on average we will face an upward sloping yield curve. The more that longer dated paper is issued overtime, the less churn there is in repricing that occurs in the portfolio, but this comes at the trade-off of higher relative yields compared with issuing shorter dated paper.

The mix between nominal and inflation indexed bonds adds another dimension for consideration because the greater the proportion of TIBs issued overtime, the greater is the proportion of the portfolio exposed to inflation, which represents an additional risk factor over and above repricing risk. In effect this gives another dimension to the risk associated with cost variability. In an environment with a benign inflation outlook this may not be of significance; however where a benign inflation setting is not the case the case the costs associated with this part of the portfolio could be appreciable. While the opportunity to issue inflation indexed bonds is generally considered important for diversifying the issuance task and the investor base, this risk factor is one reason why TIBs tend to be less than about 15 per cent of the portfolio in those sovereign debt markets in which they are issued. Limited overall demand relative to nominal bonds is also another key factor.

With these considerations about the overall costs and risks of managing the portfolio in mind we can look at the key factors guiding our current thinking on whether to further extend the yield curve in the near future, purely on the basis of portfolio management considerations.

In considering the options open to us we are interested in the possible future size of the market given that a larger market will tend to accommodate a wider range of options than will a smaller market. Based on the current Budget outlook the size of the CGS market relative to GDP is not in broad terms expected to be much different from what we see now.

Let’s put to one side for the moment the issue of whether there would be sufficient demand for longer dated CGS, other than to say that we would need to be confident (through market soundings of some type) that longer dated maturities could be established through one of either a competitive tender or a syndicated launch. We see issuance via private placements, as not being consistent with the transparency of operation investors have come to expect from us.

In terms of the length of the curve we wouldn’t see any particular reason to rule out the future establishment of a 20-year, 25-year or 30-year maturity. The choices essentially come down to an expectation as to which lines could prove more in demand – something that is difficult to gauge accurately prior to gaining firsthand experience.

Once again we understand the policy reasons for wanting to consider lengthening the yield curve (such as for example to facilitate long-term infrastructure development) but I will look more specifically at them through the prism of debt portfolio management. It should go without saying that a happy coincidence between the portfolio management and policy reasons for extending the curve further makes for more easier decision making!

Let’s say for example the objective was to establish a 25-year maturity. There are essentially two ways to achieve this – one is to build the curve out incrementally towards the objective maturity. The other would be to launch a 25-year maturity and then consider intermediate lines – or tapping the 25-year maturity as it shortened overtime. The next questions would be: how long after issuing the longest bond would we repeat the exercise (say every two to five years?); and how often would we tap the new line. Both of these questions really go to how well the bond traded after it was established.

If there were sufficient indicated demand, it would probably be possible to tap the line say once or twice a year through tender – having probably left it for a year untouched after establishment. The question of how often to repeat the issue of a longest bond goes to how many intermediate lines we would envisage as useful, but in many respects this is only a marginal consideration. This is because under present circumstances, we wouldn’t foreshadow the ability to issue large volumes in the ultra-long end of the curve, and relating to that we also couldn’t imagine there being sufficient interest in numerous ultra-long CGS maturities.

The more conservative approach to establishing a 25-year maturity would be to issue a 20-year maturity and observe how that bond was received by the market. If it wasn’t received well we could avoid further curve extension; on the other hand if it was received well it would help the pricing of a 25-year maturity.

Another option to consider is whether to lengthen the nominal curve by first lengthening the TIB curve (which is what we have done before). TIB outstandings are substantially smaller than for nominal bonds so my earlier point about reduced options in a smaller market bears relevance here. But extending the TIB curve is certainly a possibility as we have already been discussing whether or not to introduce a longer TIB maturity with some of our investors.

Taking these options into account through the cost/risk framework I outlined earlier gives rise to the following considerations. Having established a number of new issuance points out to around 15-years we are comfortable that we can probably increase duration of the portfolio sufficiently through maintaining (or even further increasing) our current average tenor of issuance. From the perspective of cost, establishing longer dated maturity points carries the possibility of steepening the yield curve, particularly if the longer dated maturity first approach is taken. Difficulty in pricing with little or no other reference points other than spreads to other sovereigns (such as US Treasuries for example) and an illiquid swap market leaves open the possibility of bonds with yields that could in effect create higher yields for other intermediate maturities than would otherwise have been the case. This would obviously increase the overall costs to the portfolio compared to a situation where the longest bond hadn’t been issued. This is what we refer to as ‘contagion risk’ – that is where initial limited demand results in smaller issuance with higher yields. The ultra-long bonds create higher yields for shorter bonds as investors look to price other maturities from the longer rather that the shorter end of the yield curve. This unfortunately we will only discover from experience. A further risk with limited demand is that small bond lines can become very illiquid and not be actively traded – thus becoming what we refer to as an ‘orphan’. In this situation there is little or no pricing benefit to the rest of the bond market from having the line ‘out there’ and it risks creating a negative aspect to part of our yield curve.

In summary, I have raised a number of questions for consideration and I know have given little by way of definitive judgement about them. The point I wanted to make though is that extending the yield curve has a number of possible benefits, but they do not come without risk. From a portfolio management perspective we have been actively considering further extensions of the yield curve but we haven’t come as yet to a landing on how far that should be and by which curve – TIB or nominal – we should lead. Both have been under consideration but the timing remains unclear if we were to consider this for purely portfolio management reasons. In the meantime we see good consolidation of our nominal bond curve extension and consider this advantageous in preparation for any further extension. As to incrementally extending versus the long one-off approach, we are by nature a relatively conservative agency and would probably see establishing an ultra-long bond initially as a bridge too far without the imprimatur of government. That said, should we believe the circumstances and risks are aligned to do that, we would actively promote this, having first discussed it with our Board.

Issuance for 2013-14

I will now turn to the issuance program for the year ahead and will start by noting that we have no intention of deviating from the pattern of weekly Treasury Bond and monthly Treasury Indexed Bond tenders. This suits our operational requirements in terms of managing cash flow needs, remaining in touch with market conditions, and spreading the risk of issuance. At the same time we understand this to meet with investor appetite for regular supply.

The TIB market had a more buoyant feel to it throughout this year compared with last year, and we expect a building bid for our TIBs based on tender outcomes and market feedback. We remain conscious of the commitment in the Budget several years back to build the market to 10-15 per cent of outstanding bonds and although there was no express time frame set out for us to reach the 10 per cent threshold, we have been looking for opportunities to increase issuance (and in doing so increase the liquidity of existing lines) and to introduce new lines.

We have considered the feedback received over past months regarding further development of the TIB market and have responded by appreciably increasing our allocation to TIB issuance for next year – up to $4 billion. While we will need to increase our tender volumes in order to achieve the announced supply, we are also conscious of the benefit of using the syndication method to generate additional market interest and are considering the merits of building liquidity in existing lines by tapping one or more of them through a book build process, as well as considering what the next new maturity should be. I can’t imagine us being short of offers for assistance in this regard but our plans don’t envisage this happening over the next few months.

Our Operational Notice of the 15th of last month should be clear in terms of intended volumes for this year and next year. As to the possibility of establishing a new TIB line in the coming year, a choice is needed between something like a 2027, or a longer maturity, say a 2035 or 2040. The latter approach would suit demand for those investors seeking duration in TIBs – of which we seem to have quite a few. This would leave open the possibility of launching a 2027 when it was about to become a 10-year bond. On the other hand we have had enough indication from other investors that another intermediate line rather than a curve extension bond would also be welcome. We will continue to consider the merits of whether a new line should be established this year, and if so what maturity we should target.

Treasury Bonds:

We currently have 18 nominal bond lines with 13 of those having more than $10 billion on issue and 8 in excess of $15 billion on issue. Market feedback consistently indicates very good liquidity in nominal CGS, these large lines an important contribution to that. Now that we have extended the curve to 15 years we have had the capacity to create new issuance opportunities, while continuing to build existing lines to deep and liquid levels. Furthermore, as we are aligning the maturity dates of new bond lines with quarterly tax collections, this will assist overtime in managing maturities – thus allowing us to build our nominal bond lines to much higher levels than has been operationally prudent in the past.

Establishment of new lines over the past year with maturities of 2024 and 2025 have from our perspective been very successful. The 2024 has been in strong demand allowing us to build this quite quickly through our regular tender process. The 2025 launch a fortnight ago garnered a strong bid with the order book reaching over $6 billion, allowing us to issue $4 billion at the mid-point of our initial pricing guidance.

Our current thinking is that we will launch a new long bond line of around 16-years in every other year to meet the aim of maintaining the 15-year curve. We are confident that demand for bonds of this tenor will continue to increase – facilitating consideration of further yield curve extensions.

We maintain the belief that the strategy to lengthen the yield curve will continue to attract a new investor base of longer fixed income investors.

Treasury Notes:

Treasury Notes continue to be highly useful in absorbing the volatility in our cash position throughout the year. The volume of Treasury notes on issue will by their nature and use, vary over the year, but we have now abandoned our commitment to keeping a minimum of $10 billion on issue. We had originally aimed to keep $10 billion on issue to help promote liquidity, but on the basis of widespread feedback it is clear that this floor was not achieving that aim. During the year we announced a reduction in our reliance on short-term funding and this will mean there could be times when there are little or no Treasury Notes on issue, and by implication this also means that there may be periods during which we are not issuing Treasury Notes at all.

RMBS

Before closing let me make some comments on the RMBS program.

On the 10th of April this year, the Treasurer announced the end of the RMBS investment program. While the AOFM will not be making further purchases of RMBS, it remains the holder of a large portfolio of Australian prime RMBS. So, our job is not quite over. That said, I will add to the voice of a few commentators who have praised the AOFM for the manner in which the investment program was implemented and acknowledge the considered and appropriate judgement that my colleagues in the agency have exercised on a consistent basis since the program began back in late 2008.

To recap our involvement, over the 4.5 years that the investment phase of the program has been running, the AOFM has invested a total of $15.5 billion – this being across 111 tranches of 67 transactions from 20 programs. During this period the smaller lenders sponsoring these programs have been able to source more than $45 billion in funding. As a broad indicator the program has directly facilitated about 245,000 home loans, while at the same time indirectly supporting market lending of more than $3 billion to small businesses.

Of the $15.5 billion of RMBS purchased, we have had $5.6 billion repaid and have sold a further $634m, leaving us with $9.2 billion currently outstanding.

Portfolio amortisation is heavily dependent on the rate of repayment in the underlying home loans, as well as whether call options embedded in the transactions are exercised. Nevertheless, we expect the bulk of the investments we hold to have been repaid by 2018, although there are some longer term investments that will continue for longer. We remain confident about the quality of our RMBS assets and can comfortably hold these to maturity, given the strong performance of this asset class to date. However, we will continue to monitor performance of the portfolio closely to ensure we are abreast of any developments that could impact the performance of the portfolio.

Holding this stock to maturity is consistent with what we have done and have been expected to do thus far, but we will also continue to consider sale opportunities as and when they arise. In doing this we will remain fully cognisant of the potential influence that us selling RMBS parcels could have on the market. As noted in our last operational notice, we “will not entertain sales that undermine the improvements seen in the market to date”. This means we may, for instance, sell in the interests of price discovery and market transparency. As an example of this, through our last trade we were able to demonstrate strong demand and significant further spread compression for 2.5 year RMBS with margins at 75 basis points. I hasten to add that we did not saturate this demand with our trade.

Given the phase where we are at with the RMBS program and the forum here today, I thought it would also be interesting to quickly touch on the topic of “free market versus government intervention” that this live example offers us.

From the perspective of our role and from observations by others, we believe this program has largely been viewed as a success, not only in terms of facilitating continued supply/demand activity in such a way that has minimised market hindrance, but also because the number of issuers in the market remains the same at the conclusion of the investment part of the program as was the case at the start. The key to our success has been in fully understanding the market dynamics at play at every decision point and making appropriate investment decisions around it. It has been crucial to have the market expertise within government to achieve this. We have been very careful to allow the free market to reign when that has been an option, and we have been ready and willing to step in as and when required in an appropriate manner.

By rolling out the program in a way as naturally aligned to market practice as possible, we believe we have had little to no dampening impact on the appetite of a market free of government intervention. Securitisation, and in particular, prime RMBS issuance, has continued to function here in Australia during this tumultuous period to the envy of offshore. As investors now emerge, they are finding a smaller world of credit investment alternatives, amongst which our market has retained its good track record and continues to offer relative value.

From almost any angle it could be argued that this targeted, timely but temporary approach has proven a worthy intervention.

Thank you

Last updated: 6 November 2013