This week's Learning Curvedeals with the types of market participant, geographic variations and two of the most popular instruments: zero-coupon swaps and year-on-year swaps. Next week's article will look at real annuities and real-coupon swaps.
According to estimates by ICAP the market in derivatives linked to inflation has grown more than 10-fold in the last two years.
There are six main categories of market participant:
* Liability managers;
* Holders of inflation-linked revenue steams (e.g. property developers, project finance vehicles);
* Issuers;
* Asset swap investors;
* Retail business; and
* Leveraged investors
Some Common Instruments
The most common products traded in the inflation-linked derivatives market are various types of swap. Commonly, one side of the trade has a dependence on a defined index and the other is a standard fixed or floating nominal stream of cash-flows. Trading is now well-established in the euro area, which is the largest market, the U.S. and the U.K. The smaller inflation markets, such as Sweden and Japan, support a proportionately less significant volume of derivative trades.
Market liquidity has improved dramatically in the major markets over the past year, but still remains dependent on the choice of index, maturity, size and complexity.
Trade maturity is one of the most important considerations for two reasons. The first, more technical, reason is the risk profile for the inflation curve is rather different to that of a yield curve. The volatility of index fixings and the vagaries of seasonality affect short-dated trades much more than longer maturities. Additionally, the paucity and illiquidity of short-dated inflation linked instruments in most markets means the inflation curve can be somewhat ill-defined in that part of the curve. Also relevant are the economic characteristics of individual markets. For example, Europe is dominated by retail activity in the five to 10 year sector while long-dated--20 years and more--liability hedging and project finance deals dictate activity in the U.K. Consequently, the concentration of liquidity is quite different in the two markets.
The most liquid market is in Europe, in the HICP ex tobacco index. There bid-offer spreads in the five to 10 year sector have compressed to 4 basis points at most for sizes up to EUR100 million, with tighter prices often available if a dealer has a specific interest. Longer dates are a little less liquid, but should still trade no more than 2-3 basis points from mid-market. U.K. and U.S. swaps will generally trade wider because there is somewhat less volume in these markets.
Construction of a particular instrument is dependent on adequate definition of the future payments. In this respect, the most important additional parameters of an inflation swap vis-à-vis a standard interest-rate swap are the choice of index and the timing of data readings.
The most common indices are those used by the bond market: Euro HICP excluding tobacco, French CPI excluding tobacco, U.K. all-item RPI, or U.S. Urban CPI NSA. However, there are active markets in several other indices and end-users should consider to which index their exposures lie.
The timing of data readings is important for settlement purposes. Data are usually fixed some time before the cash payments are settled in order to avoid confusion or delay. For example, the market in euro-area HICP conventionally fixes swaps with a three month lag between the inflation reading and payment. Some structures may require an interpolation of the published data, rather than a fixing directly from a published number.
Within this broad framework, there is considerable room for invention. However, most swap deals are based on four basic constructions.
Zero-Coupon Swaps
The simplest, most straightforward structure is the zero-coupon swap. Box 1 illustrates an example of a zero coupon swap on euro-area HICP excluding tobacco. The most common users are in the inter-dealer market (where it is the standard structure) or sophisticated ALM clients and retail banks who can manage their exposures through diversified portfolios of such trades. They are, of course, operationally rather simple, consisting of only two cash-flows. Note the lagged definition of the index. For example, if swap start is in October and swap end is in October 2009, then we might conventionally use a three month lag (for most euro trades) and fix the index data in July and July 2009 respectively. Often the index is defined in conjunction with an agreed reference page--for example Reuters page OATEI01.
Box 1: Zero Coupon Swap on Euro HICP ex tobacco | |
Trade Date | 11th October 2004 |
Swap Start | 13th October 2004 |
Swap End | 13th October 2009 (5 years) |
Notional Amount | EUR50,000,000 |
Payer of fixed | Party A |
rate coupon | |
Party A Pays | ((1+ X%) ^ 5 – 1) x € 50,000,000 |
Coupon Payment | [Swap End subject to Modified Following Business |
Date | Day Convention] |
Payer of Inflation | Party B |
Coupon | |
Party B Pays | ( (INDEX(End) / INDEX(Start)) -1) x EUR50 million |
Inflation Lag | 3 months |
INDEX(Start) | Euro HICP ex Tobacco unrevised published for the |
month 3 months prior to Swap Start | |
INDEX(End) | Euro HICP ex Tobacco unrevised published for the |
month 3 months prior to Swap End | |
Coupon Payment | [swap end subject to Modified Following Business |
Date | Day Convention] |
Source: Barclays Capital |
Year-On-Year Swaps
This structure (see Box 2) is generally what the novice user of inflation swaps might imagine to be the standard structure. Annual year-on-year inflation rates are exchanged for fixed cash-flows. In the example, the index employed is French CPI ex tobacco. Such structures are often traded as a hedge for retail products paying an inflation-linked income.
The usual convention for all French inflation swaps is to mirror the behaviour of the bonds and use an interpolated value of the index between the two month and three month lagged values. The interpolation is a simple linear calculation, but in any case the Agency France Trésor publish the daily results of the calculation on both Bloomberg and Reuters.
Box 2: Year-On-Year Swap on French CPI ex tobacco | |
Trade Date | 11th October 2004 |
Swap Start | 13th October 2004 |
Swap End | 13th October 2014 (10 years) |
Notional Amount | EUR100 million |
Payer of fixed rate | Party A |
coupon | |
Party A Pays | X% x EUR100 million |
Coupon Payment Dates | Annually from Swap Start to Swap End subject to |
Modified Following Business Day Convention | |
Payer of Inflation | Party B |
Coupon | |
Party B Pays (in Year n) | ( (INDEX(Year n) / INDEX(Year n-1)) -1) x EUR100 million |
Inflation Lag | [Conventionally three months interpolated] |
Coupon Payment Dates | Annually from Swap Start to Swap End subject to |
Modified Following Business Day Convention | |
INDEX(Year(n)) | Interpolated French CPI published for the date three |
months prior to the payment date in year n. | |
Source: Barclays Capital |
This week's Learning Curve was written by Joe Mulvey, a senior trader in European inflation derivatives at Barclays Capitalin London.