MART Position Paper – Philippine Yield Curves

This paper aims to establish the urgent need to adjust the methodology for the construction of the standard yield curve in the Philippines. In particular we recommend prioritizing the use of market derived yields using the most liquid benchmark securities along the yield curve as opposed to the current methodology where the curve is often interpolated from illiquid benchmark bonds.
Position Paper on Philippine Yield Curves

BACKGROUND AND ISSUES

Consistent with most risk-free yield curves globally, the Philippine risk-free yield curve is constructed out of the secondary market yields of government securities as traded in the market and collated by PDEX. The current curve construction involves the capture of market rates in what are defined as “benchmark maturities” which run from 1 month to 25 years. Yields in between these benchmarks are then derived through straight-line interpolation to complete the yield curve. Unfortunately, most of these benchmarks have not traded in months, even years, rendering them irrelevant to the market. Yet the market is required to price them as benchmarks and the rest of the curve is priced off them. Though this is not always the situation, it has been observed that this is often the case. The aberration exists when the more traded liquid benchmark bonds, through time, are no longer classified as such and are replaced by other issued bonds which are not liquid.

The table above illustrates the capture of the defined benchmark bonds on the 13^th of January 2011.  Most of these bonds have not been dealt in the market for days, weeks, months or even years.  The yields are often not reflective of where most bonds trade.

Adjunct to this issue is the fact that most liquid bonds in the market are disappearing due to non-replacement or non-issuance. The issue is most pronounced in the short-end of the yield curve where Treasury Bills have been in critical short supply thus creating distortion in pricing as well as limiting the price discovery process. This is a matter that can only be addressed by government as it is the only risk free issuer and cannot be supplanted by another entity. Though government may not have any need for financing in this part of the curve, it assumes another pivotal role in developing the capital markets in providing for the development of a default-risk free benchmark yield curve.

The benchmark in the short-end is not assisted by the fact that there is practically no interbank money market to speak of. Other jurisdictions would normally refer to interbank money-market benchmarks such as libor to provide the basis for the short end of the yield curve. In the Philippines, despite several attempts to build such a benchmark, an interbank money market continues to elude us which in itself shall be the subject of another paper. At best, the need for a short term benchmark is partially being addressed by PHIREF which is the implied peso rate derived from the FX forward points curve. This however, despite being the best available solution, is far from ideal in building a proper benchmark yield curve.

MARKET NEEDS

The urgent need for a realistic market determined yield curve cannot be underscored enough as it serves as basis for pricing a whole slew of financial products from loans to derivatives. It is the lynchpin for developing a robust and efficient debt capital market. It also paves the way for access to longer term capital.

The current shortcomings in the Philippine yield curve needs to be addressed immediately. Though not inherently wrong, it is lacking in information critical to the proper valuations of bonds not classified as benchmark, which as mentioned earlier, can generate a big disparity when interpolated rates are compared with secondary market rates. For one, we have a relatively large and fast growing bond market that needs to be properly valued. Earlier we mentioned that the current market process of interpolation is creating a relatively huge distortive effect on certain parts of the curve. Being a straight-line process, this distortion tends to change depending on where the current liquid issues lie along the curve and the twists in the shape of the curve. This has translated into significant variations in bond valuations in the past where portfolios are either over or under valued. Unfortunately, because of the abovementioned methodology, it is also the most liquid bonds which are affected by this distortion.

The RPGB 6.24 1/14 or series 5-67 for example, which at one time was the most liquid of bonds outstanding (enjoying peak volume turnovers of as much as Php35.6bio in one day), had an observed variation of as much as 20bps from the actual market. That would have translated to approximately Php202.4mio of over/under valuation for that one trading day for the whole market.

The impact is serious given that the largest holders of bonds, i.e. banks, insurance companies, and funds, need public disclosure of their balance sheet valuations. A classic example would be that of a bond fund that calculates its daily NAV based on the distorted curve. Should the NAV be overvalued which leads to a unit holder selling the fund, the cost of the overvaluation would be left with the remaining fund unit holders. Should the fund itself be liquidated, the final proceeds could be significantly off from the valuation on paper. This is also true with foreign investors with regard to the valuation of their onshore debt investments.
Another issue arising from this is that it puts into question market transparency. Questions normally arise from investors as to why the posted values are not reflective of where the actual market trades. At time of execution, customers are perplexed by the phenomenon and as they are made to understand the distortion they will tend to disregard the yield curve altogether. The implications on the market transparency initiatives need not be stated.

PROPOSAL

The issues are multiple and will require action from several fronts. Our proposal begins with the most fundamental aspect of yield curve construction which is to derive market based yields along the yield curve. As observed, there are many gaps in the Philippine yield curve due to various reasons discussed. Where there is no market activity, the market will look to government to provide the impetus to develop that segment of the market. Demand must be established for a market to evolve and eventually a benchmark. This is the experience of many if not most jurisdictions globally. To achieve this, we have various recommendations in terms of policy and infrastructure which in combination should lead to a well developed yield curve.

q The first recommendation is for the national government to have a consistent auction schedule covering most parts of the desired curve. Government stimulated demand will lead to the creation of a market based price discovery process around those tenor points where there may currently be none. It also sets a good precedent as the evolving market will translate to that of a default-risk free benchmark.

q Corollary to consistent issuance is a strengthening of the reissuance process and an enhancement of the switch programs with the primary objective of maintaining benchmark issues. This has been one of the most effective government initiated programs in other jurisdictions which contributed significantly to the depth of benchmark bonds thus leading to a more robust yield curve.

q To assist in generating market activity and depth, focus must be given to developing efficient and simplified financial instruments. Among those contemplated are the development of an efficient interbank repo market, interest rate forwards market, bond forwards market if not a bond futures market. These products merit individual studies and position papers not contemplated in this paper, but the development of these markets are never the less considered critical and should be prioritized by both regulators and market stakeholders. Crucial to its success though is the reduction to the barest minimum transaction costs which may involve further involvement of Congress.

Noting that the foregoing needs time to completely achieve, and that delivery will likely come in stages, we are confronted by the question of how best to represent the existing yield curve for the markets to continue functioning well.

To this end, we recommend to immediately prioritize the use of market derived yields using the most liquid benchmark securities along the yield curve. That is, we shift away from tenor based yield curve construction to a securities based methodology. This should address the immediate issues arising from valuation which have impacted market economics as well as financial/regulatory reporting and transparency. We suggest defining such securities in accordance with the existing PDEX definitions for “liquid” bonds with the end goal of establishing a bond curve that is as close to the market as possible.
Understandably, there are portions of the yield curve which may need to be supplanted by other sources. The main body of the curve though is best derived from the risk free yields provided by government bonds.

From this curve, other curves can easily be derived as the need arises and as the infrastructure develops. Various curves can be derived to suit various needs. A tenor based curve, for example, can be derived from this market based spot curve using a variety of methodologies. The development of several yield curves is an eventual possibility which brings with it the welcome sign of improving sophistication. Eventually, curves can be established to suit more specific needs to address the multitude of valuation concerns that the market will develop. This can happen if we can manage to develop a robust credible risk-free curve to act as basis for everything else.

We trust that the above has provided you a better appreciation of the issues surrounding the local debt market.