@scooby. I suppose we need to see improved liquidity across the entire bond universe to decide whether prices are market driven. There's an assertion from the NSE itself that about 70% of prices on the market are off market.

This has been a key problem for the long tenored portion of the yield curve that has seen low volumes or recent primary issues so a majority of market perception are guess-estimates.

There's no benchmark which leads me to believe that AAs benchmark is neither here or there. I talked to AA recently and they use short end on the run issues and long end off the run issues. Six, seven year yields are extrapolated.

As you mentioned the HF bond is priced based on a comparable tenor, but that in itself can be misleading as it's difficult to understand how the NSE derives it's yield curve. Bloomberg also have a yield curve derived from a daily market survey (blind leading the blind).

The only thing i like about the AA benchmark is that clean price is tracked using face value as a reference. So it's easier to track the performance of the market (rather than changes in the yield curve). According to AA the clean price was 110 at the end of 2010 and is now 89. Can we compare this to the NSE index return?

Emlyn,

I do agree with you that a passive strategy is not designed to generate alpha. And one of the ways that an investor can pursue a passive strategy is to engage in buy and hold strategy.

But there are a couple of things that I need to clarify for you, if you don’t mind.

In order to make sure that an investor’s portfolio is indeed a passive portfolio, an investor needs to compare the features of their passive portfolio with that of a benchmark portfolio.

Here are some of the features that I would look out for when reviewing a fixed income portfolio in the Kenyan context

Firstly, the weighted average duration of the portfolio vs. the benchmark. If your portfolio has a larger duration than the benchmark, there is a higher chance that your portfolio has greater fluctuations when interest rates fluctuate like the current interest rate environment which means that your tracking risk is high

Secondly is sector exposure i.e. what percentage of your portfolio is made up of corporate or treasury securities. A higher exposure for a passive portfolio than for a benchmark would mean that the investor has a greater credit risk exposure hence greater risk of defaults

Lastly is whether the passive portfolio is a barbell or bullet portfolio as compared to the benchmark. This is to evaluate the effects of a flattening or steeping of the yield curve as is the case in Kenya right now – the yield curve is inverted.

So, while you are “buying and holding” your portfolio, its characteristics could be materially different from the benchmark, thereby increasing your tracking risk.

FYI, “buy and hold” was a termed coined by accountants when they were trying to understand how financial institutions manage their investment portfolios. For instance, if a bank has a long term liability that is payable, say after ten years.

What it would ideally do (as part of its asset liability matching process) is to purchase investments with a similar duration so that that investment matures at the same time as when the liability becomes due.

Hope this helps and do enjoy your Christmas holiday.

Regards

@scooby. Nope wasn't asking that. I'm asking two questions. Can you track the performance of a bond market by looking at the weighted average clean price of a portfolio? Can the NSE yield constitute a market benchmark?

I believe NSE do have a minimum of 50 million (I think) and observe prices over a two week period. They say they ignore any price 300bps below/above market, but have no benchmark yield curve to determine where market is (so it makes little sense). AA does have a threshold, 50 million, where prices are observed over a week.

Seems a lot of what NSE says is a market driven price is very subjective (i.e., depending on how they want their curve to look). You'll notice many of the traded bond prices on the NSE have erroneous/(wrong) clean prices nway, and the level of scrutiny of this transactions is minimal on the part of the NSE.

Nway my friend have a merry merry Christmas

Scubidu,

You can track the performance over time by looking at movement in clean prices. Am wondering how you would determine the weighted average clean price...is it by the size of the issue?

And for the yield curve to constitute a benchmark, NSE should disclose which bonds they used to derive the yield curve from. I guess that what you are alluding to with your second query, right?

Regards

@scooby. Yes using the weighted average size of the issue. And yes the NSE should disclose the on-the-run issues and off-the-run estimates in deriving the yield curve. I talked to a guy whose seen their model and it has a lot of errors.

So theoretically the market only has two workable yield curves, NSE and Bloomberg and surprisingly they have a 300bps spread between them from the intermediate to long end. That's a recipe for massive mispricing if you ask me. CBK relies totally on the NSE yield curve for discounting.