Monday, February 05, 2007

Can't get enough of TIPS

A quick comment on why I started this blog. First, I have a gigantic ego, so I think that my opinions and commentary are worth other people reading. Second, I have an insatiable desire to do things myself and/or create new things. By this I mean, when I started reading investment blogs regularly, I rapidly concluded that I had to start one myself. Third, I didn't know of any blogs that focused entirely on the bond market from a trading perspective, and the blog world is all about specialization.

But probably most of all, I thought doing this would challenge my thinking. Force me to prove my ideas to such an extent that I could convince my readers. I knew that my blog would attract mostly pros, students and serious investors; who else would read this arcane stuff? So I knew if I were going to attract a loyal readership, I'd have to produce quality posts with well-reasoned arguments. I knew doing so would make me a better trader.

So thanks to all the people who have posted comments on my recent pieces about TIPS. For those who haven't already, I encourage you to go back to these posts (here, then here, then here) and see the argument unfold. We don't get many comments on this blog, which probably owes to the type of material I write about and the people who read the blog. But most of the comments we do get are in-depth and thought provoking, and that's what it should be all about.

Anyway, here are some additional comments about TIPS. I can't say they are "final" comments, because this is a debate that will continue to rage on regardless of what I say here.

1) TIPS will obviously outperform Treasuries if inflation is higher than is currently priced in over the term of the bond. Therefore if TIPS and Treasuries are your only choices, and your inflation forecast is greater than the spread between the two, buy the TIP.

2) Bear in mind that food and energy account for most of the variance in inflation figures. Therefore in order for your inflation forecast to be correct, you better have a good forecast of oil prices.

3) As I have stated a couple times, I invest based not on predictions but probabilities. I can't get to a place where my forecast for inflation is definitive enough to invest in TIPS given current spreads and given the analysis I did here. Particularly since I believe the Fed is pretty good at fighting inflation, and any spike in CPI that is not related to oil prices will be beaten down by tighter monetary policy.

4) I believe there is a "special" bid for TIPS from asset/liability matchers who have CPI-based liabilities. I think this results in TIPS being more expensive than would be in a theoretical efficient market where all agents are motivated solely by profits.

5) I don't believe in the blanket statement that any asset class with a less-than-1 correlation should be included in a portfolio. Not only is this practically impossible, but there are many types of bonds which exist to meet specific needs and/or have a "special" bid. The most egregious example would be tax-exempt municipal bonds. Introducing these bonds into a non-tax paying portfolio would be silly, and yet the CAPM extended to its logical conclusion would have you do just that. I doubt very seriously that the creators of this theory would advocate holding tax-exempt bonds in a non-tax paying portfolio. I'm not trying to contradict a Nobel prize winning theory, I'm saying that the theory makes a series of assumptions that don't hold in real life. So treat CAPM as what it is: an idea about how to build a portfolio, not as an actual strategy.


Vivek said...

Good post.

I agree with 1.

Food and energy may account for most of the volatility in the short run (certainly seasonally) but if you look at say 2 year periods, is that still the case? (That's not rhetorical. I don't know the answer.)

With regards to 3, I think that every prediction comes with inherent probabilities. And whenever you are making those predictions, you should do so with a robust, preferably purely quantitative model and mention the 95% confidence interval as opposed to simply the expected value, since the former information is 1000 times more valuable than the latter. You can't do this with many if not most things, but it's vital to do it when there is enough data to do so.

4) Very true, but as you said in 1, you should keep checking implied inflation versus expected inflation and see if you have a trading or investment opportunity.

5) Agreed. The conclusions of CAPM do not explain the movement of asset prices at all. It's a pretty model, but even Markowitz has discussed the numerous shortcomings of the model. That being said, you wouldn't stick all your money in the investment that had the highest expected return (because of--of course, probabilities). But when you diversified, correlation would matter, so if TIPS (or any other assets) are uncorrelated with the rest of your portfolio, it can be very valuable depending on how much expected return you're willing to give to get less risk.

Tom G. said...

Looking at the longer-term contribution of food and energy to CPI variance is something I'm going to look into myself. That's a great idea, and for long-term investors, its the relavant question when thinking of TIPS. Great thought.

Another good thought is the possibility that TIPS might have a particularly low correlation with some other assets you are holding. Off the top of my head I don't know what they might be, but I'm not going to ignore the possibility.

Vivek said...

I went ahead and checked out the volatility of changes in CPI. I checked 3 Mo Trailing, 12 Mo Trailing, Annual, and Biannual. Over the ones shorter than biannual, CPI vol was higher. For biannual, CPI vol was slightly higher over the past 40 years but core CPI vol was slightly higher over the past 20 years, which would seem to imply that the difference between core CPI vol and CPI vol tends to play itself out over 2 years.

That being said, currently, I do think you're right that most of the volatility will come from energy for any number of reasons.

Tom G. said...

Well, if energy prices tend to be mean reverting, then core and total CPI should converge. Which seems to be what your data is showing.

I think that comes down to your horizon. If you are trading over a 1 year horizon, then energy will be a key call. If you are a strict buy and hold investor, then maybe my point about energy isn't so relevant.