Does the rate of return for TIPS match that of inflation?

Tags: 

I would like to test an all TIPS portfolio. In ESPlanner "Inflation Indexed Government Bonds" a.k.a "TIPS" show a mean return of 2.99% in the Monte Carlo section. The benefit of TIPS is supposed to be they hedge against high inflation: the kind forecast in Kotlikoff & Burns book "The Coming Generational Storm". My question is if the return of TIPS in ESPlanner matches the inflation rate. If not, and they hold at 2.99% their benefit will not be shown in a high inflation scenario.

Comments

dan royer's picture

I'm using 2.30.3 version of the program and see a mean REAL return of 3.68% with a relative risk of .09. We will be updating this return data in the next three weeks or so for our next release.

But to your question . . . remember, you set the inflation rate in your Assumptions area. So if you leave it at 3% the real return you see on TIPS (whether it is 2.99% or 3.68%) will reflect a return that is "real" or relative to that inflation rate. If you set it at 10%, the returns will be relative to that inflation rate. Of course your living standard will go down with higher assumed inflation rates, but the distribution of that living standard as it departs, upside or downside, from the mean will be less because there's less risk in your asset allocation.

I don't know if that really answers your question. But when you look at your MC report for distribution of living standard, you'll see a benchmark of 0% real return. This would be to somehow invest in an asset class that merely keeps pace with inflation. I think the closest we might have to that is the short term bonds. TIPS have some risk, more variance even than some short term government bonds. That raises the question as to why TIPS might be thought of as a "safer" bet in a high inflation environment than short term bonds. Not sure I can answer that--and perhaps people don't think that. But short term bonds look like the least volatile (and as expected, lowest real return) asset class among those we list in the MC area of the program (aside from cash).

If you want to model a high inflation environment, change your assumption about inflation. But once you set that inflation assumption and then pick your asset class, that model is set at that inflation rate. It doesn't say what would happen at a different inflation rate. For that you'd need to change the inflation rate again.

Thanks, Dan. That basically answers my question. I think the advantage of TIPS is their "inflation factor" that accumulates as inflation rises. So if I bought a 10 year TIP for $1000 in 2015, if the inflation rate is 10% per year thereafter, when the TIP matures in 2025 I get back my $1000 in 2015 dollars, namely, $1000 X (1.1)^10. A comparable regular Treasury bond would only pay me back the $1000 in 2025 dollars. From what you describe the TIPS in ESPlanner would track inflation nicely with the additional return above inflation (2.99 or 3.68). However, it seems even the lower return short term treasury bonds would track inflation in the model per your explanation above. In reality, if I bought a 10 year treasury today and held it to maturity and the inflation rate sky rockets to 10% over that period, I would lose money in real terms at maturity. If that's the case it sounds like a shortcoming of the model in a high inflation scenario for non-TIPS bonds. Please correct me if I've misunderstood.

dan royer's picture

Well, I'm not sure that I communicated clearly, or perhaps I'm not thinking it through clearly. But what I meant to say is, in part, is that if you use Assumptions to set the inflation rate at 3% you get a model where average inflation over the next 40 years, say, (or however long you have yourself living) is 3%. That's not a model for inflation changing to 10% and so your model doesn't say anything about what happens if inflation goes to 10%. For that, you'd have to set inflation at 10% and rerun the model.

Note also that the rate of return for all the asset classes is a real rate of return. So the short term bonds is .6% real rate of return. I guess I'm not sure how that actually works in practice outside of a mutual fund where you have a duration on the fund that is cycling through that short term duration of say 3 years for example. If you bought a ten-year bond, that's different from buying a mutual fund tracking bonds with a aggregate ten-year duration because such a fund would be buying better yields on the way up and so forth. Perhaps I'm unnecessarily confusing things.

Since short term bonds mature in a short time you get to reinvest at the inflated coupon rate so you do have inflation protection in the long term, maybe not as good as TIPS but still protection.

The Monte Carlo is showing the upside and downside probabilities that are inherent in particular asset classes. It's not showing what happens in a inflation-dynamic environment. Again, in your model, inflation is fixed at whatever rate you set as is the return and variance inherent in the asset class.

So I think you might have misunderstood me. In your model, any asset class you choose is revealing a real rate of return--what changes is the variance along with that rate of return. With TIPS you see a lower rate of return as well as lower variance. But with short term bond asset class you see an even lower rate of return and I believe an even lower variance.