6.89% is a very good risk-free rate, so I've been thinking about buying I-bonds in January. I've even pondered sending a tax overpayment to the IRS so that I could buy an additional $5k with my refund (for a total of $15k per person x 2).

The Dilemma

Markets expect a recession to start in 2023, and usually these bring rapidly falling inflation. Thus I expect the interest rate on I-bonds to fall throughout 2023 and 2024 until we reach a point where there are more attractive opportunities elsewhere. I must pay my last 3 months of interest to exit my I-bonds, which is maybe not as bad as it sounds if the bonds are not earning much interest at the time I cash out.*

On the other hand, one could argue that if I'm expecting falling interest rates, then I'd be better off locking in a rate with a nominal bond. If I went longer-duration, a nominal bond could appreciate as interest rates fell, or provide years of above-market yield. Moreover, if I spent a couple years in an I-bond, I might end up with a low-yielding I-bond and a market full of low-yielding alternative investments.

What's needed is a financial model where I can input my forecasts and receive an equivalent nominal yield.

**The Question:**

Given the forecast below, what is the equivalent yield I could have earned in a nominal bond?

**Variables:**

Jan'23 - April'23 interest rate: 6.89%

May'23 - Oct'23 interest rate (est): 5%

Nov'23 - April'24 interest rate (est): 3%

Rate at which I'd cash out rather than hold: 3.75%

So in this specific forecast, I would read in Nov'23 that my new rate was 3%, then would wait 3 months and exit the position in Jan'24. I'd therefore give up 3 months of interest at 3%**. The duration of the investment would be 13 months. Different forecasts or minimum rates would output a different duration.

Method

There are 3 distinct periods when interest is paid at different rates: 6.89% for 4 months, 5% for 6 months, and 0% for the last 3 months because interest is forfeited. Interest from the previous period is added to the principal at the beginning of each new period. Rates are quoted on an annualized basis, so the interest earned is pro-rated for the number of months that rate is earned. The time@rate column represents the percentage of a year that the annualized rate is applicable for. The amount of interest earned is balance*rate*time@rate.

**Results**

Time @ Rate Balance Rate Int. Pmt. Reinvested

Period 1 0.33~ $10,000 6.89% $229.67

Period 2 0.5 $10,230 5% $255.74

Period 3 0.25 $10,485 0% $-

In the end, after 13 months I'd end up with 4.85% more than I started with (4.48% annualized). Based on this math, I'd be slightly better off in one-year treasuries, which yielded 4.61% on Friday.

Check my math and assumptions. Do you get the same result that I-bonds are nothing to get excited about any more?

**Discussion:**

For those who bought I-bonds when the yield was >9%, this same method could be used to decide when it's time to exit.

Footnotes:

*In a strategic sense, a seller of I-bonds also gives up the opportunity to build a large I-bond hedge against inflation, accumulated over multiple years. I will ignore this argument for the sake of simplicity, and assume I do nothing but chase the best deals on stocks or bonds at any given time. But there is a case to be made that people should build up an inflation hedge as part of their AA instead of hopping in and out of small denominations of I-bonds during periods of inflation. A seller of I-bonds would also lose the option value of the I-bonds becoming a lotto ticket in the event of a prolonged and out-of-control period of inflation.

**You may be wondering what if the behavior was to exit the I-bonds on Oct 31, 2023 after hearing the new rates would be 3% instead of waiting until Jan 31, 2024? Then the annualized return for my 10 month holding period would be 4.29% because I'd give up 3 months of yield at the 5% rate and shorten the holding period to make those 3 zeroes a bigger proportion of the annualized average.