Although I messed up my Treasury Bill Ladder by switching to 6-month T-Bills, they continue to be a pretty good cash investment alternative for those people in states with high income taxes, as they are exempt from such taxes. For example, the most recent 6-month T-Bill paid a rate of 4.754%. Using my Equivalent Rate Calculator, and a federal tax rate of 25% and state tax rate of 9%, that’s the equivalent of a regular 6-month bank CD paying **5.40%**. The 4-week T-bill equivalent rate is 5.07%.

For all my posts on Treasury Bills, please see here, reads bottom to top.

Interest rate is still rising, 28-day is more attractive than 182-day…

what do you think about the upcoming change to the I bond inflation index rate ? http://www.savings-bond-alert.com thinks it will drop to 0 !!

AA – Maybe, maybe not 😉 The question depends on your horizon and how much and how fast will it rise?

sid – All I know is that we’ll can figure it on around April 15th before it is actually announced, so one can better decide whether to buy in April or May, if at all.

Hey bud, I love your blog….saw it in BusinessWeek and checked every day since. One question for you: Isn’t interest on treasuries taxable for federal income tax purposes? Here is a link to a calculator I use:

http://money.cnn.com/pf/101/lessons/7/page4.html

Hey Daniel, thanks

Yes, you are only saving on state+local income tax with T-Bills, you must still pay federal taxes.

Yep, I’ve seen that CNN page, it was mentioned in the comments of my calculator post. I believe the calculator results differ from mine because it assumes that you can deduct your state taxes from your federal taxes. My previous reply:

“Like I’ve said before, I think doing that overestimates the advantage of being able to deduct your state taxes on your federal taxes.

1) To do that, you must itemize deductions. Many people do not itemize.

2) Even if you do itemize, the tax advantage of itemizing is only that which exceeds the standard deduction. This is also why some people overestimate the usefulness of the mortgage interest rate deduction.

The standard deduction for 2006 is $5,150. Say if you itemized and get $10,000 in deductions. Only the amount above the standard deduction, $4,850, actually saved you any money. So you could either look at the first half of your deductions as totally useless, or pro-rate the effectiveness of each individual deduction. I think the latter makes more sense.

But hey, it’s not that big of a difference overall. And I don’t itemize right now, so I won’t sweat it. =)”

Yes you are absolutely correct about the overestimation. It is a common mis-conception for many people.

Thanks for having this awesome page. My fiance and I are in a similar situation as you. I think she questions my fascination with tracking our net worth all the time, but as you know, it needs to be done.

You have a lot of great ideas that are right in line with a lot of ideas I have also. I’m trying to refrain from sharing until I get to read all of your past blog posts in case you have already mentioned them! Keep up the good work!

To itemizing, people really ought to consider “doubling-up” charitable deductions in a year, then forgoing the same deductions the next year if they are barely above the standard deduction (and thus, itemize).

My brother had “itemized” deductions of $4,900, so obviously he took the standard at $5,150. Those deductions were all tithing. If he had doubled-up and prepaid his tithing for 2006 in December of 2005, he would’ve had ~10k in itemized deductions. The next year he could’ve then taken the standard.

Original plan over two years: $10,300 of deductions.

Doubling up over two years: ~$15,150 of deductions.

It’s a no-brainer if you have the cash available, and you charitable organization is happy because you prepaid

Just to let everyone know, historically “laddering” short-term, 28-day, T-bills, does not materially benefit you. The historical yield is .02% higher if you do not ladder. I buy all my T-bills on the first week of every month and I have yielded a little bit more than if I laddered week to week. Save your time and just buy at the beginning of each month would be my advice. If you would like the historical quotes and my calculation please just send me an email. Thank you.

Same goes for the 182-day, 6-month, T-bills. Historically you will yield .07% more if you do not use the ladder system. Save your time and just buy at the beginning of each month would be my advice. Again if you would like the historical quotes and my calculations please just send me an email. Thank you. Now a question for Jon, did you do your research before you started buying through Treasury Direct and using the laddering system? Seems to me you could have had higher yields if you did not use the ladder system. Just curious.

I’d be interested in that info, is that based on your own research or is there a link? Isn’t buying every month still laddering, just at a wider interval (monthly instead of weekly?) Right now I’m just buying 6-month T-Bills once a month.

Also, another major benefit of laddering is that every week you will have part of your money ready for another investment if desired. For example, if you laddered four 4-week T-Bills at $1,000 each week, every week you’d have $1,000 ready in case you need it.

If you just did $4,000 once a month, you would have to wait up to a month for that $4,000 to come back to you.

I’m using this money as my emergency fund, so easy access to at least some of the money is important.

To answer your questions, its both my research in doing the actual calculations of yield, the data comes from Treasury Direct. http://www.publicdebt.treas.gov/of/ofaicqry.htm

And no buying monthly (in the 28-day T-Bill example) is not considered laddering, you buy once, one lump sum, wait until maturity then buy again, opposed to buying each (smaller amount) and every week. Way more maturities in laddering. If you placed all your money in one 6-month T-Bill you would yield .07% higher than laddering. I realize liquidity is a concern to some people, like you, with the 6-month T-Bill option. Also, in your second response, liquidity would be a lesser issue with 28-day T-Bills (only yielding .02% higher than laddering), if you had an emergency, just charge it on your credit card, when the 28-day T-Bill matures, pay with the proceeds. For the mass audience and you, historically if you lump your money together and purchase a single 28-day or 6-month T-Bill your yield will be higher than if you laddered your money. I have not yet done the research on other T-Bill option yet. Thank you for your time.

CC,

Interesting, but I disagree with your findings.

I can believe that, maybe for a specific study period, 28-day T-bills reinvested at the first of each month yielded .02% more than the same principal amount split up and invested weekly over the same period. But unless I’m missing something about your research, I’m missing how this can be extrapolated to say laddering T-bills always yields less than reinvesting a lump sum of the same total amount.

Consider this: The ladder by definition provides the AVERAGE T-bill rate over the 4-week period, while the first week of the month in your study period apparently had better yield than other weeks over that period.

So, if instead of the first week each month you chose the 2nd or 3rd or 4th for your lump sum, your results would have been different, with one or two of the weeks favoring laddering over lump sum.

But again, this is just academic, we’re talking about 0.02% anyway…

BTW, you make a good point on being able to use the CC for liquidity when the maturity’s only a month anyway.

To respond to Dan’s recent comments, I did the research (which I belive you failed to do) over a 5-year period and I realized that purchasing the first week of the month historically provides for the best results. You seem to think that .02% is not a huge difference. I disagree in the highest degree. You compound the interest and it is a big deal over the long-term. Also, with the 182-day T-Bills the difference is .07%. I propose that you run a calculation of when you purchased your T-Bills using the ladder system and a calculation of if you purchased them at the beginning of the month. Then post your results within this blog. Thanks.

CC – You’re right, I didn’t run through such numbers. But, you didn’t address the good question Dan asked.

What if you bought T-Bills every 2nd week of each month? Or 3rd week? Or 4 week? If you did that comparison as well, I’d be interested to see the results. If you didn’t, then all you have done is some restricted back-testing.

For example, with back testing I’m sure one day of the week, say Tuesday, stocks end up going up a bit more than any other day of the week. Does that mean we should only buy stocks on Tuesday?

Jon – Thank you for your agreement with my reasoning. Also my apologies on not addressing Dan’s question, I did not see any question in his response. I have done the comparisons and here are the results:

28-day (1 month) T-Bill:

Buying in the 1st week (like I recommend)- 1.68%

Buying in the 2nd or 3rd week – 1.64%

Buying in the 4th week – 1.65%

Using the ladder system – 1.66%

182-day (6 month) T-Bill:

Buying in the 1st week (like I recommend) – 1.48%

Buying in the 2nd week – 1.45%

Buying in the 3rd or 4th week – 1.46%

Using the ladder system – 1.41%

In conclusion, my initial postings are correct. Investing lump sum in the 1st week of the month yields better results than using the ladder system or any other week. The ladder system does not yield better results. However, the ladder system does add a factor of liquidity (meaningless to the 28-day T-Bills).

My recommendation remains the same, DO NOT use the ladder system when purchasing 28-day T-Bills or 182-day T-Bills. The only reason why I would recommend the ladder system for the 182-day T-Bills is for liquidity reasons only. If you do not need the liquidity factor AVOID the ladder system. Any questions/comments are always welcome. Thank you for your time.

PS To answer your question. No do not only buy stocks on Tuesdays. You cannot time the market. Long-term investing is the best recommendation, unless you have insider information and want to serve time like Martha.

CC – I thank you for doing the math, and I respect your thoroughness.

However, given that buying in a week other than the 1st week actually give worse results than laddering, I do not know if your results are truly statistically significant. How do you know it is not simply due to expected deviations based on your sample size? It may be that due to a pyschological or other reason, the rates during the first week are higher than average, or it may not.

I am not a statistician, but coming from a research background that would have to be addressed for me to really agree with not laddering.

Another example. If I flip a coin 100 times, I may get 48 heads and 52 tails. Does that prove that on the next flip it is more likely to be tails than heads? No, because of the sample size, there is an expected possible deviation.

Now, if I flip a coin 1,000,000,000 times and I get 480,000,000 heads and 520,000,000 tails, that would be much more significant.

See, this is what I had to see for myself.

0.02% / 1.68% is a difference of about 1.2%.

I did a download of your link and got about 250 4-week T-Bill auction results. The question is, is a difference of 1.2% significant across 250 samples?

I just wrote a script that made 250 random numbers from 0-100. The average was 49.18, off the expected average of 50 by 1.6%. Again, this is not the proper way to do this, but my guess would be that 0.02% is not statistically significant.

That is, there is no reason to expect that buying at the 1st week of the month will continue to yield better results than laddering, any more than flipping the next coin will be more heads than tails. In fact, due to reversion to the mean, it may be even WORSE over a long-term view.

Jon:

I felt my original sample size of 5 years, week by week, was a sufficient size. But you tended to disagree. Therefore, I increased my samle size to 10 years, week by week. The results were the same. Buying in the 1st week yields higher than laddering. I did this research before I started buying into Treasury Bills and my research has paid off. I am earning more than if I would have laddered my purchases. Therefore, I believe that my research and actual results speak for themselves.

Hey, I feel that this has been a good discussion. Of course, we should all do what we want, but having a good debate leaves us all better off. I’ll stick to laddering, myself.

Jon:

Sure has been a good discussion. Just wanted to bring new light that laddering does not always produce the best results. You stick with laddering, below par results, and I will stick with purchasing on the 1st week of the month, better results. I feel great knowing that I am outperforming. Thanks again.

LOL. You sure seem confident. Banks and financial institutions, many foreign, are the largest purchasers of Treasuries, much more than individual buyers like us. Do you really think they would let such an easy-to-exploit market inefficiency exist?

Let’s agree to compare again in 20 years. That will be the real test, no?

Sure, I am confident, numbers do not lie. True banks and institutions are the largest purchasers; however liquidity is a HUGE factor in their case. Statistically there is more money “thrown” into the Treasury at the beginning of each month than any other week. Do you think maybe that is why the 1st week we can have a better rate than any other week? Sure we can compare in 20 years. Let’s see who can reach the goal of 1M first. That will be the true test.

Sorry, I’m not interested in a pissing contest. There is no need to make this personal, we’re talking about 0.02% here. 0.02% of $100,000 is $20. 0.02% of $10,000 is 2 bucks. Everyone has different earning potentials, and I do not compare my net worth with anyone else. Tons of people outearn me, and have net worths higher than me. My goals are personal ones.

Just to bring you back to reality, we are not just talking about .02% (in the case of 28-day T-Bills) or .07% for 182-day T-Bills. It may be only .02/.07 today but compounding that over the long term and it is much more than the way you are seeing it. True everyone has different earning potentials and its good not to compare net worths. My goals are also personal. Let’s strive to meet our goals and do the best we can!!!

Again, interesting stuff!

But CC, my point still stands on the 28-day T-bills. Over any study period the weekly laddering method will yield the average (or technically, a very small amount below average because T-bills are discounted) of all T-bills over the study period. I assume from your results above, that you don’t disagree with this.

Picking one week out of the four for a continually reinvesting lump sum is no guarantee that it will outperform weekly laddering.

If I understand correctly, you are suggesting to buy 28-day bills in the “first week of the month.”

But while there are 12 calendar months in the year, there are acutally 13 28-day periods. So say you started reinvesting a lump sum of T-bills on the first calendar week of a particular month, after a few months you’d find yourself in last calendar week of the month. So really, you are aribitrilly picking one week out of the four-week cycle as your Week #1.

Your results show that over your study period, the week you happened to call Week 1 was the top week. But someone else could start reinvesting on the first week of a different calendar month and find themselves calling a different week in your cycle Week#1.

Everything I said above applies only to the 28-day T-bills of course. The only way to assuredly average out the 6-month T-bill rate would be to buy a T-bill every week over the entire period, and my allocation to T-bills simply isn’t that big! I ladder them monthly, which is what I believe you said you do. If in your research you’ve observed some advantage on the first week of the month with six-month T-bills, well, that happens to be exactly how I scheduled my ladder anyway, so COOL DEAL!

Dan:

I would still have to disagree with you. I did not say that I use the ladder system. I strongly am against it. If you placed all your money (lump sum) in at the first week of the month you would outperform the ladder system. So you are not realizing the cool deal that I have. You have an extremely cold deal.

I’m trying to schedule a repeating transaction at TD. It seems you have to enter in the number of repeat transactions you want. Am I doing it wrong, I thought you could just select “monthly” and let it roll until you wanted to cancel it?

thanks

CC,

I’ve really enjoyed this thread, but it’s lasted quite a while. So I’ll post just one more time on this one. I believe to some extent where we don’t agree is purely semantics.

I’m purchasing six-month bills the first week of each month. Because I’m doing it each month, I am calling it a ladder. It sounds like you are doing the same thing, but not calling it a ladder. (Please correct me if I’m wrong.)

As far as the 28-day bills go, I’ll restate my argument a different way:

There is no such thing as “The first week of the month” for investing a lump sum into 28-day t-bills and reinvesting upon maturity.

Why on earth would I say that? Because, assuming you’re automatically reinvesting it each time it matures, three months later you’re purchase date will no longer be on the first calendar week of the month. (12 calendar months per year vs. 13 T-bill 28-day periods per year.)

This means that, over the study period you documented, what you called “Week 1″ was in reality an arbitrarily chosen week. Someone taking your advice to buy on the first calendar week three months later actually would be in “Week 4″ of your cycle, which you’ve stated above underperformed the ladder system over your study period.

LM,

Sorry, meant to respond to you in the same post above.

I came to the same conclusion you did: You must enter a number of repeating transactions.

However, this doesn’t really stop you from doing what you’re trying to do, because you can just pick a number that puts you way into the future anyway, and cancel it later if you want the reinvestments to stop.

Dan:

I am not doing the same thing you are doing. Again, let me say I DO NOT LADDER. I am putting all my money (lump sum, $6,000) into one 6-month T-Bill and just leaving it. You are laddering no matter how you look at it. In my case it would be buying (6) 6-month T-Bills spread out over a 6-month period. I only have (1) T-Bill. I AM NOT laddering. Like I have said before laddering involves way more maturities. I will just continue to gain the benefits you are ignoring. Best wishes!!!

CC,

Thanks for clarifying the 6-month bill confusion. In a post above, you advised people to buy 6-month bills “at the beginning of each month”, so I was just saying that happens to be exactly what I already do.

However, you still haven’t responded to my comment in the last post about the 28-day bills.

Dan:

Yes you are buying at the beginning of the month, but you are not maximizing your return. Though, I do understand that if liquidity is your preference than you are doing the right thing in my opinion, even though you may be giving up several percentage points.

This is what Dan was referring to when asking about 28-day T-Bills:

I would put it this way. Say you buy a 28-day T-Bill on the first week of Month 1. 28 days later, your bond matures. But now, as months vary in length, you are not necessarily at the first week of the next month yet! So your choices are

1) Go ahead and buy it anyways, but you would lose your supposed benefit of buying at the 1st week of the month.

2) Wait until the first week of Month 2. Presumably you’d be putting your money somewhere else during that time, a few days to a week, earning less interest. Again, you would probably lose any 0.02% edge doing so.

Again, I have only $6,000 in T-Bills right now so I really don’t care either way, this is just for my personal knowledge.

Jon,

I set up a 28 day once a week purchase ladder for similar reasons, liquidity, security and ease-of-use. For someone that doesn’t want to tie up emergency cash for too long I find the Treasury Direct website easy to use and the equivalent returns adequate for my purposes. Thanks for the information.

I can’t say I like his attitude or respect for others, but I do have to confirm he’s right. I ran through the last 5 years of 28-day T-bills (that’s as far back as I could get data for 28-day bills) and the first week of the month does outperform the average by 0.02% I than followed this up by determining the statistical significance for a population with an unknown mean and unknown variance. I get a t-value of 34.66. P(t>=34.66) is zero beyond Excel’s precision. Thus, the 0.02% difference is real in the world of statistics.

How does this play out in your lost profits? If you bought in the first of every month instead of every week for the last five years, that difference compounds to 1.207% Now, it’s up to you to decide if that 1.207% makes any difference to you. Please let me know if you find faults in my statistics, I’m an engineer, not a statistician. Finally, CC, don’t be an ass man, nobody has anything here to prove.

Damnit I hate excel, the actual probability that the 0.02% difference is a statistical anamoly is 42.55%, not zero. Invest whenever you want to.

Ryan, Excellent analysis!

But still, I have to be a broken record and remind about the calendar month versus 28-day cycle issue.

As Jonathan noted, you would presumably be holding matured T-bills in a bank account for a week every three months or so in order to make sure the following purchase will occur on the first calendar week of the next month.

In the current interest rate environment, even if you used a hi-yield savings account, that would total to a month out of each year earning (usually slighlty lower) interest that is also subject to state taxes, which presumably is what made T-bills a more attractive cash investment in the first place.

If there were a way to factor in how the occasional week of not automatically reinvesting the T-bill would affect returns, my best guess is that the 1.207% over five years would decrease considerably.

Grrrr… this is the last time, I swear. I knew that t-number looked incredible large compared to the standard deviation… the correct t number is 0.19 (I had my Excel cells all mucked up.) P(t>=0.19) = 42.55%, ie, there is a 42.55% chance this was just a statistical anamoly.

Dan, I think you’re absolutely correct about the time between first week investments. If you wait for the first week of the month, you get exactly twelve T-bills in a year. If you instead bought T-bills every 28 days (I’m not sure if you can do this, I’ve never bought them) you’d get an extra T-bill every year. Even at the 0.02% lower rate, over the course of the year, you’d gain 2.16% by that extra compounding.

Regardless of the extra compound, the 0.02% difference is completely (statistically) insignificant. The moral of the story is, buy them whenever you want. I’m sorry for changing my story so many times, but I promise you this is the final time.

Yep, that’s how they sell them. They’re “28-day bills” not “1-month bills”.

Continually reinvested, as you said, 13 T-bill cycles would fit in each year.

Don’t apologize – thanks for doing the research!