Effectiveness of swapping between muni & treasury money markets

[hkcj]

I noticed in another thread (that I now cannot find) that some people swap back and forth between the muni (e.g. VFYXX) & treasury funds (VUSXX) depending on which has a higher 7-day yield. Given that there's a lag due to the 7-day average and that by the time you switch there's some lost income (i.e. by the time you know the tax-equivalent yields have swapped places, you were actually under-earning for the past 7 days), has anyone actually computed whether this strategy earns you any extra money? How much, really?

[Steadfast]

I have not computed it, but as a muni investor I do watch these things go back and forth a bit. For me, what few dollars I might be able to squeeze out by constantly calculating tax-equivalent yields and performing the necessary transactions, is definitely not worth my time.

[erp]

I back-tested this using FFTXX and GOTXX which are MMFs available at Merrill, and which publish both 1-day and 7-day yields to get an idea of the gain from switching and also how much error everyone else was getting by relying on Vanguard funds which only publish 7-day yields.

I thought it was significant in the relative sense - at my tax bracket you lose 30% of the gain by swapping 2-4 days late everytime. But in the absolute sense it was small - $72 vs $50 gain on $100k being swapped over a period of 7/1/22 to 2/16/23.

Another person did the same analysis and decided it wasn't even a factor (ie the late swapping). I think he either used a higher tax rate or made a mistake by using 7-day yields in the accrual calulation though.

[retiringwhen]

I back-tested this using FFTXX and GOTXX which are MMFs available at Merrill, and which publish both 1-day and 7-day yields to get an idea of the gain from switching and also how much error everyone else was getting by relying on Vanguard funds which only publish 7-day yields.

I thought it was significant in the relative sense - at my tax bracket you lose 30% of the gain by swapping 2-4 days late everytime. But in the absolute sense it was small - $72 vs $50 gain on $100k being swapped over a period of 7/1/22 to 2/16/23.

Another person did the same analysis and decided it wasn't even a factor (ie the late swapping). I think he either used a higher tax rate or made a mistake by using 7-day yields in the accrual calulation though.

I used 1-day yields comparing FDLXX, and FTEXX at 40.8% rates and 10.75% state. There was a difference but it was inconsequential. $34/year extra return on a $100,000 investment with a total after tax return of approximately $1,200. BTW, swapping itself really isn't worth in the large. It may earn at best $100/year on the total return at the cost of approximately 10 swaps a year. Lots of labor for very small return.

[Walkure]

I back-tested this using FFTXX and GOTXX which are MMFs available at Merrill, and which publish both 1-day and 7-day yields to get an idea of the gain from switching and also how much error everyone else was getting by relying on Vanguard funds which only publish 7-day yields.

I thought it was significant in the relative sense - at my tax bracket you lose 30% of the gain by swapping 2-4 days late everytime. But in the absolute sense it was small - $72 vs $50 gain on $100k being swapped over a period of 7/1/22 to 2/16/23.

Another person did the same analysis and decided it wasn't even a factor (ie the late swapping). I think he either used a higher tax rate or made a mistake by using 7-day yields in the accrual calulation though.

I used 1-day yields comparing FDLXX, and FTEXX at 40.8% rates and 10.75% state. There was a difference but it was inconsequential. $34/year extra return on a $100,000 investment with a total after tax return of approximately $1,200. BTW, swapping itself really isn't worth in the large. It may earn at best $100/year on the total return at the cost of approximately 10 swaps a year. Lots of labor for very small return.

Two things that jump out at me. 1) recent backtests cover a period of very low interest rates in general. It may make a bigger difference going forward if short term rates are higher for an extended period of time. 2) At 40%+ marginal tax, I imagine it is very rare for treasuries to overcome the TEY of munis, and the short windows where they do will create whiplash as the advantage is quickly lost, hence needing 10 swaps a year to obtain minimal gain. Someone who has a total tax rate nearer a more typical 30% might see a more persistent trend before the yields cross, with more to gain and fewer transactions.

[retiringwhen]

I back-tested this using FFTXX and GOTXX which are MMFs available at Merrill, and which publish both 1-day and 7-day yields to get an idea of the gain from switching and also how much error everyone else was getting by relying on Vanguard funds which only publish 7-day yields.

I thought it was significant in the relative sense - at my tax bracket you lose 30% of the gain by swapping 2-4 days late everytime. But in the absolute sense it was small - $72 vs $50 gain on $100k being swapped over a period of 7/1/22 to 2/16/23.

Another person did the same analysis and decided it wasn't even a factor (ie the late swapping). I think he either used a higher tax rate or made a mistake by using 7-day yields in the accrual calulation though.

I used 1-day yields comparing FDLXX, and FTEXX at 40.8% rates and 10.75% state. There was a difference but it was inconsequential. $34/year extra return on a $100,000 investment with a total after tax return of approximately $1,200. BTW, swapping itself really isn't worth in the large. It may earn at best $100/year on the total return at the cost of approximately 10 swaps a year. Lots of labor for very small return.

Two things that jump out at me. 1) recent backtests cover a period of very low interest rates in general. It may make a bigger difference going forward if short term rates are higher for an extended period of time. 2) At 40%+ marginal tax, I imagine it is very rare for treasuries to overcome the TEY of munis, and the short windows where they do will create whiplash as the advantage is quickly lost, hence needing 10 swaps a year to obtain minimal gain. Someone who has a total tax rate nearer a more typical 30% might see a more persistent trend before the yields cross, with more to gain and fewer transactions.

I would strongly suspect the effect is in reverse or immaterial. The lower the tax rate, the less value for 1-day swapping. With 7 day yields, the value of swapping goes down to about half of the value for people at the highest tax rate. Really, all this swapping is not really worth. I suggest you play with my tool and find a place/time where it might be meaningfully valuable. I am using 7 yields in the tool, but I have sheet of data from earlier this month to test Fidelity funds.

[anon_investor]

The swapping isn't really going to yield a big amount. I have done it a few times at Fidelity since rates started going up but for me it doesn't take much time to check 1 day yields/swap.

[retiringwhen]

Two things that jump out at me. 1) recent backtests cover a period of very low interest rates in general. It may make a bigger difference going forward if short term rates are higher for an extended period of time. 2) At 40%+ marginal tax, I imagine it is very rare for treasuries to overcome the TEY of munis, and the short windows where they do will create whiplash as the advantage is quickly lost, hence needing 10 swaps a year to obtain minimal gain. Someone who has a total tax rate nearer a more typical 30% might see a more persistent trend before the yields cross, with more to gain and fewer transactions.

I would strongly suspect the effect is in reverse or immaterial. The lower the tax rate, the less value for 1-day swapping. With 7 day yields, the value of swapping goes down to about half of the value for people at the highest tax rate. Really, all this swapping is not really worth. I suggest you play with my tool and find a place/time where it might be meaningfully valuable. I am using 7 yields in the tool, but I have sheet of data from earlier this month to test Fidelity funds.

As an aside, I am running the history over the dates 2/18/22 to 2/16/23 There were significant movement and differences in the rates for 11 of the 12 months. Regardless, as with any sine wave vs. straight line, the returns will be very dependent upon start/end dates. I am using 12 months to reduce the likelihood of such a time dependent issue. I am working on the assumption that the seasonal swings in Muni's is pretty regular on a year to year basis and that seems to be a pretty reasonable assumption.

I just re-ran my Fidelity study with 24% Fed and 6% State. The improvement for 1-day yields was $35 on $1,400 total after-tax return on a $100K investment. On a percentage basis the improvement was less than at the higher rates 2.8% vs. 2.5% improvement. On a total return basis it only improved returns by 3 basis points.

At 40%/10.75 rates, VUSXX beat VMSXX about 33% of the time in the past year. Swapping is by far the most valuable at the highest rates as the sine wave on Muni's gives much bigger dips to buy Treasuries.

[hkcj]

This is an amazing set of information and resources - thanks everyone. The MM Optimizer especially seems like a great tool, and I had no idea other funds published a 1-day yield I could use.

I now have a whole different set of questions, specifically:

• Which was the higher easy (no-switching) return at 40.8%? (Treasury or MM?)
• Do the swapover points seem to align across Fidelity and Vanguard? E.g. can I rely on Fidelity 1-day yields for Vanguard switches?

[retiringwhen]

• Which was the higher easy (no-switching) return at 40.8%? (Treasury or MM?)
• Do the swapover points seem to align across Fidelity and Vanguard? E.g. can I rely on Fidelity 1-day yields for Vanguard switches?

- Muni's are almost always best at the 40.8% rate
- Most general purpose Muni funds all track pretty closely to the SFIMA index, so the difference are pretty minimal.

[hkcj]

Muni's are almost always best at the 40.8% rate

Thanks - I was curious because the rate seemed to dip pretty low on VNYXX... but I suppose I could have figured that out .

[retiringwhen]

Muni's are almost always best at the 40.8% rate

Thanks - I was curious because the rate seemed to dip pretty low on VNYXX... but I suppose I could have figured that out .

Try this tool, put in your rates and look at the history. you'll see that the rates vary a lot, but if you aren't into swapping, then the Muni's almost always win at the rate and for New Yorker's at least last year the in-state Muni was a winner (that is not always true in other states like CA).

MM Optimizer v4

[scoothome]

retiringwhen,

I just started looking at MM Optimizer v4. Nice tool! I think you have a couple typos on the My Best Now and Optimize Study sheets where you have "VUSXX - VSMXX". The last ticker should be VMSXX.

[retiringwhen]

retiringwhen,

I just started looking at MM Optimizer v4. Nice tool! I think you have a couple typos on the My Best Now and Optimize Study sheets where you have "VUSXX - VSMXX". The last ticker should be VMSXX.

Ahh , yes I added that column yesterday. I will fix

[sugobw]

Hi all, this MM Optimizer is confusing for me, and that's probably just because I am a bit dense...

If it's 40% (37% but add the 3% on top for NIIT, not sure if that is correct either?) + 10.75% state, it seems like VUSXX is "~60$" worse than VMSXX? Am I reading this correctly? And that's a marginal difference and that's with maximizing swaps?

So is it right to say that VMSXX is slightly better but barely. If I'm understanding this correctly, for me, personally, I don't want to swap and do all that extra work for very marginal gains, so sticking with VUSXX still ok?

Thanks all

[retiringwhen]

Hi all, this MM Optimizer is confusing for me, and that's probably just because I am a bit dense...

If it's 40% (37% but add the 3% on top for NIIT, not sure if that is correct either?) + 10.75% state, it seems like VUSXX is "~60$" worse than VMSXX? Am I reading this correctly? And that's a marginal difference and that's with maximizing swaps?

So is it right to say that VMSXX is slightly better but barely. If I'm understanding this correctly, for me, personally, I don't want to swap and do all that extra work for very marginal gains, so sticking with VUSXX still ok?

Thanks all

Not sure exactly what you are reading, but for NJ tax rate of 10.75 and 40.8% tax bracket (Tranch A on the Optimizer Tab), if you just held VUSXX vs. just held VMSXX, you would be $81 ahead after taxes over the past year with VMSXX (Cell AA10).

At that tax rate, if you swapped approximately 11 times over the last year between VUSXX and VMSXX, you with end up another another $115 (cell T10) over just holding VMSXX.

Choosing the Muni at that rate is a no brainer. Spending the time swapping is dependent upon the value of your time, each swap would have netted you $10 per $100K invested over the past year (cell V9).

https://docs.google.com/spreadsheets/d/ ... sp=sharing

[sugobw]

When you say choosing the Muni at this rate, what exactly does that mean? Just keep VMSXX and be done with it?

I'm not ever going to do Swaps. So I get your point, if I just keep VMSXX I could've been ~81 ahead (for last year). I understand this year could be slightly different, of course. It's a marginal difference, but still a difference by doing nothing (no swaps), which is what I would plan to do.

Thanks for your help

[retiringwhen]

When you say choosing the Muni at this rate, what exactly does that mean? Just keep VMSXX and be done with it?

yes.

I'm not ever going to do Swaps.

Good time management skills there!

[sugobw]

Haha, yes, I guess my philosphy is at this tax bracket, doing anything extra for 100$, or even upto a few hundred, it's really just not worth the time. Wonder if others in this bracket do feel the same.

Also, last question, I understand that VUSXX is "safe" as they come given majority is in Treasury, but the Muni (VMSXX) is just 100% tax exempt securities, so that is just as "safe"? The obvious answer here is yes of course, but essentially securities are just as good as treasury basically?

It's very interesting stuff though, because we're seeing VUSXX at ~4.5% versus VMSXX at ~2.2% (with a higher expense ratio!), which is a pretty big diff in yield, but still interesting how VMSXX comes out on top... I'm going to read more about this to further educate just out of the interest, but thanks for giving me the straight answer. Makes sense though why the higher brackets should use this because of the federal exempt. This community is quite helpful and nice.

[retiringwhen]

Haha, yes, I guess my philosphy is at this tax bracket, doing anything extra for 100$, or even upto a few hundred, it's really just not worth the time. Wonder if others in this bracket do feel the same.

Also, last question, I understand that VUSXX is "safe" as they come given majority is in Treasury, but the Muni (VMSXX) is just 100% tax exempt securities, so that is just as "safe"?

It's very interesting stuff though, because we're seeing VUSXX at ~4.5% versus VMSXX at ~2.2%, which is a pretty big diff in yield, but still interesting how VMSXX comes out on top... I'm going to read more about this to further educate just out of the interest, but thanks for giving me the straight answer. This community is quite helpful and nice.

It is a matter of debate just how risky Muni's are, but considering their very small premium vs. Treasuries and the fact that most people consider Treasuries the "risk free"* alternative, it could argued that Muni's are not worth the potential risk. I am in a much lower bracket, but even back when Muni's paid better for my rates, I kept them to a very small part of my portfolio (less than 20% of fixed income.)

The cyclical nature of the Muni market is actually kind of fascinating, but also surprisingly hard to exploit for any material gain. That was the whole point of my creating the Optimize sheet. I was testing potential situations where swapping may actually be valuable. Turns out, the value is pretty darn small in the absolute sense. If I had $1M in money market funds, maybe swapping would be worth it. People have suggested several optimizations, but each one adds effort and small marginal gain at best.

*These days "risk" is breaking out in hives everywhere. I wouldn't consider Treasuries to be in-fact risk-free.

[sugobw]

Yes, this was why I was asking, because if Munis are arguably 'more' risky than Treasuries, for such a nominal difference of <100$, I rather just keep it in what's "Safer," hence VUSXX.

[8000m]

The amplitude of the previous yield swing cycle & the current one actively hunting a March low have been fascinating to observe. The late-Jan/early-Feb muni yield lows dropping below their early-Dec yield lows was probably the most interesting bit to my mind's eye.
Looking forward to seeing how this March roundtrip plays out!

KevinM's google sheet & Retiringwhen's optimizer sheets (even the early versions) are amazing resources for those who find this topic interesting. Thank you both for putting that time in & letting others benefit, much appreciated. Each doc has been bookmarked for me since inception.
@Retiringwhen, your continued iteration on optimizer & the resulting workbook is particularly impressive.

[Hogan773]

Where do you think we are in the VMSXX swings right now? Should I bother switching to VUSXX for a bit? I think even at top rates we are well below VUSXX tax adjusted yield

[sugobw]

I just kept it in VUSXX and forgetting about it. Even though according to the MM Optimizer sheet, VMSXX is the best (for my bracket/state anyways), but it's really a marginal difference by the end of it all, and munis are arguably "less safe" than treasuries. Not thinking about it anymore - pay a few extra bucks on tax, but it is what it is. Good enough for me.

[Hacksawdave]

I own all three of the Vanguard CA tax-exempt offerings. I have never swapped between a taxable and tax-exempt MM fund, and never will. I have owned the CA MM and intermediates since 1997 and added the CA long-term in 2018 before I was retired.

Why does one use them? The most talked about benefit is taxable versus tax-exempt yields for high income. I use them for other downstream benefits besides reducing ordinary income taxes:

• Ability to use tax-free dividend money to exchange into a different fund through auto-exchange.
• Better exploit the zero capital gains rate with lower AGI.
• Larger CA itemized deduction allowance through lower AGI exclusion amount.
• Better cashflow in retirement with less to no estimated tax payments.
• Ability to take more tax deferred 401k distribution before RMDs at a lower tax rate.

I use taxable MM funds in tax-deferred and Roth accounts for staging backdoors, bucket transfers, and automatic exchanges.

[Hogan773]

@RETIRINGWHEN

Wow what a cool spreadsheet! Thanks

So I took a look and in Illinois and assuming your "Tranche 2" scenario of just flipping between VMSXX and VUSXX it takes 11 swaps (presumably well-timed) and on $100K that yields extra $132. Or if we are talking about $500K then $660, or a million dollars then $1320, and so forth.

If we had to pick one or the other though is it the case that VMSXX itself will do better or will VUSXX. I need to go open up your sheet again to see

[retiringwhen]

@RETIRINGWHEN

Wow what a cool spreadsheet! Thanks

So I took a look and in Illinois and assuming your "Tranche 2" scenario of just flipping between VMSXX and VUSXX it takes 11 swaps (presumably well-timed) and on $100K that yields extra $132. Or if we are talking about $500K then $660, or a million dollars then $1320, and so forth.

If we had to pick one or the other though is it the case that VMSXX itself will do better or will VUSXX. I need to go open up your sheet again to see

Yes. you get the idea. Make sure you have both state and Federal rates correct. But if you are a high federal rate taxpayer in IL, then yes, the VMSXX will probably be the best deal. IL has relatively low flat rate, so the muni funds will work best for most high income investors. If you really have $1M of cash to put in a money market, then timing swaps may actually be worth your while.

Generally, anyone with a MFJ taxable income over $400K in IL will be better off with the muni fund, if you don't want to swap. But if you swap, you would add some value.

[muffins14]

Surprisingly for me, FDLXX at 4.27% is 2.528% after all taxes, but the NY muni FSNXX is only 1.98% right now (7 day SEC yields)

I’m at 37% fed plus 3.8%, and 10.7% state plus local

[Hogan773]

@RETIRINGWHEN

Wow what a cool spreadsheet! Thanks

So I took a look and in Illinois and assuming your "Tranche 2" scenario of just flipping between VMSXX and VUSXX it takes 11 swaps (presumably well-timed) and on $100K that yields extra $132. Or if we are talking about $500K then $660, or a million dollars then $1320, and so forth.

If we had to pick one or the other though is it the case that VMSXX itself will do better or will VUSXX. I need to go open up your sheet again to see

Yes. you get the idea. Make sure you have both state and Federal rates correct. But if you are a high federal rate taxpayer in IL, then yes, the VMSXX will probably be the best deal. IL has relatively low flat rate, so the muni funds will work best for most high income investors. If you really have $1M of cash to put in a money market, then timing swaps may actually be worth your while.

Generally, anyone with a MFJ taxable income over $400K in IL will be better off with the muni fund, if you don't want to swap. But if you swap, you would add some value.

So if I wanted to try to swap, what is the most effective way to watch and anticipate the swap points? Should I follow the Sifma index and use that as a predictor because VMSXX SEC yield as reported on Vanguard ultimately follows it with a lag? So when the SIFMA index is heading upward and equals the tax adjusted yield on VUSXX for example then swap out of VUSXX and into VMSXX and vice versa?

[retiringwhen]

So if I wanted to try to swap, what is the most effective way to watch and anticipate the swap points? Should I follow the Sifma index and use that as a predictor because VMSXX SEC yield as reported on Vanguard ultimately follows it with a lag? So when the SIFMA index is heading upward and equals the tax adjusted yield on VUSXX for example then swap out of VUSXX and into VMSXX and vice versa?

Other folks have been trying to figure that out, but I think watching the weekly on Wednesday update of the SIFMA index is probably the best approach IF you are willing to do the homework to setup the necessary calculations.

The problem is how do you calculate after tax rates for the SIFMA rate mapped to your actual fund (remember expense ratios will lower your fund yield vs. the index).

You will also need to do the same with the swapped fund (e.g., VUSXX). But since it is a 7-day average, it is "behind". so you need to decide how to handle that.

All this is complicated and for very little difference. I did a back-test using one day returns for Fidelity funds and the SIFMA rates and it made almost no difference.

I personally would just watch the daily reported after-tax rates on my sheet and swap when it passes one direction or the other. Really, I wouldn't even both unless you have a couple hundred thousand $ sitting around in cash.

[Hogan773]

So if I wanted to try to swap, what is the most effective way to watch and anticipate the swap points? Should I follow the Sifma index and use that as a predictor because VMSXX SEC yield as reported on Vanguard ultimately follows it with a lag? So when the SIFMA index is heading upward and equals the tax adjusted yield on VUSXX for example then swap out of VUSXX and into VMSXX and vice versa?

Other folks have been trying to figure that out, but I think watching the weekly on Wednesday update of the SIFMA index is probably the best approach IF you are willing to do the homework to setup the necessary calculations.

The problem is how do you calculate after tax rates for the SIFMA rate mapped to your actual fund (remember expense ratios will lower your fund yield vs. the index).

You will also need to do the same with the swapped fund (e.g., VUSXX). But since it is a 7-day average, it is "behind". so you need to decide how to handle that.

All this is complicated and for very little difference. I did a back-test using one day returns for Fidelity funds and the SIFMA rates and it made almost no difference.

I personally would just watch the daily reported after-tax rates on my sheet and swap when it passes one direction or the other. Really, I wouldn't even both unless you have a couple hundred thousand $ sitting around in cash.

Okay thanks makes sense

Still amazed by the magnitude and regularity of the swings in that yield. Seems odd

[scoothome]

@retiringwhen

You've created such a cool tool in MM Optimizer. I suggest you start a master thread for it, similar to what Ben Mathew did for TPAW viewtopic.php?t=331368, longinvest did for VPW viewtopic.php?t=120430, BigFoot48 did for RPM viewtopic.php?t=97352, etc.

That way you can announce version updates, users can report bugs and request features, etc. Or maybe you have already started such a thread and I missed it?

[retiringwhen]

@retiringwhen

You've created such a cool tool in MM Optimizer. I suggest you start a master thread for it, similar to what Ben Mathew did for TPAW viewtopic.php?t=331368, longinvest did for VPW viewtopic.php?t=120430, BigFoot48 did for RPM viewtopic.php?t=97352, etc.

That way you can announce version updates, users can report bugs and request features, etc. Or maybe you have already started such a thread and I missed it?

I have a version 6 brewing with some thoughts on T-Bill ladders and auto-rolls. Once that settles down, I will start a dedicated thread when I announce that version. Thanks!