Why Investors Should Fear The Permanent Portfolio

[stone]

Clive, thanks for explaining the rate tart ladder so clearly. I'm still puzzled though about it not mattering loosing the capital gain. I perceived a key benefit of the LTT capital loss/gain being to "crowd out" potential misallocation to the other three assets when paying in money and to direct which asset to draw down when consuming the PP. Does it make any difference if  the PP is being paid into for a few decades then sold off for a few dacades rather than just being maintained with an infinate time horizon? If I had most of my savings in a PP, I would also find it attractive to be able to totally liquidate it at a moments notice if needs must. I thought that in practice people do actually end up needing to liquidate just when things are at their worst (in order to emigrate or whatever). In such a scenario wouldn't lack of overall portfolio volatility really be a godsend?

[doug6zj9]

Daily volatility is just a paper value fluctuation. What matters at the end of the day is to have some actual gains banked.

Agreed; gains banked or losses averted.  My point is I can't predetermine end of "which" day.  Thereby short and medium term stability (of the package as a whole) is of value.

Don't bank on LTT's bailing you out. Rising base rates/inflation for instance could see both stocks and LTT's down, perhaps substantially. The fact that their prices cycled around each other during that dive and smoothed the line wouldn't be of much comfort.

A. of course, and declining base rates/inflation could see them up...
B. that is why the other 50% is gold and cash
C. I suggest that for real people in real time, the parts cycling around each other and smoothing the line does provide much comfort.  That is the point of the PP and it's diversification.

[Storm]

The rate tart and 5 year approach however will tend to work both when interest rates are rising or in decline. The ST/LT barbell however more generally only works when interest rates are in decline. The ladder can also be fully rolled out of in 5 years or less (return of capital), whilst a ST/LT barbell could be locked into capital losses for an extended period of time.

Clive,

Please forgive me if I'm being dense, but I think I understand what you are saying that the rate tart 5 year ladder generally gives you a 1.2% advantage over standard yields, so if yields are averaging 2-3%, and core CPI is 4.6%, isn't there a real risk of having negative real returns with a 5 year rate tart ladder?  During periods of 7.5% interest and 3% inflation, you could really make a killing, but it seems like now we are in periods of 2-3% interest and 4-6% inflation.  Please feel free to correct me if I am wrong here.

Of course you don't have to stick to ladders (or barbells) of Treasury's alone. You might include some more riskier alternatives, maybe looking to boost yields by taking on an element of additional risk. Here in the UK for example some lenders are being paid 8% to 10% via such options as http://www.fundingcircle.com/ - where borrowers are matched with lenders, or a single lender can spread there lending across multiple borrowers to reduce the risk from defaults. If the defaults are few enough then you likely make better gains. If defaults are high then you might lose a substantial amount. Over time from what I've seen it all tends to balance out and lending more speculatively encounters failures (defaults) that wipe out the higher rewards, that overall you'd have done equally as well by investing in safer Treasury's.

I would strongly recommend against any of the peer-to-peer lending.  Prosper.com is an example of one company that offers unsecured loans, and claims that if you spread the risk among enough borrowers, you effectively reduce your overall risk of default.  The problem is that these are unsecured loans to borrowers that couldn't qualify for financing at a normal bank.  The default rate was as high as 20% per year.

I'm a little shocked, to be honest, that you would recommend unsecured peer-to-peer loans as an alternative to AAA government debt.

[Bongleur]

I suppose you can simulate a ladder with funds ?

... having say several months, 1 year, 3 year, 5 year maturities... but how would you roll out of them?  Other than by diverting dividends from reinvestment to another fund...

[stone]

Clive, as the long term bonds component of the PP is there to deal with a Japanese style deflation scenario, I guess Japan is the best example to check whether they do actually work as intended. From what I can see if someone bought 1000 Yen of 20year JGB in 1990, then they would have got a 7.5% yield. As you say, at that time the yield curve was inverted so 1year JGB yielded more at 8.6%. BUT that 20y JGB yield of 75 Yen per year would be locked in for then on as the 2010 maturing bonds were replaced by 2015 maturing bonds etc. As interest rates fell away to next to nothing, someone with short term JGB would by contrast have ended up getting next to no interest. Much more significantly the 1000 Yen of 7.5% 20y JGB bought in 1990 and replaced with new long term JGB (so as to keep the duration long) would have had a capital gain to over 9000 Yen when 20y JGB yields fell to 0.8% in 2003. Isn't that 9x capital gain (over and above any interest payments) the be all and end all of why the PP has long term bonds? Am I in a muddle or missing something?

I do very much take the point that the PP is very precariously exposed to a scenario of sustained increases in short term rates. I can imagine a Volcker 1981 style down year for the PP. I find it harder to imagine a long series of such years. It gets awkward for the government to pay real rates of interest (inflation adjusted) on outstanding debt above the rate of GDP (or rather tax intake) growth. I guess though the government can reconfigure debt arrangements any which way and so make anything possible.

[doodle]

I keep jumping back and forth between the original PP and the 5 Year rate tart ladder. I am trying to preserve my parents retirement money and I can't conclude which of the two ways would be the best to do this. 

The strongest argument against the possibility of strongly rising rates is that it would be so crushing to our economy that it would probably send us back into a deflationary tailspin in short order. The level of debt that our government is carrying would also become very unwieldy if rates rose dramatically and so it is definitely against central bank interests to let rates rise too precipitously.

In terms of preservation of capital where do the majority lean in terms of the bond barbell vs. the 5 year ladder?

[stone]

For Japanese 20year goverment bonds doesn't the 1990 (7.5% yield) to 2003 (0.8% yield) period amount to an annualized total return of about 19% per year for a thirteen year period? In addition inflation was minimal. Did any major asset class anywhere in the world do remotely as well over that thirteen year period? I'm really struggling to understand how in a deflation, the LTT component can be thought of as anything less than a cornerstone of the PP. Arguably 2008 was the first fully globalized debt fueled bubble burst and post 2008 is the first fully globalized debt deflation. Attempts may be being made to hold that deflation at bay but it seems to me very bold to total faith that it will not kick in at some point.

[stone]

I think part of the problem is that in the event of a Volcker 1981 style rate hike when cash was the only positive asset, cash doesn't do well enough for a 25% cash portion to avoid an overall portfolio loss. It hasn't mattered so far because we have only had one such year. People are touting ditching the LTT so as to increase the cash-like portion simply because only Japan has seen a scenario for which the LTT were crucial. No doubt if this was 1998 we would be talking about ditching the gold portion.
As a pie in the sky query, is there some way to make the 25% cash portion act as "hyper-cash"? Would a small portion of something like a curency ETF that was short the South African Rand give volatility that saved the day in the event of a Volker 1981 style rate hike?

[Gumby]

FYI, here are some 3-year charts — one showing a 4x25 PP and the other showing a 5-Year Ladder Rate Tart PP.

Both portfolios were started on March 18, 2008 — at the 2008 peak for the PP. All dividends are included and were reinvested into cash. All assets were rebalanced once, on Dec 29, 2008 (when it was recommended by craigr).

4x25 PP (25% GLD, 25% TLT, 25% VTI, 25% SHY)




5-Year Ladder Rate Tart PP (50% FSBIX, 25% GLD, 25% VTI)



On Dec 29, 2008 a 4x25 PP holder — having nearly erased their entire 2008 loss — would have sold their winning Long Term Treasuries to rebalance into stocks and gold.

On that same day, a Rate Tart PP holder would have had to sell from their Short Term Treasuries in order to purchase stocks and gold. As you can see from the chart, the recovery of the Rate Tart PP may have seemed a little more sluggish than that of a 4x25 PP.

From a psychological standpoint, I think it's worth noting that a Rate Tart PP holder might have blinked at the thought of rebalancing out of their cash holding into stocks and gold when things were still looking rather dark on 12/29/08 and Long Term Treasuries were the only winning investment.

On the other hand, it's also worth noting that in 2008 the 4x25 PP holder probably would have had to pay capital gains tax upon selling their winning LTTs (unless held in an IRA, of course). Whereas the Rate Tart PP holder would have had very little capital gains tax upon selling their Short Term Treasuries in 2008.

Yet, as of today, the 4x25 PP would have outperformed the Rate Tart PP by +3.44%.

MAX DRAWDOWN — using a 3/18/08 inception date:

        4x25 PP: -12.86% (Nov 11th, 2008)
 Rate Tart PP: -13.77% (Oct 28, 2008)

[align=right]DISCLAIMER: PAST PERFORMANCE DOES NOT GUARANTEE FUTURE RESULTS[/align]

[MediumTex]

Great charts Gumby!

Thanks for posting.

[Gumby]

I'd be interested to know how you calculated the rate tart ladder (RTL) values Gumby.

Sure.

You defined a treasury rate tart ladder as:

Rate tarting is a name with give over here (UK) to seeking out the best current cash deposit rates and jumping from one bank to another if more can be earned (having no loyalties and selling yourself to whoever is providing the best rates). I've just in effect adapted/extended that term to also include treasury's.

I simply used FSBIX — Fidelity's 1-5 Year Short Term U.S. Treasury Fund — as an approximation for the 5 year rate tart ladder. At least 80% of the fund's assets are invested in U.S Treasuries (currently 99.5% of the fund is invested in US Treasuries). I would have to assume that the Fund manager is doing everything in his power to get the best Treasury yields for his investors (i.e. rate tarting is his job).

Here's the fund info:

Spartan Short-Term Treasury Bond Index Fund (FSBIX) - SUMMARY
Spartan Short-Term Treasury Bond Index Fund (FSBIX) - MONTHLY HOLDINGS REPORT

I plugged the portfolio into the Google Finance portfolio tracker (which includes dividends, accrued as cash). Of course, in my examples, the dividends from both portfolios weren't reinvested into Short Term Treasuries — since the payments just accumulated as cash. However, that's true of both the 4x25 PP chart and the RTL PP chart.

If you know of a better way to (easily) track a rate tart portfolio, do let me know.

[stone]

Grumby, its great that you have cranked out the hard facts for USA 2008. Both the rate tart ladder and the 4x25% clearly worked OK then. For Japan 1990 to 2004 am I correct in thinking that using the 4x25 with 20year (or longer once available) JGB would have been much much better? On the face of it the 7.5% (sept 1990) to 0.8% (may 2004) drop in 20year JGB yields looks like an astonishing capital gain (900%). I know yields can go up as well as down but the same can be said for any asset prices and isn't the whole point of the PP  to capture such volatility by re-balancing between the uncorrelated assets.

[Gumby]

If you know of a better way to (easily) track a rate tart portfolio, do let me know.

Thanks Gumby. No I don't know an easier way. I tend to also use the 5 year as a guide for the rewards, but with it in mind that the volatility is 2.5 year and risk is cash like. Actual values are laborious to calculate for comparison purposes  

Thanks again.

Good to know!

Just curious, but how laborious is it to implement a rate tart ladder? I'm just trying to imagine a retiree who might want to spend more time enjoying golfing and less time rate tarting during their retirement years. Seems like a 1-5 year Treasury Fund would be less effort.

[Gumby]

For Japan 1990 to 2004 am I correct in thinking that using the 4x25 with 20year (or longer once available) JGB would have been much much better?

I can't say. (You'd have to ask Clive). I don't believe that Japan is a usable example of a US PP. One would never hold a US PP with 20 year Treasuries. Never. It just wouldn't work as Harry Browne intended.

And Japan would have looked very different with 30 year Treasuries back then. There's just no usable PP comparison between the US and Japan. Think of what the high-yielding 30 year Japanese Gov bonds sold in 1980 at 11% or 12% would have done to the economy (and everyone's portfolios) when they became 5 year Gov bonds! Comparing a traditional 4x25 PP to a Japanese PP really makes no sense from my perspective.

[Gumby]

Overall the two PPs — as shown in the charts — did roughly the same over a three year period. However, what is clear is that the Long Term Treasuries provided a much needed breath of life to the 4x25 PP in December of 2008. Whereas the Rate-Tart PP had nothing to lift it out of the doldrums during that same period.

We can hem and haw about what might have been in Japan — of course, we'll never know for sure — but we do know that the LTTs made a significant difference during December 2008 in the US. There's no denying that.

[stone]

Clive, 30year JGB had 2.7% yield in Feb 2002 and 1.0% yield in May 2003. Isn't that a >100% capital gain ie a "much much better result"?

[Gumby]

Overall the two PPs — as shown in the charts — did roughly the same over a three year period. However, what is clear is that the Long Term Treasuries provided a much needed breath of life to the 4x25 PP in December of 2008. Whereas the Rate-Tart PP had nothing to lift it out of the doldrums during that same period.

We can hem and haw about what might have been in Japan — of course, we'll never know for sure — but we do know that the LTTs made a significant difference during December 2008 in the US. There's no denying that.

But equally that rise was followed by subsequent declines. So it generally only helped smooth the volatility in the paper valuation of the total PP for a relatively brief period of time (in terms of the mid to longer term investment horizon).

Right. And that's why the 4x25 rebalanced out of LTTs at just the right moment — staying ahead of the RTL PP the entire time. Clearly LTTs are incredibly useful in deflationary crashes and black swan scenarios. You make it sound like Oct 2008 — March 2009 would have been an easy time to hold a RTL PP. I disagree. The RTL PP made you sweat it out for the entire time, with no relief. That's when people are most likely to bail. The 4x25 PP provided some relief and a slightly shallower loss.

You really don't see the usefulness in LTTs in a situation like that? We're talking about an extremely stressful time to be an investor in stocks and gold.

[stone]

Clive, thanks for showing me that calculator. My impression of the 30year JGB volatility was exaggerated but even the smaller reality still comes across as dramatic to me.

When you say:  "In effect the LTT's zigzag around more, but got to the same destination in the end."-

-Stocks have zigzagged around for the last decade without getting anywhere. Gold has zigzagged on a long time scale without getting anywhere (in real terms). From what I can see the essence of the PP is that it captures such zigzags by rebalancing. After all, many companies' shares do not pay dividends and companies generally go bust after 30years or so. In effect such a share is a 30 year zero coupon bond that amortizes at maturity. The only way share holders get anything from earning such shares is by rebalancing.

[Gumby]

I don't recall reading about anyone having actually rebalanced out of LTT's at the 2008 peak in practice.

Clive, as I mentioned in my previous post. On December 29, 2008, Craigr recommended, on CrawlingRoad.com, that everyone rebalance their PPs:

https://web.archive.org/web/20160324133 ... rebalance/

If you read crawlingroad.com at the time, and followed his advice, you would have absolutely rebalanced at the perfect time. It wasn't luck. It was a simple rebalancing band — and that's why I included that date as the rebalancing date.

Unless I missed it, I'm not sure I understand if you've made it clear when an RTL PP investor should rebalance their holdings in the middle of a (deflationary) crisis. Is there a hard rule you've devised that makes the RTL PP easy to follow?

[Bongleur]

>
One would never hold a US PP with 20 year Treasuries. Never. It just wouldn't work as Harry Browne intended.
>

Really?  what is the precise definition of long term then?  How often must one sell their off-the-run bonds in order to buy bonds that are still long enough to work?

Going forward you are probably not going to make the roll yield anymore, so that's a loss with every trade.

[AdamA]

How often must one sell their off-the-run bonds in order to buy bonds that are still long enough to work?

Going forward you are probably not going to make the roll yield anymore, so that's a loss with every trade.

You're supposed to buy them as close to 30 years as you can get and then sell them when they have 20 years left to maturity.

Why are you less likely to make the "roll yield" today than in the past?

[stone]

If the coupon payments are small, then the duration of the bond will be almost the time to maturity. If the coupon payments are large then the difference will be greater. So if a UK PP holder has 50year gilts with 8% coupon payments, they would not need to replace them very often whilst if they have 50year gilts with 1% coupon payments, they would need to be replaced more often in order to keep the duration as long as possible. I was thinking that with 2060 maturing 4% gilts I wouldn't need to replace them until 2030 for the PP. As I pay in to the PP, I would always get the longest dated gilts available. Have I got the right end of the stick about this? The commission is £5.99 to sell and £5.99 to buy unlimited amounts and spread is small.

[Bongleur]

How often must one sell their off-the-run bonds in order to buy bonds that are still long enough to work?

Going forward you are probably not going to make the roll yield anymore, so that's a loss with every trade.

You're supposed to buy them as close to 30 years as you can get and then sell them when they have 20 years left to maturity.

Why are you less likely to make the "roll yield" today than in the past?

The bull market in bonds is probably over; new bonds will have a higher interest rate (ie cost less to buy) so you will take a loss when you sell your old 3%  bond and buy a new bond paying 8%.

The comment about low coupons having long duration is salient.  If you presume interest rates will rise at some rate, and pick a rule about when to trade the bonds based on duration, you can estimate the loss from rolling the bonds over.

Also, if you have a 20 year bond paying 3%, but the market for 30 years is up to 8%, how well does your bond protect you from deflation?

If bonds cost money to roll over instead of make money then that cost needs to be factored into the PP performance projection.

EDIT:  The duration of the bond is not the only thing to account for.  If one holds old bonds that are now down to 20 years or so, their interest rate is still far above the current rate for 30 year bonds.  So does that mean that those high-rate bonds are even better for deflation protection than current 30 year bonds?

[AdamA]

The bull market in bonds is probably over...

Why do you say that?

[Gumby]

Clive, I've been giving the RTL PP some thought, and in a way, it has some characteristics of PRPFX (such as no Long Term Treasuries).

FYI, this is what PRPFX's targets are:

----------------------------------------------------------
Investment Category            Target Percentage
----------------------------------------------------------
Gold..................................................... 20%
Silver..................................................... 5%
Swiss franc assets.................................. 10%
Stocks of U.S. and foreign real estate
 and natural resource companies.............. 15%
Aggressive growth stocks......................... 15%
Dollar assets.......................................... 35%
================================
Total                                                   100%

PRPFX uses various ladders for its Treasuries (according to Cuggino), and the PRPFX prospectus describes its "Dollar assets" as:

Dollar assets include cash, U.S. Treasury bills and notes and U.S. Treasury bonds, and may include other U.S. dollar-denominated assets such as the obligations of U.S. government agencies, high-grade, short-term, corporate bonds and banker’s acceptances which, in the opinion of the Portfolio’s investment adviser, are secure enough to escape default even under deflationary economic conditions. The average length to maturity of the Portfolio’s net dollar assets will not exceed fifteen years and corporate bonds will have a Standard & Poor’s rating of “A”? or higher and a remaining time to maturity of twenty-four months or less.

and in case you were wondering, the "Swiss Franc assets" can be anything (as a hedge against the dollar):

Swiss franc assets consist of deposits of Swiss francs at Swiss or non-Swiss banks and the bonds and other securities of the federal government of Switzerland of any maturity;

...So, indeed, it looks like you may may have found a nice and simple alternative to PRPFX and the PP. I would imagine that upping the stock to a 30% allocation would give you performance that is quite similar to PRPFX. (FYI, I had included PRPFX in my charts that I posted, so you can see the comparison).