Donate to Remove ads

Got a credit card? use our Credit Card & Finance Calculators

Thanks to johnstevens77,Bhoddhisatva,scotia,Anonymous,Cornytiv34, for Donating to support the site

Ladder using stock/gold/bonds and relative valuations

Stocks and Shares ISA , Choosing funds for ISA's, risk factors for funds etc
Investment strategy discussions not dealt with elsewhere.
1nvest
Lemon Quarter
Posts: 4323
Joined: May 31st, 2019, 7:55 pm
Has thanked: 680 times
Been thanked: 1316 times

Ladder using stock/gold/bonds and relative valuations

#398625

Postby 1nvest » March 24th, 2021, 12:11 pm

As a follow up to viewtopic.php?p=398502#p398502 thought I'd post about a ladder approach.

More conventionally a ladder will, for instance for a 10 year ladder, buy a 10 year gilt, hold that to maturity and then roll the proceeds from the maturing bond it into another 10 year gilt, doing so for each of 10 rungs. So each year one bond/gilt matures and gets rolled into the ongoing 10 year gilt yield. There are two ways that can be valued, you can look at the actual price of the bonds each year and tally up all of the capital gains/losses and add in interest received ...etc. i.e. 'marked to marked'; Or as you're not selling/trading bonds - just holding to maturity, you can approximate the yearly gain as the average of the ten 10-year-yields i.e. the yield that each bond provides when held to maturity, which is known at the offset (time of buying the bond). Those that like to hold such ladders suggest that you get the average of the ten year yields with a average maturity of 5 years, along with very smooth valuations over time (no negative years).

That aside, here I'm looking at a similar ladder/rungs approach, but rather for a 3 rung (3 year) approach, and instead of rolling into another 3 year bond as each rung matures we roll into the asset with the better valuation at the time, either stock, 10 year bond or gold. Here we are marking to market and trading i.e. holding stock for a fixed 3 year time period if stock is the indicated better valued choice at the time ... or gold ... or bonds.

For the relative valuation I'm using a simple comparison/method of opting for gold if the Dow/Gold ratio is 20 or higher, otherwise picking either stock or gold according to whichever has the higher yield when comparing stock earnings yield (inverse of PE) to the prevailing 10 year Gilt yield at the time. If for instance at the time the stock PE is 25 ... so a 4% earnings yield (inverse of PE) whilst 10 year gilts are paying 3%, then stock is considered as being the better valued choice.

So to initially load into that we use the relative valuation comparison as per 2 years ago for the first rung, the valuation 1 year ago for the second rung, and the current valuation as at the start date ... to determine what assets each of the three rung will hold, initially allocating a third of total capital to each. We then count that first rung as though it had already been held for 2 years so we're just going to hold that for a year before considering it as having matured and due to be sold and the proceeds rolled into whatever the valuation indicates at that time. Similarly we consider the second rung as having 2 years remaining before we 'rotate' it (consider it as having matured) ...etc.

For testing purposes I used US data due to better data availability, running a backtest from the start of 1974. Making some observations ...

Overall that achieved a better reward than all-stock (11.8% annualised versus 11.2% annualise - note that these are all total returns, dividends/interest reinvested) and did so with less volatility, and broadly averaged (weighted average) overall exposure of 40% stock, 40% bonds, 20% gold. Similarly it also beat constant 40/40/20 stock/bond/gold (yearly rebalanced) by a considerable margin (9.6% versus 11.8% annualised).

Looking deeper into the gain attributes it became apparent that the individual rungs saw quite dissimilar overall results. The best rung achieved 13.1% annualised whilst the worst provided 9.8% annualised. The reason for that was that the best case had a good start and saw a sizeable up/gain the first year, whilst the worst rung saw a loss in its first year. The circumstances were such that it was like the best rung having twice as much as the worst rung a year after having started. Which reasonable became locked in and reflected in overall longer term annualised rewards.

My take from that study is that rather than starting with what was indicated as the choice three years ago for one rung ...etc. that perhaps simply starting with a third in each of stock/bond/gold is more inclined to reduce early years sequence of returns risk. And then a year later decide which of the three you might consider as being the one that had 'matured' and roll that into whatever was being indicated as the better valued choice at that time. Leaving 2 candidates from which the next years review might select one of the two as being the 2nd ladder rung.

Primarily for that 1974 based backtest it was a case of gold doing well, stocks doing poorly in the early years. For other start dates such as 1980 when the Dow/Gold ratio was down at near 1.0 levels (suggestive of gold expensive/stocks cheap) and that would be more inclined to see a reversal i.e. the rung started with gold doing relatively poorly compared to the rung started with stocks.

Such a ladder might be considered as being more appropriate for lump sums. When you're young, accumulating, you'll add savings over time - cost average in which reduces/eliminates the risk of lumping in at a relative high, or lumping out (selling) at a relative low (when you retire you'll draw funds yearly - cost average down/out). However equally a existing buy and hold position is no different to costlessly selling and lumping all-in again each and every day, and when a sizeable amount any additions (savings) averaged in on top might be near insignificant amounts relative to the whole.

How well did the relative valuation method work, well it can at times end up all in one asset and that was the case for gold for the years 2000 to 2006 inclusive. A period that was bad for stocks (2% annualised) whilst gold gains were reasonable (12%). 2007 and gold was reduced to add around 16% stock, 2008 saw that continue to 65% stock, 35% gold and at the end of 2009 it was a case of 75/25 stock/bonds being held. i.e. it averaged into stocks across the 2008/9 financial crisis period. Remained at pretty much at 75/25 stock/bond to the end of 2012 before then pretty much having been all-stock since. But with a close threat of moving some into gold (Dow/Gold ratio up near 20 levels) - perhaps suggestive of relatively high stock/low gold prices/valuations.

All told and it seems like a reasonable/simple model for a valuations based ladder might be employed to good effect. However I'm not keen on the idea that it can be all-in on single assets at times - too much concentration risk for my liking (concentration risk is one of the greatest of risks). Maybe then as just part of a overall allocation. Core holdings of 25% bonds, 25% stocks, and 50% allocated to the above ladder style. Which might align with Ben Graham's advice of no more/less than 75/25 or 25/75 stock/bonds. Based on the ladders weighted averages of 40/40/20 stock/bond/gold that would have averaged 45/45/10 stock/bond/gold overall since 1974, but perhaps with rewards bolstered by the valuation method as per how it bettered constant weighted by more than 2% in the above backtest (i.e. which when weighted to 50% of the total portfolio = 1% portfolio uplift which would have closed the gap compared to all-stock in that link). Whilst having lower portfolio volatility (better risk adjusted reward).

tjh290633
Lemon Half
Posts: 8209
Joined: November 4th, 2016, 11:20 am
Has thanked: 913 times
Been thanked: 4097 times

Re: Ladder using stock/gold/bonds and relative valuations

#398772

Postby tjh290633 » March 24th, 2021, 10:56 pm

My way is to construct a cash flow for each year. A new gilt is bought with the previous year's sale, the interest from the gilts held is disbursed and at the end of the year the maturing gilt is sold. On this you can superimpose portfolio value at start and finish, if you wish.

This can be built into a longer term cash flow if required, which does not need the starting and closing valuations for each year, as they cancel each other out.

TJH

spasmodicus
Lemon Slice
Posts: 259
Joined: November 6th, 2016, 9:35 am
Has thanked: 65 times
Been thanked: 117 times

Re: Ladder using stock/gold/bonds and relative valuations

#398843

Postby spasmodicus » March 25th, 2021, 10:19 am

Thanks for those observations, 1nvest.

For those unfamiliar with the Dow/Gold ratio (that includes me), here is a link to a useful illustration
https://www.bullionvault.com/gold-news/dow-gold-122020181

The surprising thing about this is
a) the long term running running average of the ratio has changed comparitively little in the las 100 years, from about 10 in 1920 to around 15 today.
b) the difference between the maximum of around 42 and the minimum of about 2 is around a factor of 20

The periodicity is very long, with about three major cycles in 100 years. A period of 30 years or more is quite long compared with the length of many people's investing careers. In this respect it has something in common with the Schiller CAPE method fo evaluating the USA stock market, or maybe these are both just examples of reversion to the mean?

regards,
S

1nvest
Lemon Quarter
Posts: 4323
Joined: May 31st, 2019, 7:55 pm
Has thanked: 680 times
Been thanked: 1316 times

Re: Ladder using stock/gold/bonds and relative valuations

#398919

Postby 1nvest » March 25th, 2021, 2:25 pm

Pre 1971 was a different paradigm where the price of gold was very stable - i.e. fixed and you could exchange US$ into gold and vice versa at a constant fixed rate. Under such circumstances it made more sense to hold money deposited earning interest that could then buy back more ounces of gold at any time at the fixed rate. Money and gold were in effect the same thing. Since that coupling was broken (by president Nixon as a means to pay down the cost of the Vietnam war) the price of gold has been massively volatile. Better partnered 50/50 with stocks, a barbell of two extremes that combines to form a central 'bullet' as in how a 20 year and 1 year gilt barbell combine to form a 10 year gilt bullet. As such its best to just look at since 1972 Dow/Gold variability.

With recent historic low interest rates prices of stocks, bonds, houses etc. might be considered as being pretty high. Aided by progressive declines from high 1980's interest rates down to present low rates. Buffett considers interest rates to be financial 'gravity'. As such forward rewards are inclined to be a lot lower than since the 1980's, perhaps even with rising rates seeing a negative pressure on prices.

1M of wealth invested with a view to drawing 4% at present valuations has higher risk than 500K invested with a 4% SWR when interest rates are much higher. A rise in inflation could see 50% or more of values disappear. We've transitioned from a era of targeting growth, to a era where wealth preservation is more the priority.

If you have enough that 1M and a 2% SWR is enough, then fine. If valuations do halve and your income rate doubles to 4% you'll be fine, just on paper 50% poorer than before. 4% at present valuations halving and requiring 8% income is much more of a risk.

As a wealth preserver a asset allocation such as US stock, UK stock, gold, cash, bonds is more resilient to downsides. 20% in bonds might be drawn at 2%/year for 10 years to leave UK stock/US stock/gold/cash 25% weightings, a form of 50/50 portfolio and where gold is the asset that is more inclined to spike when the others fall (such as via large negative real interest rates).

Whilst the gold cycles do indeed tend to be long over the decade(s) of gold being low/down reducing stocks/bonds that have done well to add more ounces of gold can pay back massive dividends sooner or later.

Consider a poor 10 years for stocks/bond where the actual portfolio value sees -4% real (after inflation) annualised outcome, whilst consumer prices are rising at perhaps +4%. After 10 years with the same inflation adjusted income having been drawn along with the low/poor returns from stocks over that decade you might be down at 25% to 33% of the former inflation adjusted portfolio value remaining. Typically under such situations in the past gold has tended to do very well, perhaps treble in value.

Whilst gold is a drag factor over some periods, it can be a saviour in other periods. Insurance. Overall that insurance cost washes - mathematically. In practice however if those without insurance see their portfolio values fall to unsustainable levels, fail, then it matter not that stocks might subsequently make great compensatory gains as even 1000% on 0% capital is still nought.

Your house might only have a risk of burning down once in a century or less, but fire insurance is good policy as you just never know.

The Golden Butterfly for a UK investor might be 20% in each of FT250, US S&P500, gold, cash, gilts ... quite resilient whilst rewards are still reasonable (here's a US based example click the inflation adjusted tickbox and look for instance at 1972 to 1982 when all stock total returns with dividends reinvested was down -20% over that decade and significantly more if you were also taking a income from the portfolio). Rather than just fixed weighting each of those assets you might apply more discretionary measures, overweight some, underweight others in accordance with perceived relative valuations. For instance still continuing to hold gold when it took just a little over a ounce to buy the Dow is indicative of either very high gold prices or very low stock prices.

Over the 1980's and 1990's you might broadly have seen gold halving in nominal price and some shares repeatedly being sold to add to gold such that you may have ended up with 10 times more ounces of gold being held. Whilst that dragged down gains compared to had no gold been held, rewards were still reasonable. Then from 2000 to 2011 the price of gold spiked increasing around four-fold, notably over years when stock gains were relatively low. Those sorts of progressions aren't fluke, rather they're circumstantial. When stock/bonds/cash outlooks all look bleak and interest rates are intentionally being kept low in anticipation of inflation eroding away such capital, then investors will tend to favour gold. When stocks are doing well and growth prospects look good, investor will tend to flight from gold and into stocks. If as policy to permanently hold/trade (rebalance) gold, then you'll provide liquidity to both sides of those movements and where sooner or later the insurance qualities might pay off - in a big way.

Looking at US stock (for UK investor/UK inflation) and gold as just being two 'assets' and you might average the yearly years best and worst of the two. The average of each decades best and worst figures provides a indicator of the overall average real reward those assets provided. This image further shows how ounces of gold and number of shares drifted over time from 'trading' i.e. rebalancing back to 50/50 each year those assets.

Image

It matters little what the actual yearly best or worst assets actually where, or how often one was the best or worst. The far right 'average' real figure is the main indicator of outcome and the consistency of such outcomes. In that images case gold was best in around 20 out of 48 total years. In other cases that might flip the other way around. Sometimes you might be able to 'predict' which may be the better - such as in 1980 when the Dow/Gold was down at near 1 levels, or in 1999 when the Dow/gold was up at 40 levels. And if your predictions turn out to be more right than wrong then adjusting weighting accordingly might add to overall rewards.

torata
Lemon Slice
Posts: 521
Joined: November 5th, 2016, 1:25 am
Has thanked: 203 times
Been thanked: 210 times

Re: Ladder using stock/gold/bonds and relative valuations

#399137

Postby torata » March 26th, 2021, 9:37 am

tjh290633 wrote:My way is to construct a cash flow for each year. A new gilt is bought with the previous year's sale, the interest from the gilts held is disbursed and at the end of the year the maturing gilt is sold. On this you can superimpose portfolio value at start and finish, if you wish.

This can be built into a longer term cash flow if required, which does not need the starting and closing valuations for each year, as they cancel each other out.

TJH


Terry

I have a distinct recollection that you had tried a gilt ladder but given it up, or at least the 5-year version.
Is my memory wrong or did you go back to it? Or is this a '1-year' gilt ladder, or in official parlance, a gilt stool? :)

torata

tjh290633
Lemon Half
Posts: 8209
Joined: November 4th, 2016, 11:20 am
Has thanked: 913 times
Been thanked: 4097 times

Re: Ladder using stock/gold/bonds and relative valuations

#399282

Postby tjh290633 » March 26th, 2021, 4:20 pm

torata wrote:
tjh290633 wrote:My way is to construct a cash flow for each year. A new gilt is bought with the previous year's sale, the interest from the gilts held is disbursed and at the end of the year the maturing gilt is sold. On this you can superimpose portfolio value at start and finish, if you wish.

This can be built into a longer term cash flow if required, which does not need the starting and closing valuations for each year, as they cancel each other out.

TJH


Terry

I have a distinct recollection that you had tried a gilt ladder but given it up, or at least the 5-year version.
Is my memory wrong or did you go back to it? Or is this a '1-year' gilt ladder, or in official parlance, a gilt stool? :)

torata

I have just updated it as the anniversary date has just passed. The XIRR figure for the past 12 months is 0.31%, the first time it has been positive for 8 years. It's a 5-year ladder.

TJH

Hariseldon58
Lemon Slice
Posts: 835
Joined: November 4th, 2016, 9:42 pm
Has thanked: 124 times
Been thanked: 513 times

Re: Ladder using stock/gold/bonds and relative valuations

#399320

Postby Hariseldon58 » March 26th, 2021, 6:11 pm

1nvest wrote:


For the relative valuation I'm using a simple comparison/method of opting for gold if the Dow/Gold ratio is 20 or higher, otherwise picking either stock or gold according to whichever has the higher yield when comparing stock earnings yield (inverse of PE) to the prevailing 10 year Gilt yield at the time. If for instance at the time the stock PE is 25 ... so a 4% earnings yield (inverse of PE) whilst 10 year gilts are paying 3%, then stock is considered as being the better valued choice.





I can see picking Gold if its cheap compared to the Dow, but then we are back to Stock or Gold....


"otherwise picking either stock or gold". Should this reference to Gold be Gilts ?

1nvest
Lemon Quarter
Posts: 4323
Joined: May 31st, 2019, 7:55 pm
Has thanked: 680 times
Been thanked: 1316 times

Re: Ladder using stock/gold/bonds and relative valuations

#399405

Postby 1nvest » March 27th, 2021, 12:02 am

After some deep/broad analysis (over a century of data) ... conclude that it all tends to wash, might as well just allocate to fixed weightings, periodically rebalanced to realign risk/exposure and just run with that.

Periodic negative 10 year periods occur relatively infrequently, 1 in 50 years or so, but in the scale of investment horizons that becomes more like a > 50/50 chance of encountering such during your investment lifetime.

If you're drawing 4% SWR and you see a annualised -5% for a decade then at the end of the ten years you'll be down to 20% of the inflated adjusted start date amount remaining. Pretty much a failure situation. The events driving such declines pretty much hit most assets, but where gold can be the exception.

With a rule of moving into portable physical assets such as silver/gold ..etc. during World Wars, along with a third each UK stock, US stock and gold, then such risk was pretty much eliminated. Perhaps surprisingly the 'cost' of carrying gold was negative i.e. such a three way split on average across all ten year periods hit a 0.6% annualised HIGHER reward than all UK stock. All stock however did have better 'great' periods. Of the order 15% annualised versus 10% annualised real gains at times. However that was a rare event - that few actually would have seen but that biases the broader compounded figure for all-stock when measured across a long period of time that include that 'right tail' (good period).

Lowering the SWR figure can also make a huge difference. At 3.5% for instance then historically you were into perpetual/heirs type territory where even a bad decade is a relatively mild event for the UK/US/gold asset allocation. On average supplemented with 4% real gains on top (i.e. optional/discretionary additional spending; Or that surplus real could have been left to expand real wealth considerable for heirs).

Relative value based rotating/adjustments didn't broadly add value, but increased risk (lower risk adjusted reward) :(


Return to “Investment Strategies”

Who is online

Users browsing this forum: No registered users and 7 guests