Why does anyone invest in bonds long term?

[audreyh1]

Quote:

Originally Posted by Lsbcal Audrey, you are making good points here. I had brought up the 1929 case because of the original thread title and direction. I don't think VPW presents a magic formula but it is great for me because I can follow individual worst case performance. I think the developer (Longinvest) was quite careful in his historical data set. Plus it is an open tool ... pretty unique. VPW offers a decent starting strategy and those interested should head to the excellent Bogleheads forum discussions on this plus the current spreadsheet source.

VPW is designed to draw the portfolio down to zero over the time period given, and as such tends to use higher withdrawal rates. So I guess I don't understand your concern of seeing a % remaining portfolio drop in half or even 60%, over a 15/16 year time period when it ends up recovering to being only ~down half in the worst case for the 4.35% withdrawal case.

Lets look at the 1906 case and compare the sequences and see what happens at the worse case of 16 years (end 1920/beg 1921).

Looking at VPW for 50/50 for 30 years starting 1906, by year 16 the $1M portfolio is down to $295K, so that's >70% down in 16 years before it starts to recover a little in real terms, but as planned it gets spent down to zero at the end of 30 years. Income has dropped form $51K to $24K by year 16 (withdrawing 7.9%). The withdrawal rate starts at 5.2% (and keeps increasing) which is why the portfolio drops so much.

Compare this to the 5% remaining portfolio withdrawal method which drops to $374K after 16 years before recovering. Income starts at $50K and drops to just under $19K before recovering to higher income, and the portfolio recovers to a $735K terminal value at 30 years.

We can also compare it to the 6% withdrawal case which withdraws higher income. This drops to $319K at 16 years which is closer to the VPW low point at 16 years of $295K. It starts at $60K income and drops to $19K income before recovering to higher income, and ends with a $535K portfolio after 30 years.

So - I guess I don't get your point about concerns over large portfolio drops and a retiree's reaction. You are certainly going to see them with VPW as it is designed to increase withdrawals and drive the portfolio to zero. Bad sequence of returns can still drive the VPW portfolio down to a pretty low level by the half-way point. It obviously has higher total income over the period compared to the % remaining case which is designed not to go to zero.

With VPW if you increase your depletion years such that the portfolio depletes after 40 years instead of 30 so that you have a higher terminal value at 30 years - you end up with $518K which is closer to the 6% remaining portfolio method terminal value. You start with a 4.5% withdrawal rate, your income starts at $45K and drops to $21K at year 16 before increasing again. The portfolio has dropped to $364K by that point, so it's still ~64% down before recovering somewhat. This is somewhere between the 5% and 6% remaining portfolio methods which both start at higher initial income and end with higher portfolio values after 30 years.

Not seeing any magic bullet here. VPW is a good choice if you want to be aggressive about spending down the portfolio in later years and your income won't drop quite as much after a bad sequence because the withdrawal rate increases every year, but you still could be looking at real income drops of ~50% and drops of ~64 to 70% in portfolio value at half time if you stumble into a particularly a nasty scenario.

[audreyh1]

Quote:

Originally Posted by Lsbcal I don't think VPW does avoid heavy drawdowns. I use 1966 as my worst case because it seems to represent the US bias to avoid depression at the expense of some inflation. Also it is closer to our economy then the more agrarian one of the 1930's. I think in the 1930's something like 30% of the workforce was on the farm. Audrey, we have had a discussion on another thread about your withdrawal strategy. In the context of VPW, I've used its algorithm to suggest how to build up a reserve fund for the bad downturns. So although VPW suggests ramping up the withdrawal rate, in practice I withdraw at more like a constant 3.2%. The balance then goes into a low risk account (like short term bonds) for filling in the lean income years. For an example, here is a VPW 60/40 with a start year of 1966 and a depletion of portfolio by age 110. The portfolio low comes 17 years into the retirement at about 41% of the start portfolio. In our personal case, VPW says we can draw 4.4% in 1966 but we would have spent 3.2% in 1966 and kept the extra 1.2% for future income needs should the portfolio experience a bad decline. One probably doesn't have to build up this reserve indefinitely. There are lots of ways to manage this, but the basic idea is not to have too small an income if the portfolio declines severely.

I was working on my prior comparison post before I saw your 1966 scenario post.

I have also studied the 1966 case extensively. Looking at that showed me to expect draws down to 45% of the original portfolio for draws of 4.35%, lower for higher withdrawal rates [40.6% for 5%, 34.3% for 6%], at around as you say the 17 year mark. The 1966 portfolio worst drops end at around 5% higher than the pre 1920 cases. So our two methods seem to track more or less, and both of us have come up with schemes to manage the income drop while the portfolio recovers. Mine also counts on letting excess income (from spending less than I draw) accumulate in short term reserves, although I've already established what I think is a reasonable reserve to provide me a floor and gradually drop down and hold at 75% of our current budget during the nastiest run using the 3.5% of remaining portfolio withdrawal case.

I like the early nastier runs simply because to me they give cases that outlie recent history and some answer to "what if it's worse in the near future than the recent past?" question. If maybe the US mechanisms for managing the economy weren't operating so smoothly anymore or we lost our global edge or whatever, and we returned to much nastier times. The answers told me to plan for worst case draws down to ~40% rather than ~45% (4.35% case), so some additional hedging.

[Lsbcal]

Quote:

Originally Posted by audreyh1 ... I like the early nastier runs simply because to me they give cases that outlie recent history and some answer to "what if it's worse in the near future than the recent past?" question. If maybe the US mechanisms for managing the economy weren't operating so smoothly anymore or we lost our global edge or whatever, and we returned to much nastier times. The answers told me to plan for worst case draws down to ~40% rather than ~45% (4.35% case), so some additional hedging.

I had not done the 1906 case because it seemed too far back to be relevant today. But using that date with the VPW + smoothing, it seems like a 45% worst case portfolio low (16 years into the simulation) would be about right.

If someone wants me to, I will post the detailed simulation here.

[Huston55]

Quote:

Originally Posted by ERD50 Not that I'm recommending 100% equities (but I wouldn't argue against it either), but what did the HELOC cost him? I'm pretty confident that 100% stocks minus the cost of the HELOC put him way ahead of a more conservative AA in the long run. -ERD50

IIRC, there are a few ER.org members who do something similar.

I think Nords uses home equity as investment leverage but, don't recall the details

[audreyh1]

Quote:

Originally Posted by Huston55 IIRC, there are a few ER.org members who do something similar. I think Nords uses home equity as investment leverage but, don't recall the details

Nords has spending covered by pensions, so they could get by without income from their investments if necessary. At least that was his situation many years ago.

When someone has all expenses covered by pensions, annuities and/or SS, they are in a different ball game with respect to their investments.

[audreyh1]

Quote:

Originally Posted by Lsbcal I had not done the 1906 case because it seemed too far back to be relevant today. But using that date with the VPW + smoothing, it seems like a 45% worst case portfolio low (16 years into the simulation) would be about right. If someone wants me to, I will post the detailed simulation here.

With 60/40 VPW I'm seeing a portfolio drop to $376K in year 16 (a 37.6% worst case low). I'm pretty sure my VPW spreadsheet matches yours (except that I set SS to zero) because my 1966 case matches yours.

So I guess your smoothing scheme is taking care of limiting the drop somewhat.

PDF attached. Worst case year is highlighted in yellow. The remaining portfolio in year 31 is marked in green to compare to remaining portfolio after 30 years for other withdrawal methods.

[Lsbcal]

Quote:

Originally Posted by audreyh1 With 60/40 VPW I'm seeing a portfolio drop to $376K in year 16 (a 37.6% worst case low). I'm pretty sure my VPW spreadsheet matches yours (except that I set SS to zero) because my 1966 case matches yours. So I guess your smoothing scheme is taking care of limiting the drop somewhat. ...

The smoothing scheme is nothing more then taking some unspent VPW allowed withdrawals and putting it into a low duration account. That could be money market, CD's, short term bonds, etc. Thus if the bond markets decline or equities decline the smoothing fund is not much affected.

In the case of the 1906 series, the money could build up over several years before it is needed. So that account could grow to be maybe 15% of the portfolio which is not market sensitive. That would explain why the overall portfolio is drawn down less. See the last column in example below.

Here is a screen shot of the VPW + smoothing example. I've just taken VPW and added some columns to demonstrate the usage. It's going to be a very individualized set of numbers for each of us ... different risk tolerances, different portfolio sizes, different spending needs, etc.

[audreyh1]

Quote:

Originally Posted by Lsbcal The smoothing scheme is nothing more then taking some unspent VPW allowed withdrawals and putting it into a low duration account. That could be money market, CD's, short term bonds, etc. Thus if the bond markets decline or equities decline the smoothing fund is not much affected. In the case of the 1906 series, the money could build up over several years before it is needed. So that account could grow to be maybe 15% of the portfolio which is not market sensitive. That would explain why the overall portfolio is drawn down less. See the last column in example below. Here is a screen shot of the VPW + smoothing example. I've just taken VPW and added some columns to demonstrate the usage. It's going to be a very individualized set of numbers for each of us ... different risk tolerances, different portfolio sizes, different spending needs, etc.

So basically you are starting with a 3.1% withdrawal rate instead of the 4.4% withdrawal rate allowed by VPW. And that difference is being allowed to accumulate as a reserve?

In general it looks like you are cutting your initial income available for spending to ~70% so that you can set aside 30% in short-term reserves for the bad years? Then when the portfolio drops enough you are spending the full amount and supplementing from these reserves?

So instead of dropping to $376K, the portfolio only drops to $453K at year 16, a worst case draw down of only 55% instead of 62%. But this is accomplished by much lower income available for spending during the first decade+. One thing I don't see is the annual spending income in this scenario. Does this improve the average or median spending over say the first 30 years?

[Lsbcal]

Quote:

Originally Posted by audreyh1 So basically you are starting with a 3.1% withdrawal rate instead of the 4.4% withdrawal rate allowed by VPW. And that difference is being allowed to accumulate as a reserve?

Yes, nothing magic here.

Quote:

In general it looks like you are cutting your initial income available for spending to ~70% so that you can set aside 30% in short-term reserves for the bad years? Then when the portfolio drops enough you are spending the full amount and supplementing from these reserves?

Right, the spending in this fictitious example is set at 65,000 inflation adjusted. VPW suggested 78,000 in the first year but cannot provide the 65,000 in some later years (like years 13 to 18 in the example).

Quote:

So instead of dropping to $376K, the portfolio only drops to $453K at year 16, a worst case draw down of only 55% instead of 62%. But this is accomplished by much lower income available for spending during the first decade+. One thing I don't see is the annual spending income in this scenario. Does this improve the average or median spending over say the first 30 years?

The spending is 65,000 inflation adjusted. I don't know about improving the median spending but it does cushion the portfolio in a severe downturn. To me the need to spend at a certain level is more important then to spend much higher some years.

Audrey, I thought you were doing something like this though not necessarily using the VPW algorithm i.e. you set up an off-portfolio smoothing account which is safe from market forces.

For us, at our current portfolio size a 3.2% spending level looks good with plenty of fun spending money. It could be 3.0% or 3.5% but somewhere in there. I am just trying to find a regular way of setting up a smoothing fund while the market and our portfolio are fairly high. Yes, this means we won't spend the 4.4% VPW allowed spending but we fortunately are not in need of all of that. So again, nothing really magic here.

I suppose one could just set up a smoothing fund initially and use the VPW algorithm with the rest of the portfolio. Then draw from it if needed or spend it if the market moved sideways or up.

[audreyh1]

Quote:

Originally Posted by Lsbcal Audrey, I thought you were doing something like this though not necessarily using the VPW algorithm i.e. you set up an off-portfolio smoothing account which is safe from market forces. I suppose one could just set up a smoothing fund initially and use the VPW algorithm with the rest of the portfolio. Then draw from it if needed or spend it if the market moved sideways or up.

I have been accumulating unspent funds in short-term reserves simply because we aren't spending as much as we are withdrawing, but being fully aware that those funds could really help if things went south for a few years. I only formalized this recently, setting aside a reserve specifically for supplementing bad years and investing it in long-term CDs. I already had these funds available, so I don't need to build the reserve from future draws.

I have come up with a combination of using a short-term reserve and belt-tightening to 75% of budget as a way of handling the absolute worst case scenarios. Looking at the worst case scenarios and our current spending levels, I calculated that an amount around one year's budget was sufficient.

At current portfolio levels and our withdrawal rate of 3.5%, we have only been spending at most 80% of the after tax income drawn from our portfolio over the past five years.

So the portfolio could drop ~20% before our current spending is even affected. Obviously, in the worst case scenarios it can drop more than that. So I created the reserve to supplement income in the worst cases, and this doesn't kick in until after tax income drops below the current 80% level. Once the reserve drops by 1/3, we have to cut our budget 12.5%. If the reserve keeps dropping to the 2/3 point, we have to cut our budget another 12.5%. Modeling several bad years it looks like that is as far as we need to cut, and that the portfolio will start to recover before the reserve runs out. During recovery we don't increase spending until we replenish the reserve to 2/3 down point, and then to the 1/3 down point.

We don't need to tighten our belt as much if we set aside a larger reserve, but I decided that it was a good idea to do a combination because a large reserve would only be needed in the few truly horrific scenarios. Hopefully little belt tightening will be needed, and if so it will be brief.

[rayvt]

[ERD50 mentions, but does not recommend, a 100% stock portfolio]

There is another way to protect your portfolio than just having bonds and rebalancing. Consider that the main benefit of the bond allocation is as a safe-harbor for your money when stocks are tanking.

But there is another possible safe-harbor when stocks are tanking. Instead of mechanically swithing to bonds as a side-effect of rebalancing a 60/40 or 80/20 allocation, skip bonds altogether and use Meb Faber-type SMA signal to move from 100% stocks to 100% cash.

I plugged into my spreadsheet a 60/40 portfolio (S&P500/T-bills) starting Jan 1966, initial balance $100,000, at 4% SWR withdrawals.

After 30 years (Dec 1995) the portfolio value was $133,000.

A 100% stock portfolio went into a unremitting decline starting Sept 1987, at the 20th year, and went to $zero on Dec 1994. Not good.

A 100% stock portfolio with a 10-month SMA timing signal had ending 30-year balance of $825,000.

$133,000 vs. $825,000.

So you *can* safely have a 100% stock allocation, as long as you have a method that takes you out of stocks in a bear market.

That method could be mechanically moving money to bonds by rebalancing regardless of market conditions -- or mechanically moving money to cash using a moving average timing signal.

The spreadsheet: https://www.dropbox.com/s/cbzvg74iye...d-IUL-test.xls

[Digger1000]

I guess I don't get some of these calculations. This weekend I started in 1966 and went year by year entering the yearly return(up or down)for the S&P 500 and also entered my yearly withdrawal. I went from 1966 to 1983 only because as I mentioned earlier I had done this exercise too for retiring Jan 1,2000 up until today which was 17 years. I ended up nearly doubling my money by being 100% in stocks from 66 to 83. As you know the market performed remarkedly well in the time beyond where I stopped in 83. The worst point in the exercise I was down 41%.

[audreyh1]

Quote:

Originally Posted by Digger1000 I guess I don't get some of these calculations. This weekend I started in 1966 and went year by year entering the yearly return(up or down)for the S&P 500 and also entered my yearly withdrawal. I went from 1966 to 1983 only because as I mentioned earlier I had done this exercise too for retiring Jan 1,2000 up until today which was 17 years. I ended up nearly doubling my money by being 100% in stocks from 66 to 83. As you know the market performed remarkedly well in the time beyond where I stopped in 83. The worst point in the exercise I was down 41%.

Were you accounting for inflation? That's what drives funds to zero in the 4% constant spending case.

Here is the FIRECALC output for the 4% constant withdrawal case - withdrawal is adjusted for inflation each year to maintain constant spending. Portfolio is 100% stocks using total stock market. Starting year is 1966. Portfolio drops below 42% in 1977, and is drawn down to 0 in 1986. I added the Portfolio Remaining column at the end. Edited to add: I did a quick check and with 100% S&P500 instead of total stock market the portfolio didn't go to zero until 1988.

In contrast, the 60% total stock market, 40% intermediate bonds doesn't run out until 1991. Still runs out, but buys you another 6 years. Edited to add: or 4 more years compared to 100% S&P500 portfolio.

I'm not sure why FIRECALC shows the starting portfolio to be $958,616 instead of $960,000 = $1,000,000 - initial $40,000 draw. $958,616 means withdrawal of $41,383 was used instead. This is $40K with the 1966 inflation factor applied which to my mind would be the withdrawal used in 1967. Maybe someone else here knows the reason for that discrepancy. But that is tiny compared to the overall portfolio behavior.

[Lsbcal]

Quote:

Originally Posted by rayvt [ERD50 mentions, but does not recommend, a 100% stock portfolio] There is another way to protect your portfolio than just having bonds and rebalancing. Consider that the main benefit of the bond allocation is as a safe-harbor for your money when stocks are tanking. But there is another possible safe-harbor when stocks are tanking. Instead of mechanically swithing to bonds as a side-effect of rebalancing a 60/40 or 80/20 allocation, skip bonds altogether and use Meb Faber-type SMA signal to move from 100% stocks to 100% cash. ...

I could see this market timing method used as a partial bond alternative. Just about all retirees would not be keen on the implementation details:

1) Must look at the end-of-month comparison details on a spreadsheet or other tool. Use a consistent set of rules to get in and get out of equities.
2) Must be ready to act immediately, no second thoughts.
3) Must not let life events or vacations get in the way of implementation at the end of each month.
4) Must not be upset by occasional whipsaws (last one was in late 2015 I think).
5) Must be willing to follow through on this plan for decades and not be bothered by the long periods where buy-hold might have better returns.

Then if one felt all this is doable, maybe set aside 10% to 20% of the portfolio for this.

[ERD50]

Quote:

Originally Posted by rayvt [ERD50 mentions, but does not recommend, a 100% stock portfolio] ...

Just to be clear, by "not recommend" I mean only that. IOW, I'm agnostic about it. I wouldn't "un-recommend" it either.

From what I recall, the people on this forum that are 100% equities understand why they are doing it, and accept the volatility and lack of cushion. So I say fine (not that it matters what I say!). But I wouldn't go out of my way to recommend it to anyone either, I personally feel a bit more comfortable with the middle area of the FIRECalc AA chart, flat from about 40/60 - 95/5, and I like being on the high side of that, around 75-80 equities.

-ERD50

[OldShooter]

Quote:

Originally Posted by rayvt ... So you *can* safely have a 100% stock allocation, as long as you have a method that takes you out of stocks in a bear market. ...

Absolutely true, the key words being "as long as you have."

I invested a small amount of Google time in this "Meb Faber-type SMA signal" thing and was unable to find any examples where it successfully predicted future market action over a reasonable period. That didn't surprise me because, if there were a scheme that really worked, enough of the 100,000+ active managers would be using it that we would see their effect in the market. And, ultimately, it would be defeated simply by the volume of people trying to exploit it.

Finding schemes for market timing and for individual stock strategies that backtest well is easy. They are a dime a dozen and not even worth the 0.83¢ price. I'm absolutely certain that I could find a few variables like the price of salt in Fiji, coal shipment loadings in Australia, and the ruble/rupee exchange rate which, when plugged into an equation of my creation, would backtest to near perfection. Clearly, it would fall on its nose as a predictor, however.

So ... if a good market predictor even exists it is either still undiscovered or the person that has discovered it is secretly using it to make him/herself rich. Either way, it is not going to be available on the internet for free.

Not to insult, but I don't believe in the Easter Bunny either.

YMMV, of course.

[Lsbcal]

Quote:

Originally Posted by OldShooter Absolutely true, the key words being "as long as you have." I invested a small amount of Google time in this "Meb Faber-type SMA signal" thing and was unable to find any examples where it successfully predicted future market action over a reasonable period. That didn't surprise me because, if there were a scheme that really worked, enough of the 100,000+ active managers would be using it that we would see their effect in the market. And, ultimately, it would be defeated simply by the volume of people trying to exploit it. ...

My view would be that trend following systems try to capture momentum in the markets. They are not really predictive but the better ones are trying to quickly capture momentum before other future events occur that have their own momentum affects (possibly in the opposite direction).

Momentum has been studied by academics and Wall Street types. It appears to be an acknowledged force in the markets. Can it be exploited by trading is the question.

[OldShooter]

Quote:

Originally Posted by Lsbcal My view would be that trend following systems try to capture momentum in the markets. They are not really predictive but the better ones are trying to quickly capture momentum before other future events occur that have their own momentum affects (possibly in the opposite direction).

I dunno. What I do think I know is that there are no effective predictive systems that are common knowledge. Either they, if they exist at all, are undiscovered or they are the "secret sauce" of a small enough number of investors that they have not affected markets detectably and certainly not affected markets in a way that negates the secret system.

Quote:

Originally Posted by Lsbcal Momentum has been studied by academics and Wall Street types. It appears to be an acknowledged force in the markets. Can it be exploited by trading is the question.

Yes. I think that is roughly where Eugene Fama is. There is a worthwhile 37 minute video here: INVESTORS FROM THE MOON: FAMA | Top1000Funds.com He touches on momentum about 24:15.

[Lsbcal]

Quote:

Originally Posted by OldShooter I dunno. What I do think I know is that there are no effective predictive systems that are common knowledge. Either they, if they exist at all, are undiscovered or they are the "secret sauce" of a small enough number of investors that they have not affected markets detectably and certainly not affected markets in a way that negates the secret system. ...

Yes, I would guess there a great many "secret sauce" systems out there. Some might even work but they don't have much invested in them to affect markets. Some old geezers having fun .

I don't know much about the hedge fund industry but I'd guess they have their systems and the better ones might have a variety of systems that get discarded when the system stops being effective.

[rodi]

Very interesting discussion encompassing withdrawal strategies, earmarking dividends for withdrawals, specific years of down markets... etc.

But back to the question in the subject line of this thread. I'm in bonds for diversity. Just as I have some international equities, some small cap, and a lot of large cap... I also have some bonds.... and we own a rental property that provides a stream of income.

It's just part of the general diversification of my portfolio. Helps me sleep better at night.