ERN - Early Retirement Now blog

Your favorite books and links
Post Reply
User avatar
Fish
Posts: 196
Joined: Sun Jun 12, 2016 9:09 am

ERN - Early Retirement Now blog

Post by Fish » Thu May 25, 2017 7:13 pm

https://earlyretirementnow.com

Big ERN has original perspectives on investing for FIRE. He has written a series on safe withdrawal rates which is also posted as a SSRN paper. Could be a decent addition to the ERE blogroll. We discussed a few of his posts earlier but I think this deserves its own thread.

The latest installment is about sequence of returns risk which includes the finding that SRR is the dominant parameter affecting SWR and the unexpected conclusion that exposure to SRR risk can be minimized by using low-interest debt to fund retirement savings early in the working career.

classical_Liberal
Posts: 178
Joined: Sun Mar 20, 2016 6:05 am

Re: ERN - Early Retirement Now blog

Post by classical_Liberal » Thu May 25, 2017 11:24 pm

Great to see a post from @fish, I think I have missed you :D

I only read the article linked, not the entire series. His conclusions on sequence risk are correct. The idea of “mortgage your retirement” is essentially just low interest, long-term, non-callable leverage. It’s pretty obvious to me this is a great option without needing further collaborative research.

Unfortunately, the only way to really get this type of leverage is through non market forces. Mortgages and student loan rates are artificially low because the government encourages education and home ownership. If one finds themselves in a position with this debt or if one wants an education or home, with current rates It’s a very good idea to invest aggressively rather than repay those debts, the longer the time frame the better. The only issue becomes cash flow. As long as you can make the minimum payments you’re golden. Real estate mortgage has an added value because the lender assumes a portion of the risk and none of the gains. Meaning, if your home value increases with or better than inflation, you get all gains; if it’s value gets cut in half, in most states you can just walk away and be rid of the debt, but keep your market gains. Sign me up! Except I really don’t want to be tied to a home and I see no other way to get this type of leverage.

From a different standpoint, look at tyler900’s philosophy. Instead of trying to hedge your bets with time, do it with non-correlated asset diversification to smooth out returns as best as possible. Smooth, lower returns beat high CAGR, but volatile returns for predictable WR’s any day of the week in my book. You simply have to look past the standard financial planner model of only owning a generic stock/bond split. Of course, with either strategy, you are banking on the future being reasonable similar to the past.

My personal strategy has evolved to the acceptance (actually preference) that my ERE will have some, perhaps sporadic, wage income as one of the legs to my stool. This will allow for a more variable withdrawal rate that can flux to a higher degree than the bogleheads example. This, coupled with a Tyler9000esque strategy of non-correlated passive investing should make the savings leg of my stool relatively free of sequence risk. Although it has the disadvantage (or hormesis advantage) of putting some pressure to earn more wage income in any early periods of portfolio drawdown.

oldbeyond
Posts: 136
Joined: Thu Nov 29, 2012 10:43 pm

Re: ERN - Early Retirement Now blog

Post by oldbeyond » Fri May 26, 2017 2:29 am

It's a great resource, he really has been digging deep. AFAIK academics are still stuck on the Trinity-level?

Here's my thread about his backtest: viewtopic.php?t=8635

Here's an epic read about an econ grad(yes, it confirms all of Talebs preconceptions) attempting MYR using revolving 0% interest credit card offers on the eve of the Great Financial Crisis:
https://www.bogleheads.org/forum/viewtopic.php?t=5934

ThisDinosaur
Posts: 586
Joined: Fri Jul 17, 2015 9:31 am

Re: ERN - Early Retirement Now blog

Post by ThisDinosaur » Fri May 26, 2017 8:43 am

Just read the boglehead thread by the econ major "market timer" and the ERN link on SRR. I have questions.

They both make the case that you can avoid sequence of return risk by taking out low interest debt and buying an all-equity portfolio. Then use your accumulation phase to pay of the debt. So your end-portfolio is fully invested the whole time. They assert a lump sum investment is not subject to sequence of return risk.

Here's my question; if you take out a "mortgage your retirement" loan, invest in equities, and the market drops, you can't use your income/savings to buy into the low market, because you still have the mortgage to pay.

Question the 2nd; T9K and ERN both show that a low CAGR portfolio can have a high SWR. And Tyler's site also shows you can beat Sequence of Return risk by being hedged in gold and treasuries at all times (accumulation and distribution phases equally.) That would imply you should take out a mortgage to buy debt (bonds). If your bond value rises, you can rebalance (or contribute less to bonds vs. gold &equity). This seems incorrect.

wrt that 2nd point, I have a mortgage at 2.75% interest, and VLGSX has a YTM of 2.9%. But that's subject to income tax, and mortgage interest is usually tax deductible. So, you know, wtf?

User avatar
Fish
Posts: 196
Joined: Sun Jun 12, 2016 9:09 am

Re: ERN - Early Retirement Now blog

Post by Fish » Mon May 29, 2017 12:31 pm

@CL That's a good observation that ideal leverage (low-interest, long-term, non-callable) is usually only available as a consumption loan. So the idea is to take the loan in instances where you would have paid cash, and instead funnel the cash to investments. If consumption is very low (ERE), your funds are invested anyway without the drag of paying interest on the loan. It's more applicable to those with normal to high consumption, as an option to have improved investment returns with a risk of delayed retirement if the market doesn't perform as expected.

Additionally, ERE'ers (and even MMM'ers) will benefit less because high savings rates are the driving force to FI. With a short timeframe to FIRE there's much less benefit from compounding, while MYR will still expose your leveraged investments to SRR risk as there's no "in the long run" to smooth things out. If the market tanks before the loan is repaid then retirement date gets pushed out accordingly.


@oldbeyond I see this has been done before! That was an interesting read and I really liked this quote at the end:
market timer wrote:Accumulating wealth is not something to build a life around, though. The desire to accumulate is just a debt that cannot be repaid. I now appreciate how, even with every advantage, it is easy to make a mess of your life, accumulating all different kinds of debts. When the stars align and you have your health, wealth, and freedom, you have to follow your passion.
That was a rather unfortunate experience, but it appears that this was a situation where many people were able to learn the lesson from one person's experience. Just imagine if he had timed it well and halved his time to FI, inspiring a bunch of copycats to enter with adverse timing and get hit by the black swan.


@TD Regarding your 1st question, I agree an early market drop will lead to suboptimal portfolio performance. However, MYR assumes "in the long run" and its purpose (as I understand it) is to reduce the influence of SRR on time to FI while taking advantage of expected market returns in excess of the cost of borrowing. For shorter accumulation periods introducing leverage only increases variance due to SRR.

I'm not sure I follow the 2nd question. Do the same conclusions hold when leverage is used? I'd expect there was a time in the early '80s there was a period where it would have been favorable to deleverage double digit mortgage debt instead of investing it in (US) stocks and bonds even though those subsequently performed really well.

oldbeyond
Posts: 136
Joined: Thu Nov 29, 2012 10:43 pm

Re: ERN - Early Retirement Now blog

Post by oldbeyond » Fri Jun 02, 2017 12:48 pm

As far as I understand it, the point is that you get more time in the market, and therefore get closer to the long term returns(https://portfoliocharts.com/portfolio/l ... m-returns/). Say you aim for a 4% SWR with a 60% savings rate. You take out your(magical low interest, non-callable, freely available)loan for 10x income and then spend the next ten years paying it off. We're being value agnostic in this thought experiment, but of course you might be better served by not taking out the loan in the real world at certain points in time. If you take out the loan, you get the 10-year return. If we pretend that the future can't be worse than the past, you know your worst case scenario - the worst ever 10 year return. For a lot of portfolios that number is positive, and if it's negative it'll be a lot smaller than if you do not take the loan and get the full exposure to the portfolio only after 10 years. You minimize the risks that hefty drawdowns pose in the early drawdown years by "diversifying across time".

As for taking out the loan and buying bonds, it would be nuts/redundant to do so if the bonds had the same characteristics as your loan. Say you get 4% for your rate and a duration of 10 years. If you then go ahead and buy a MYR bond with the same rate and duration composed of people with the same credit risks as you have, that'd be useless. But if you buy 30 year treasuries say, you make a duration and credit quality play. The loan and the bonds do not completely cancel out.

That is my understanding of it, feel free to tear it down if its flawed :ugeek:

In practice it's of course very hard to get credit like that. Some countries have very favorable terms for student loans(we get to borrow at 0.7x the 10-year yield on government bonds, with principal pay down over 30 years. I've borrowed and invested). You get to take out roughly 5 jacobs in our local currency, not adjusted for COL. You could get unsecured consumer debt up to roughly the same amount, but the terms are much, much less favorable of course. Then there's not paying down mortgages faster than you have to.

Of course, this is all betting on business as usual, were assets continue to deliver decent returns.

User avatar
Dragline
Posts: 4450
Joined: Wed Aug 24, 2011 1:50 am

Re: ERN - Early Retirement Now blog

Post by Dragline » Fri Jun 02, 2017 2:24 pm

These are the loans you are looking for (margin at IB currently 1.16% - 2.41%): https://www.interactivebrokers.com/en/index.php?f=1595

You can hold a much broader portfolio than "all or mostly equities" if your borrowing costs are very low, and still have a decent CAGR. And I would think holding "all equities" on margin would be a massive mistake because of the risk of margin calls. Rather, you would want a diversified portfolio that has a sufficient CAGR, but minimizes draw-downs.

But you have to have already accumulated a sufficiently large stash to take advantage of these kinds of loans. Or engage in some kind of double leveraging -- e.g., take out a home equity and deposit it in IB which can then be used to support a large margin at a lower rate.

There are methods to this madness, or just madness to these methods, depending on your perspective.

ThisDinosaur
Posts: 586
Joined: Fri Jul 17, 2015 9:31 am

Re: ERN - Early Retirement Now blog

Post by ThisDinosaur » Fri Jun 02, 2017 2:30 pm

Leveraging your retirement is to take a low interest loan and invest in a high-expected return investment. Portfolio Charts shows that a portfolio of uncorrelated assets can have a higher return than is implied by its constituent assets. So, if I have a mortgage and student loan debt, I may be better off DCAing into my high CAGR portfolio, even if it includes bonds with a lower YTM than my debt. This seems very counterintuitive to me, so I'm hoping someone can explain what I'm missing.

User avatar
Fish
Posts: 196
Joined: Sun Jun 12, 2016 9:09 am

Re: ERN - Early Retirement Now blog

Post by Fish » Tue Jun 06, 2017 2:37 pm

oldbeyond wrote:
Fri Jun 02, 2017 12:48 pm
If you take out the loan, you get the 10-year return. If we pretend that the future can't be worse than the past, you know your worst case scenario - the worst ever 10 year return.
These steps of reasoning might be really obvious but writing it out really helped me better understand the MYR concept. Thanks!
oldbeyond wrote:
Fri Jun 02, 2017 12:48 pm
As for taking out the loan and buying bonds, it would be nuts/redundant to do so if the bonds had the same characteristics as your loan. Say you get 4% for your rate and a duration of 10 years. If you then go ahead and buy a MYR bond with the same rate and duration composed of people with the same credit risks as you have, that'd be useless. But if you buy 30 year treasuries say, you make a duration and credit quality play. The loan and the bonds do not completely cancel out.
@ThisDinosaur See above. If YTM is the only source of investment return, I would agree that it does not make sense to buy bonds that yield lower than the interest rate on your loan. I think of it as making a duration/quality play in case treasuries appreciate, allowing a rebalance into equities. You're covering yourself for that possible future.

Where the mortgage and treasuries are asymmetric is that you are able to make extra principal payments on the mortgage and reduce future interest paid (at your option, and when conditions are favorable to you as the issuer of the debt) while treasuries don't carry this risk of early repayment of principal. With a mortgage, you don't carry the downside risk of being locked into paying all the interest on your now-high interest loan if rates go down. That is, assuming you are able to pay it off or refinance. Let's not forget MYR is really a consumption loan being disguised as a clever investment strategy. ;)

ThisDinosaur
Posts: 586
Joined: Fri Jul 17, 2015 9:31 am

Re: ERN - Early Retirement Now blog

Post by ThisDinosaur » Tue Jun 06, 2017 3:26 pm

@Fish and oldbeyond. I think I maybe understand what you are saying but not quite.

As I see it, the only reason I would be holding treasuries in this scenario is to prepare for the following: Interest rates fall dramatically (so the bonds I'm holding jump in price) simultaneously with stock prices plummeting. I then rebalance and get a great deal on stocks.

I can think of the following objections to holding bonds, then:
-1-It's unlikely that interest rates will fall that much during my accumulation phase or during my mortgage paying phase. So don't buy bonds.
-2-If interest rates rise, my bonds are worth less, my stocks probably drop in price. Since I still have an income, I can buy both at a better yield.

Incidentally, in a MYR scenario, if that second thing happens, you can't buy into the cheap market, because you are paying back the MYR loan with your income.

An objection I have to the MYR idea is that it *assumes* "stocks always go up" in the long run. I see that view as very subject to recency bias. Tyler's site shows that you may often need to liquidate LTTs or gold when currency fluctuations are more important than nominal market returns.

User avatar
Fish
Posts: 196
Joined: Sun Jun 12, 2016 9:09 am

Re: ERN - Early Retirement Now blog

Post by Fish » Tue Jun 06, 2017 4:56 pm

First, I would decouple the MYR concept from the investment strategies that have been used to illustrate how it works.

MYR really has 2 requirements: 1) an investment strategy that benefits from "time in the market" and 2) expected CAGR that exceeds the cost of borrowing.

It's easy to imagine that some strategies are not really compatible with MYR, such as buy-and-hold 100% long-term treasuries a la the original YMOYL. The challenge is determining how to make productive use of leverage, and whether it even makes sense in the first place. An accumulation/investment strategy may need to be modified for the use of leverage as your portfolio example and interest rate outlook demonstrates.

This is a long-winded way of saying that you've entered "exercise for the reader" territory. :D

oldbeyond
Posts: 136
Joined: Thu Nov 29, 2012 10:43 pm

Re: ERN - Early Retirement Now blog

Post by oldbeyond » Thu Jun 08, 2017 1:18 pm

Fish wrote:
Tue Jun 06, 2017 2:37 pm
These steps of reasoning might be really obvious but writing it out really helped me better understand the MYR concept. Thanks!
You're welcome! I had to meditate on all of this for a while, but I think I get it now...

On one level, this whole MYR-edifice is rather trivial, "it's expensive to be poor". SRR is interesting and easy to overlook, but the part about borrowing like 25x earnings for normal people is rather like saying "you'd have more money if you were rich". It is not feasible on a personal level, and on a systemic level it'd amount to a massive transfer of wealth to the borrowers(why would the lender not invest the money himself?). Sure, there's carrying a mortgage and investing your surplus in the stock market instead of paying it down, but it's not like that is completely riskless either, see 2008. You still have to manage the "margin call" on the house, which you of course can by carrying a small enough mortgage, but this and other "good debt" will likely only get you a fraction of your nest egg. Still, it's better than nothing of course.

Another interesting post is ERN:s analysis of long term bond data*. Especially this chart** is a nice counterweight to the recency bias that is reinforced by the more popular portfolio backtesters only having data going back to the 70's(not complaining, I understand if good data is hard to come by/certain assets not going back further than that, but it's easy to mistake the rightmost upward slope in the chart for the whole picture!). You have a worst 40-year real return of -2% for bonds or 83 years of 0 real return from 1898-1981!

I must say that I find it odd that ERN along with many of his readers seem to go for US markets only. Not that it's impossible to argue for such a play, but it seems a bit out of sync with the rest of his philosophy, especially with world CAPE at 21.7 compared with 27.5 for the US**.

* https://earlyretirementnow.com/2016/05/ ... tock-risk/
** https://earlyretirementnowdotcom.files. ... .png?w=809
*** http://www.starcapital.de/research/stockmarketvaluation

ThisDinosaur
Posts: 586
Joined: Fri Jul 17, 2015 9:31 am

Re: ERN - Early Retirement Now blog

Post by ThisDinosaur » Thu Jun 08, 2017 1:45 pm

Fish wrote:
Tue Jun 06, 2017 4:56 pm
An accumulation/investment strategy may need to be modified for the use of leverage as your portfolio example and interest rate outlook demonstrates.
I've seen it written elsewhere that using leverage is akin to reducing your bond allocation into the negative. To my mind this seems to still be a bet for a particular interest rate trajectory. Which is verboten in "passive" investment strategies. This is the sort of exercise that helps me find the holes in my understanding.
oldbeyond wrote:
Thu Jun 08, 2017 1:18 pm
Another interesting post is ERN:s analysis of long term bond data*. Especially this chart** is a nice counterweight to the recency bias that is reinforced by the more popular portfolio backtesters only having data going back to the 70's
A couple things about this. First, there is also a country bias in favor of the US S&P500 in that chart. Other countries' past (and possibly the US's future) may show bond returns better than stock returns. Second, I'm not sure if the returns on bonds in isolation matters when looking at their impact on a diversified portfolio's performance. I have pestered Tyler9000 about this question a lot all over this forum.

As for ERN on CAPE, I'm also confused why he doesn't talk more about non-US stocks. But he has shown that current US CAPE can historically support a 3% SWR for 60 years.

oldbeyond
Posts: 136
Joined: Thu Nov 29, 2012 10:43 pm

Re: ERN - Early Retirement Now blog

Post by oldbeyond » Thu Jun 08, 2017 2:55 pm

Here's the data on a country basis I've been able to locate: http://monevator.com/world-stock-markets-data/
It's a summary of this paper from Credit Suisse: https://publications.credit-suisse.com/ ... B5D14A7818

It is true that the US is an outlier in terms of real return, and that the difference between stock and bond returns are lower for some countries, but the relationship of lower bond returns holds. It also illustrates geographic risk pretty clearly! And of course it still understates it. Not all german investors in 1900 got that 3.2% return over this century...

ERN does a good job of detailing the effects of different stock/bond mixes, but of course he doesn't use any gold or commodities in his backtests(gold would be hard to backtest before the 70's I guess, and I don't know how far back there's decent data on commodities). There is of course a rebalancing effect, and of obviously an effect on the SWR if adding real assets makes the rider smoother. Still, I think something like the Desert portfolio(60% int. bonds, 30% stocks, 10% gold) would've had a hard time during a prolonged dry spell for bonds.

Post Reply