Clippings 2017

Recommended reading, economic debates, predictions and opinions.
longinvest
Veteran Contributor
Veteran Contributor
Posts: 3956
Joined: 10 Sep 2012 17:26
Location: QC

Re: Clippings 2017

Post by longinvest »

ghariton wrote: 26 Dec 2017 20:01 But why not purchase inflation protection directly, if that is what you want?
+1 :thumbsup:
Variable Percentage Withdrawal (finiki.org/wiki/VPW) | One-Fund Portfolio (VBAL in all accounts)
Park
Veteran Contributor
Veteran Contributor
Posts: 1724
Joined: 13 Aug 2010 20:51

Re: Clippings 2017

Post by Park »

longinvest wrote: 26 Dec 2017 19:30
Park wrote: 26 Dec 2017 19:26 Quote from James Montier:

"one of the great institutional constraints that we suffer is, of course, that everybody obsesses about short-term performance. If you could buy a set of stocks today and bury them for five years, you'd be laughing. But the trouble is there's very few institutions who can behave in that fashion."
Where's the promise, on my stock certificates, that their 5-year total return will always be higher than other investments? Maybe I should ask a Japanese domestic stock investor...

By the way, the appeal to authority (James Montier, in this case) doesn't make the argument true. See: https://en.wikipedia.org/wiki/Argument_from_authority
Actually, the reason I quote Montier was not in response to Longinvest's post. Longinvest quite rightly points out that appeal to authority doesn't make the argument true. But I can see where Longinvest felt it was a response to their post.

The reason for the quote was to point out that individual investors do have some advantages. At times, it can feel like institutional investors hold all the cards. As I've learned more about personal finance, I've found that that's not completely true. I divide assets into cash, fixed income and stocks. When it comes to cash, individual investors have HISA. When it comes to fixed income, they have GICs and CPP. When it comes to stocks, they can more easily own small stocks.

Montier is pointing out that time horizon can be a problem for institutional investors. I've read similar comments from Joel Greenblatt. So individual investors, who are patient enough, have another advantage over institutional investors.
Park
Veteran Contributor
Veteran Contributor
Posts: 1724
Joined: 13 Aug 2010 20:51

Re: Clippings 2017

Post by Park »

ghariton wrote: 26 Dec 2017 20:01
Park wrote: 26 Dec 2017 18:43 But in Canada, there aren't many inflation indexed bonds, and those that exist are long duration.
The issue isn't so much that RRBs are long duration. After all, the 2021 matures in less than four years. Rather, the problem is that maturities are spaced five years apart. In withdrawal stage, that means using something else, like a nominal bond, for an average of some two and a half years. That's an inflation risk I can live with.
Liquidity is an issue.
RRBs are less liquid than the corresponding nominal government bonds. But I have found that brokers have a very good inventory and will quote a price quite readily. It's true that the bid-ask spread is bigger than I like, reaching up to 3 per cent. But since I hold RRBs to maturity -- they are a way of accumulating a nest egg, not of funding day-to-day living expenses -- then I only have to pay half the spread, or 1.5 per cent. Amortized over ten years, say, that's fifteen basis points. Not too bad.
In a tax advantaged account, you'll hedge out the inflation risk. But in a taxable account, you've lost the inflation hedge.
RRBs are an effective inflation hedge in both types of account. But I wouldn't hold them in a non-registered account because income from inflation indexing is taxable as accrued, and that means that I would be prepaying some taxes. That and the administrative hassle of keeping the requisite records.
A 2.5% real return on an RRB is good. But right now, RRBs have 0.56% real return. It's a high price to pay for an inflation hedge.
Interest rates are low all around. But last week's inflation report shows a year-over-year increase of 2.1 per cent. If that is a good forecast of inflation over the next few years -- and some here have argued that it is, because it is very close to the Bank of Canada's target -- then that RRB has a nominal return of 2.66 per cent. The nominal return on a comparable nominal bond is 2.25 per cent. So I'm expecting the RRB to do better -- and I get "free" insurance against unexpected inflation, thrown in as a bonus.
Stocks are claims on real assets, and other than the short term, tend to keep up with inflation. They also have the potential to do better than inflation and are more tax friendly than bonds.
Indeed. But they are also more risky than bonds. That's why few of us hold 100 per cent equity portfolios. For me, the point of holding bonds is not to increase returns, but rather to reduce risk.
As mentioned in the last paragraph, don't expect stocks to keep up with inflation over the next 5 years. For that, I would use cash and short term fixed income. HISAs and short term GICs are my plan.
During long periods of time after World War II, Canadian short term treasuries yielded negative real returns. If unanticipated inflation were to materialize, I would expect that to happen again. Yes, liquidity can serve as partial protection against inflation. But why not purchase inflation protection directly, if that is what you want?

George
Assume tax rate of 40%. Assume inflation rate of 10%. In a tax advantaged account, your RRB has kept up with the 10% inflation. In a taxable account, your RRB has kept up with 6% inflation. Please correct me, if I'm wrong.

Your point about negative real returns is well taken. In a taxable account, they become even more negative. That's one reason why I plan to keep cash/fixed income just enough to tide me over about 5 years of a stock bear market. And here individual investors have an advantage. Your comment is about Canadian short term treasuries, and presently the 1 month bill is yielding 0.91%. Canadian Tire's HISA is yielding 1.5%, and the highest I've found is 2.3%, albeit probably a teaser rate.

About liquidity serving as a partial protection against inflation, could you got into more detail?

Vanguard has a paper on the net, which shows that cash and short term fixed income have historically been good inflation hedges. Once again, that doesn't mean that they will in the future. The Canadian government has guaranteed that RRBs in the future will hedge inflation.
User avatar
ghariton
Veteran Contributor
Veteran Contributor
Posts: 15954
Joined: 18 Feb 2005 18:59
Location: Ottawa

Re: Clippings 2017

Post by ghariton »

Park wrote: 26 Dec 2017 21:19 Assume tax rate of 40%. Assume inflation rate of 10%. In a tax advantaged account, your RRB has kept up with the 10% inflation. In a taxable account, your RRB has kept up with 6% inflation. Please correct me, if I'm wrong.
Look at it this way. On a pre-tax basis, RRBs keep up with inflation in both a taxc-advantaged and a non-registered account. In a TFSA you pay no more tax, and your after-tax return equals your pre-tax return, i.e.10.56 per cent (say 10 per cent for simplicity). In both your RRSP and your non-registered accounts, the 10 per cent return will be taxed fully, say at 40 per cent. The difference is in the timing. In the non-registered account, you will be taxed every year on the gain in that year. In the RRSP your taxes will be delayed until withdrawal. Depending on how your tax rates change over time, you will probably be better off holding the RRB in an RRSP (where I keep mine). It's not about the putative tax liability, which is the same. Rather, it is about tax deferral (which is valuable in itself).
Your point about negative real returns is well taken. In a taxable account, they become even more negative. That's one reason why I plan to keep cash/fixed income just enough to tide me over about 5 years of a stock bear market. And here individual investors have an advantage. Your comment is about Canadian short term treasuries, and presently the 1 month bill is yielding 0.91%. Canadian Tire's HISA is yielding 1.5%, and the highest I've found is 2.3%, albeit probably a teaser rate.

Yes, if the comparison was to longer-term treasuries, the real return wouldn't be as negative and would be positive in many of those years. But if the purpose is to build a hedge against unexpected inflation, you want very short term instruments, not longer term ones. The reason is that you are expecting that, as inflation takes hold, interest rates will rise to compensate. But to take advantage of thwese new higher rates, you must be able to roll over your fixed income instruments without a loss. This is possible for very short term instruments, not for longer term ones.
About liquidity serving as a partial protection against inflation, could you got into more detail?
See above.

The theory, going back to at least Irving Fisher, is that nominal interest rates are (approximatdly) the sum of real interest rates and the rate of inflation. As inflation goes up, so do nominal interest rates. But to be able to tzke advantage of that, you must be liquid, i.e. holding instruments that you can turn to cash without any (big) loss.
Vanguard has a paper on the net, which shows that cash and short term fixed income have historically been good inflation hedges. Once again, that doesn't mean that they will in the future. The Canadian government has guaranteed that RRBs in the future will hedge inflation.
I haven't got my copy of The Triumph of the Optimists to hand, but if you do, I think that you will find the negative real rates for Canada shown there.

George
The juice is worth the squeeze
Park
Veteran Contributor
Veteran Contributor
Posts: 1724
Joined: 13 Aug 2010 20:51

Re: Clippings 2017

Post by Park »

Assume tax rate of 50%. I put $10K in an RRSP. Of that $10K, $5K in my money and $5K is the CRA's that they didn't tax. Assume that $10K grows to $100K. I withdraw the money. Assume tax rate hasn't changed at 50%. $50K is my money. $50K is the CRA's. So what the RRSP has done is allow tax free growth of my money. So if I had invested the $10K in RRBs, my $5K would have grown tax free. I

What complicates the situation is that the tax rates at the time of contribution and withdrawal may not be the same. But the basic principle still holds.
longinvest
Veteran Contributor
Veteran Contributor
Posts: 3956
Joined: 10 Sep 2012 17:26
Location: QC

Re: Clippings 2017

Post by longinvest »

Park wrote: 26 Dec 2017 22:50 Assume tax rate of 50%. I put $10K in an RRSP. Of that $10K, $5K in my money and $5K is the CRA's that they didn't tax. Assume that $10K grows to $100K. I withdraw the money. Assume tax rate hasn't changed at 50%. $50K is my money. $50K is the CRA's. So what the RRSP has done is allow tax free growth of my money. So if I had invested the $10K in RRBs, my $5K would have grown tax free. I

What complicates the situation is that the tax rates at the time of contribution and withdrawal may not be the same. But the basic principle still holds.
This time we agree. :)
Variable Percentage Withdrawal (finiki.org/wiki/VPW) | One-Fund Portfolio (VBAL in all accounts)
longinvest
Veteran Contributor
Veteran Contributor
Posts: 3956
Joined: 10 Sep 2012 17:26
Location: QC

Re: Clippings 2017

Post by longinvest »

ghariton wrote: 26 Dec 2017 22:39 But if the purpose is to build a hedge against unexpected inflation, you want very short term instruments, not longer term ones. The reason is that you are expecting that, as inflation takes hold, interest rates will rise to compensate. But to take advantage of thwese new higher rates, you must be able to roll over your fixed income instruments without a loss. This is possible for very short term instruments, not for longer term ones.
I disagree. Here's part of a post I wrote on the Bogleheads forums to deconstruct the "nominal bonds are vulnerable to inflation" investment myth:

https://www.bogleheads.org/forum/viewto ... 1#p3648951
longinvest wrote: Let me take the opportunity to attack a big modern investing myth, that "nominal bonds are extremely vulnerable to inflation". Nominal bonds are no more vulnerable to inflation than other assets which are not indexed to inflation, such as stocks. The explanation is actually pretty simple.

Let's forget about bonds, for a small moment, and try to understand how cash in a savings account behaves in relation to inflation (in a tax-advantaged account, so that taxes don't erode value). How would an investor fight inflation with cash? There are two main approaches.
1- A speculative approach, which is based on trying to find a savings account which has a higher rate than whatever inflation will happen to be during the time the money will sit in the savings account.
2- A trailing approach, which is based on putting the money into a savings account which has a posted rate as least equal to the last reported 12-month inflation rate. If one keeps doing that, one will be certain that his money will keep pace with inflation with a one year lag.

The speculative approach is anxiety generating, because one has no control. So, we reject it and forget about it.

The trailing approach, on the other hand, keeps the investor mostly in control. If the interest rate on the savings account drops, the investor can simply move his money to a different institution which offers a savings account with a higher interest rate. If trailing inflation moves up, one has to make sure that his money is moved or left into a savings account with an interest rate that matches it or beats it.

The only cause for the money not to keep pace with inflation would be the inability of the investor to find a savings account with a rate that matches trailing inflation. But, this is not something that happens in the "future", it is something that happens in the "present". The investor knows about it when it happens. The investor has the opportunity to sidestep this problem by moving his money out of savings accounts and seeking another type of investment security, such as a 1-year CD or something else, which promises a high-enough rate to match trailing inflation. If the investor fails to find such an alternative investment, or if he is too lazy to search for one, he can decide to simply let his money lose ground to inflation while knowing the exact loss rate, which is the difference between trailing inflation and his savings account rate.

In other words, if inflation was 3% over the last 12 months and the savings account interest rate is 2%, the money will be losing ground to inflation at the rate of ((1.02/1.03)-1) = approximately -1% as long as the situation persists. It all happens in the "present", except that we're evaluating the "keep pace with inflation" on a 1-year lag basis.

It's now easy to project the same thinking into bonds. It would be relatively easy for a 10-year nominal bond ladder investor to make sure his money keeps pace with inflation (with a lag) using the trailing approach. The bond ladder investor only has to make sure that the yield on the new 10-year rung he buys every year matches or beats trailing 1-year inflation. The marked-to-market value fluctuates and, as a result, total return from year to year fluctuates, sometimes up, sometimes down. But, in the long term, total return will keep pace with inflation (with a lag) as long as the "reinvest at trailing 1-year inflation rate or better" rule is respected.

In the history of U.S. nominal bond returns, there was a very particular period of time when bond yields fell significantly below trailing 1-year inflation for a number of years. This happened in the 1940s and early 1950s. Note that it wasn't the result of free bond market pricing; on the contrary, it was the result of a deliberate government action, as I've learned from Barry Barnitz:
Barry Barnitz wrote: Please note that interest rates in the 1940's were subject to price controls and interest rate pegging. See --->

Interest Rate Controls: The United States in the 1940s on JSTOR
Before the Accord: US Monetary - Fiscal Policy 1945 -1951
In a period of 20% inflation, all yield (short to long) were kept artificially low, near 2%. A nominal bond investor, at the time, could have sold back his bonds to the government, at their artificially inflated prices the government was willing to pay*, and move his money elsewhere. The bond investor was in control! Of course, the investor didn't necessarily know that stocks would go up like they did (especially after the two consecutive bear markets of the early and late 1930s), but he had a choice not to leave his money into bonds (or cash) with significantly lower yields than trailing 1-year inflation.

* One of the ways the government kept yields low was by buying back bonds at inflated prices, relative to what the market would have paid!

When looking at historical bond return charts, people who don't understand this think that bonds went into a 40-year bear market. But, actually, what happened is that bonds lost something like 35% relative to inflation in the 1940s-early-1950s, due to interest rate pegging, and that was it. People don't realize that in the 1970s, bonds did pretty well, overall, with less volatility than stocks, keeping pace with inflation (with a lag). Their yields mostly kept going up with trailing inflation, thanks to free markets (investors not willing to accept less), and bonds behaved like bonds. In other words, value went down so that yields immediately took effect.

Here's a historical inflation-adjusted bond total-return chart for 1940 to 1985:
Image
(Source data for constructing the chart: VPW backtesting spreadsheet)

We see the loss of purchase power in the 1940s, due to yields significantly trailing inflation. We also see a few years dip, in the late 1970s, when inflation was raging up and yields were following up. As expected, it took a few years lag for the impact of higher yields to be fully reflected into total returns; nothing surprising, here. What's actually surprising is how yields kept a big spread over trailing inflation afterwards, explaining a gain equivalent to the losses of the 1940s in a few years, in the early 1980s. My point is that both events are completely unrelated. There was a government-imposed real loss in the 1940s. In the 1980s, bond investors became very demanding for higher yields.

In the current situation, nominal bonds and even high-interest savings accounts seem to be yielding at least as much as trailing inflation. So, as an investor, I'm not worried. Of course, I do diversify my investments (domestic and international stocks, nominal and inflation-indexed bonds, and high-interest savings account for money needed in the short term). I invest into stocks with the hope of higher returns, but I am fully aware that no stock certificate contains a promise of future returns that will beat inflation. As for bond investments, I am also fully aware that their total return will be mostly determined by future bond yields (or, equivalently, future bond prices and coupons).
Last edited by longinvest on 27 Dec 2017 09:24, edited 1 time in total.
Variable Percentage Withdrawal (finiki.org/wiki/VPW) | One-Fund Portfolio (VBAL in all accounts)
Park
Veteran Contributor
Veteran Contributor
Posts: 1724
Joined: 13 Aug 2010 20:51

Re: Clippings 2017

Post by Park »

http://theirrelevantinvestor.com/2017/0 ... ther-side/

"As Patrick O’Shaughnessy has shown: “If you isolate all 12-month periods where the market is down, the worst 20% of stocks in the S&P 500 account for 81.2% of the markets total loss.”"

This is an explanation for why equally weighted fund underperform during bear markets. From the peak to nadir of a bear market, equally weighted funds keep buying the worst 20% of stocks, whereas market cap weighted funds don't.
Park
Veteran Contributor
Veteran Contributor
Posts: 1724
Joined: 13 Aug 2010 20:51

Re: Clippings 2017

Post by Park »

longinvest wrote: 26 Dec 2017 19:30
Park wrote: 26 Dec 2017 19:26 Quote from James Montier:

"one of the great institutional constraints that we suffer is, of course, that everybody obsesses about short-term performance. If you could buy a set of stocks today and bury them for five years, you'd be laughing. But the trouble is there's very few institutions who can behave in that fashion."
Where's the promise, on my stock certificates, that their 5-year total return will always be higher than other investments? Maybe I should ask a Japanese domestic stock investor...
Since the end of 1989, I believe that the Japanese stock market is still down, after taking into account inflation and dividends. And the US market had 0% return from 1966-1982 and 1929-1945. For the Japanese stock market since 1989 and the American market from 1966-1982, I've seen evidence that value investing improved return. I don't think that's true for the American market from 1929-1945, but there was significant deflation from 1929-1932, and value stocks tends to be more levered: leverage and deflation don't mix well.

Is this relevant to today?

https://www.quandl.com/data/MULTPL/SHIL ... o-by-Month

The above link states that the present CAPE of the US market is 31.29. The 1966 peak was 24.6. The 1929 peak was 32.54. That's why I think about value investing as an alternative to market cap indexing.

One saving grace is that in about 2 years, the earnings decline associated with the 2007-2009 bear market will be gone, and that will tend to decrease CAPE.
longinvest
Veteran Contributor
Veteran Contributor
Posts: 3956
Joined: 10 Sep 2012 17:26
Location: QC

Re: Clippings 2017

Post by longinvest »

I don't believe that simplistic measures (YTM for bonds, CAPE for stocks) can predict future long-term returns. Actually, here's a mathematical proof that they just can't:

The Futility of Predicting Future Returns - Bogleheads.org
longinvest wrote: Here is a proof that the future returns of broad bond and stock funds cannot be predicted with any kind of accuracy. In other words, any "good" prediction (because an appropriate accompanying range is provided) will be mostly useless to predict the highest returning asset.


The uncertain future return of a broad bond fund

Some people predict the future return of a bond fund using its SEC Yield or average yield to maturity (YTM).

Using a fund's YTM as a prediction assumes that the fund will hold all bonds until maturity and distribute back the principal of maturing bonds to investors, in addition to coupons, until the fund's duration becomes 0 (all bonds have matured) and the fund is closed. This is not how a bond fund works!

A bond fund keeps a relatively constant duration. In simpler words, it reinvests the principal of maturing (or sold) bonds. To make an accurate prediction, one must predict the yield at which principal will be reinvested (unknowable in advance). Note that when the yield curve is steep, reinvestment yield is always higher than the fund's YTM. For all those investors worried about higher yields: higher yields imply a higher reinvestment yield, which eventually translates into a higher total return.

Accurately predicting the exact future return of such a complex bond fund as a Total Bond Market fund over any future period of time (1 day to 60 years) is simply impossible. In the short term (1 day), a change in the yield curve (unknowable in advance) dominates total return. In the long term (2 X duration and more), future reinvestment yields (unknowable in advance) dominate total return. In between, both factors impact total return.

Because bonds are contracts to pay precise amounts of money on specific dates (coupons, principal), assumptions about reasonable limits on yield movement allows the use of mathematics to compute a likely range of possible future total returns over a specific time period. But, this range is far from tight.

Now, don't interpret the above to mean that bond returns are completely unpredictable, that there is a possibility of a Bondageddon. Mathematics allow us to easily realize that the price volatility of an aggregate bond fund is limited, because the fund contains a significant amount of short and intermediate-term bonds. In other words, bonds are not stocks!

In summary, any precise broad bond fund return prediction (e.g. "expect SEC Yield") without an accompanying range or with a tight range is misleading.


The uncertain future return of a broad stock fund

Some people predict the future return of a stock fund using its price-earnings ratio (P/E) or a dividend discount model (DDM).

Unlike a bond, a stock is not a contract to deliver coupon payments on specific dates and pay back the principal at maturity. A stock is a title of partial ownership in a company entitling the investor to future dividend payments. There is no provision for paying back the invested principal nor any contractual precision about the amounts and dates of future dividend payments. (Most common stocks provide a voting right which gives the investor an indirect influence on the management of the company).

As a stock provides no guarantee of precise future payments, human psychology has a significant impact on the market price of the stock. If investors believe that a company has solid assets (financial, human), competent management, and a good prospect of growth and future profits, they will accept to trade its stocks at a higher price than if they expect future losses and no growth.

Any metric, such P/E or DDM, cannot account for the infinite range of possible scenarios that can unfold in real-life. For example: How do you use P/E or DDM to predict the future returns of Enron a few years before its bankruptcy? Or to predict the future returns of Apple on its initial public offering (IPO)?

A stock fund is just a collection of stocks. One could make an argument that the law of averages will make P/E or DDM somewhat more precise on a large collection of stocks. Yet, it is simply impossible for any stock metric to provide any kind of good precision over any time period (1 day to 60 years).

In summary, any precise or moderately precise stock return prediction (e.g. "future stock returns will be lower than in the past century") without a very wide accompanying range is misleading.
Variable Percentage Withdrawal (finiki.org/wiki/VPW) | One-Fund Portfolio (VBAL in all accounts)
User avatar
ghariton
Veteran Contributor
Veteran Contributor
Posts: 15954
Joined: 18 Feb 2005 18:59
Location: Ottawa

Re: Clippings 2017

Post by ghariton »

longinvest wrote: 26 Dec 2017 23:40 I disagree. Here's part of a post I wrote on the Bogleheads forums to deconstruct the "nominal bonds are vulnerable to inflation" investment myth
Interesting argument. Let me restate it in my own words, to see if I have understood it correctly.

First, cash (actually, a savings account). Your scheme is to use a savings account whose interest rate incorporates the actual inflation rate for the period just past. (The interest ate will include both inflation and real return, but let's put the latter at zero for simplicity.) So your return will always lag inflation by a period (let's say a year for convenience, although it could be much shorter). In a year when inflation increases, you get a loss equal to that inflation increase for that year. Then you track the new inflation rate until the year in which inflation declines again. At that point you make a gain equal to the inflation decrease. Over time, you should even out.

One nitpick. This assumes that you can find an issuer (of a savings account) whose return always tracks past inflation.When inflation is increasing, this shouldn't be a problem. But when inflation is decreasing, it may be, at least in a timely way. Since your savings account is liquid, you can always look at alternatives. But (a) this involves transaction costs of searching (b) it introduces an element of risk that you will not succeed.

Still, all in all, I agree with your cash analysis. I note that it works well, in my opinion, because you have a very liquid instrument.

Then you apply the analysis to a ten-year bond or GIC ladder. I agree that the analysis carries over for the first rung, i.e. the one with a one-year maturity. But what about the other rungs? As the extreme case, take the last rung, the one with a ten year maturity. Say inflation doubles from 5 to 10 per cent. You cannot roll over than rung for another nine years. When inflation increases from 5 to 10 per cent, you are stuck with receiving 5 per cent for the next nine years. Then you roll over to a 10 per cent bond, and enjoy that for 10 years, even though inflation may then drop back to 5 per cent. But now the time value of money matters. You have to wait nine years to "get your money back", and that will hurt.

The problem IMHO is that most rungs of a ten-year ladder are just not that liquid. That lack of liquidity will impose a penalty, which you will not be able to recover for a long time.

Where did I go wrong?

George
The juice is worth the squeeze
User avatar
ghariton
Veteran Contributor
Veteran Contributor
Posts: 15954
Joined: 18 Feb 2005 18:59
Location: Ottawa

Re: Clippings 2017

Post by ghariton »

Park wrote: 27 Dec 2017 08:18 "As Patrick O’Shaughnessy has shown: .
Mark Hulbert makes essentially the same argument at Market Watch.
So there is nothing magical about equal-weighted funds. An equal-weighted version of an index will outperform the cap-weighted version when small caps are beating large caps, or when contrarian approaches are beating momentum, or both. It will underperform when the opposite of those two conditions prevails.

One important additional factor to keep in mind about equal weighted index funds: expenses. These funds have greater transaction costs because of the frequent rebalancing needed to bring each stock’s allocation back to equal weighting. This involves selling portions of their outperforming stocks and buying more of their underperformers. No rebalancing transactions are required by a cap-weighted fund.

Transaction costs involve more than just brokerage commissions, Quantal’s Tint noted. Since equal-weighted index funds will be trying to buy and sell the same stocks at more or less the same time, they can be expected to receive more competition and higher market-trading impact on their trades than would otherwise be the case.
George
The juice is worth the squeeze
longinvest
Veteran Contributor
Veteran Contributor
Posts: 3956
Joined: 10 Sep 2012 17:26
Location: QC

Re: Clippings 2017

Post by longinvest »

ghariton wrote: 27 Dec 2017 12:37
longinvest wrote: 26 Dec 2017 23:40 I disagree. Here's part of a post I wrote on the Bogleheads forums to deconstruct the "nominal bonds are vulnerable to inflation" investment myth
Interesting argument. Let me restate it in my own words, to see if I have understood it correctly.

First, cash (actually, a savings account). Your scheme is to use a savings account whose interest rate incorporates the actual inflation rate for the period just past. (The interest ate will include both inflation and real return, but let's put the latter at zero for simplicity.) So your return will always lag inflation by a period (let's say a year for convenience, although it could be much shorter). In a year when inflation increases, you get a loss equal to that inflation increase for that year. Then you track the new inflation rate until the year in which inflation declines again. At that point you make a gain equal to the inflation decrease. Over time, you should even out.
Exactly.
ghariton wrote: 27 Dec 2017 12:37 One nitpick. This assumes that you can find an issuer (of a savings account) whose return always tracks past inflation.When inflation is increasing, this shouldn't be a problem. But when inflation is decreasing, it may be, at least in a timely way. Since your savings account is liquid, you can always look at alternatives. But (a) this involves transaction costs of searching (b) it introduces an element of risk that you will not succeed.
I mentioned this difficulty, in my post. In other words, there's always a risk that the investor would be unable to find a savings account with a high-enough interest rate to match trailing inflation. But, when this happens, the investor is aware of it. He can choose to try his luck at seemingly higher-yielding investments. Or, if the difference between the account's rate and trailing inflation is small enough, he can just accept a small loss of purchase power (might be balanced, if he's lucky, by the other years when he got a higher rate than trailing inflation).
ghariton wrote: 27 Dec 2017 12:37 Still, all in all, I agree with your cash analysis. I note that it works well, in my opinion, because you have a very liquid instrument.
It actually works well because of mathematics.

Given a sequence of inflation rates, i_prev*, i_1, i_2, i_3, ..., i_n, representing inflation during the year that preceded the investment, and during each of the n years of investment; given a sequence of interest rates, r_1, r_2, r_3, ..., r_n, representing the interest rate of a high-interest savings account during each of the n years of the investment, where r_(x) >= i-(x-1); we can calculate a lower bound on the cumulative total real return as follows:

* i_prev would better be described as i_0, but I find i_prev more intuitive.
  • To get a lower bound, we assume that r_(x) = i_(x-1), in other words, that r_1 = i_prev, r_2 = i_1, ..., r_n = i_(n-1)
  • The total cumulative real return is: (1 + r_1)/(1 + i_1) X (1 + r_2)/(1 + i_2) X (1 + r_3)/(1 + i_3) X ... X (1 + r_n)/(1 + i_n) - 1
  • This can be rewritten as: (1 + r_1)/1 X (1 + r_2)/(1 + i_1) X (1 + r_3)/1 + i_2) X ... X (1 + r_n)/(1 + i_(n-1)) X 1/(1 + i_n) - 1, thanks to the commutativity of multiplication and division.
  • But, as r_2 = i_1, r_3 = i_2, and so on, the above is equal to: (1 + r_1) X 1 X 1 X ... X 1 X 1/(1 + i_n) - 1 = (1 + r_1)/(1 + i_n) - 1
What this tells us is that if we succeed at always finding a high-interest savings account with a rate matching exactly trailing 1-year inflation, the cumulative total real return on our investment will be determined by the ratio of the first year's growth over the last year's inflation minus one. The annualized real return is the nth root of this ratio minus one. The ratio converges to 1 as n grows to infinite. In other words, the real return converges towards 0% the longer the investment is held.

In non-mathematical words, the longer one holds the investment and succeeds at matching trailing inflation, the smaller the impact of the last year's inflation on annualized total returns.

This is not perfect. But, let's not forget that our principal worry as investors is the insidious cumulative impact of inflation on the purchase power of our investments. If we can restrict this cumulative loss, over a lifetime, to a single year's inflation rate, then we've pretty much achieved our goal of preserving most of the purchase power. In other words, we've got a good enough solution.

This sets the stage to address your worries about bond funds (represented by a 10-year bond ladder below).
ghariton wrote: 27 Dec 2017 12:37 Then you apply the analysis to a ten-year bond or GIC ladder. I agree that the analysis carries over for the first rung, i.e. the one with a one-year maturity. But what about the other rungs? As the extreme case, take the last rung, the one with a ten year maturity. Say inflation doubles from 5 to 10 per cent. You cannot roll over than rung for another nine years. When inflation increases from 5 to 10 per cent, you are stuck with receiving 5 per cent for the next nine years. Then you roll over to a 10 per cent bond, and enjoy that for 10 years, even though inflation may then drop back to 5 per cent. But now the time value of money matters. You have to wait nine years to "get your money back", and that will hurt.

The problem IMHO is that most rungs of a ten-year ladder are just not that liquid. That lack of liquidity will impose a penalty, which you will not be able to recover for a long time.
Assumptions

I'll use a mathematical argument that sidesteps liquidity issues (or that simply considers them as part of market pricing).

To keep the model simple, we'll approximate a bond fund using a 10-year bond ladder where, each year, all coupons (paid by all rungs) and the principal of the maturing rung are reinvested into a new 10-year bond. We'll assume that we have non-callable and non-defaulting bonds which pay a single annual coupon. We'll also assume that bonds are bought at par and never sold (except for initiating the ladder and liquidating it).

Preliminary notions

Let's pick a single 10-year bond with an x% coupon bought at par. If we hold this bond until maturity and don't reinvest its coupons, we know that its internal annualized rate of return will be x%. Note that this isn't the total return which assumes that coupons are reinvested at whatever the residual yield-to-maturity (YTM) happens to be at the beginning of each of the 2nd, 3rd, ..., 10th year. In other words, the total return depends on the prices of the bond at the beginning of each of those years. The smaller the coupon, the less total return can differ from the initial YTM of the bond. Often, investors use the initial YTM of the bond (which is equal to the coupon, in our case, as we bought the bond at par) to approximate its effective future total return, knowing that actual total return might differ but not too much for a single non-callable non-defaulting bond held to maturity. For a single bond, YTM is a good enough estimate of its possible future annualized total return.

So, let's assume, for simplicity, that this 10-year bond will effectively deliver an annualized x% return over its lifetime. We know that it won't deliver it as a smooth x% annual return each of the 10 years; that annual returns will fluctuate. Given a steep (where long-term yields are higher than shorter-term yields) and fixed yield curve (where the yield curve does not change over time and all rates stay fixed), returns will be higher in the beginning and get lower in subsequent years, as the yield for the residual maturity of the bond gets lower, such that the cumulative growth over 10 years is (1 + x%)^10. In other words, (1 + r_1) X (1 + r_2) X ... X (1 + r_10) = (1 + x%)^10 where r_1, r_2, ..., r_n are the total returns of year 1, year 2, ..., year 10. Obviously, given the static yield curve (utopic) assumptions, it must be that r_1 > x% and r_10 < x%.

In real life, the yield curve isn't static. Yields move around. Actually, it would be more accurate to talk about bond prices as the "yield curve" is just an approximation which fails to fully explaining bond prices! For curious readers, I've explained this on the Bogleheads: What determines the shape of the yield curve?

The bond ladder model

Our 10-year bond ladder never buys anything but 10-year bonds (except when initiated and liquidated). If the yield curve was steep and fixed, the return of each rung would fluctuate, but we know that the ladder's main driver of return would be its x% coupons. Remember that any reinvested coupons also generates x% coupons. So, total return has to be close to x% over a long period of time; there's no escaping it.

<OFF TOPIC>
This often surprises new bond investors who think that the weighted average YTM of the ladder would be a better approximation of the ladder's long-term return. This is not so. Why? Because when the ladder is first bought, its average YTM is the weighted average of the YTMs of each of the 10 rungs, which happen to be x%, and 9 additional smaller rates (given our steep yield curve assumption). Worse, because it's a weighted average, shorter-term rungs which have inflated prices (e.g. higher than par) due to their lower yields weight more, making the weighted average YTM even lower than a simple average and farther from the x% long-term return estimate.
</OFF TOPIC>

If we forget about the initiation and liquidation of the ladder, for a moment, but allow for a fluctuating yield curve, we realize that the total return is only affected by reinvestment rates.

Let's pick a random year. That year, we use the maturing rung's matured bond capital and combine it with the coupons paid by the 10 rungs to buy a new bond which goes on delivering 10 y% coupons during the next 10 years and paying its principal back at maturity. The next year, we buy a new bond which goes on delivering 10 z% coupons... And, so on.

One can see that part of the ladder delivers x% for 10 years. Another part delivers z% for 10 years, and so on. Other than for the initiation and the liquidation of the ladder, total return is only determined by the coupons. A good enough approximation is to say that the annualized total return of the ladder is equal to the average of 10-year coupons rates over the holding period.

But, what about initiation and liquidation? At initiation, we buy 9-year, 8-year, ..., 1-year bonds (which were initially 10-year bonds but which have aged). In a steep and fixed yield curve environment, they would cost more than par and have lower yields than the fixed 10-year yield. But, at liquidation, a similar situation would happen: we would sell 9-year, 8-year, ..., 1-year bonds (which were initially 10-year bonds but which have aged). In a steep and fixed yield curve environment, they would have a higher value than par due to their lower yields than the fixed 10-year yield.

I'll now use the concepts introduced in the "Preliminary notions" section. Even if the initially bought 9-year bond goes on to deliver (1 + r_2) X ... X (1 + r_10), the missing (1 + r_1) growth to get the (1 + x%)^10 cumulative growth is balanced by the 10-year bond that was bought one year prior to liquidating the ladder. That 10-year bond delivers our missing (1 + r_1) growth. The same goes with the other rungs.

If we drop the fixed and steep assumption, and adopt a real-life varying yield curve, we see that the annualized total return could differ from the average coupon by the nth root of the gain or loss factor, minus one, due to differences in shape between the initial and liquidation yield curves. The longer the time frame, the bigger n, the closer the nth root to one, and the closer the difference in annualized returns to zero.

All I'm trying to say, here, is that over n years where coupons are c_1, c_2, ..., c_n, the annualized total return of the ladder has to be close enough to the average of c_1, c_2, ..., c_n.

It should be easy enough, now, to see that if c_1 >= i_0, c_2 >= i_1, ..., c_n >= i_(n-1), where c_(x) are the 10-year coupons and i_(x) are the annual inflation rates, that the bond ladder should track inflation well enough.

Of course, the ladder's market value is subject to interest rate risk. It fluctuates. It's not a cash investment, so the investor isn't immune to short-term losses (the magnitude of which can be approximated by the average duration). But, as long as one pays attention to reinvestment rates, future inflation shouldn't be a worry for a nominal total-market bond fund investor. If the 1940s situation was to happen again with yields significantly lower than trailing inflation, it would mean that bonds are overvalued (low yields equal high prices, by definition!) and it would be a good time to sell the bond fund to buy something else that yields more. That's the nice thing: prices would have to be high, otherwise yields would be higher!

It's counter intuitive, but a rational nominal bond fund investor wishes for steeper short-term losses when inflation goes up. In the 1940s, in the US (and I suspect in Canada), the government artificially kept bond prices high, insulating bond investors from short-term losses but exposing those who didn't seize the sell opportunity to a deep long-term loss of purchase power.
Variable Percentage Withdrawal (finiki.org/wiki/VPW) | One-Fund Portfolio (VBAL in all accounts)
User avatar
ghariton
Veteran Contributor
Veteran Contributor
Posts: 15954
Joined: 18 Feb 2005 18:59
Location: Ottawa

Re: Clippings 2017

Post by ghariton »

Okay, I think I see what you've done (and a nice job, too.) As you say, your analysis side-steps questions of liquidity.

But, unless my understanding is incomplete (which is entirely possible), I'm still left with two issues.

The first can be illustrated with your "cash" example. At the end. the return is equal to [1 + c(1)] / [1 + i(n)] -1. While intermediate coupons are assumed to be equal to or greater than the previous year's inflation, the first year's coupon isn't anchored that way. Indeed, in a period of rising inflation (which is what ny original point is all about), c(1) will be smaller, and likely much smaller, than i(n). Subtract the 1, and you get a negative return, which makes intuitive sense to me. It's true that, as n increases, taking the n-th root means a smaller and smaller effect. But as n increases, the chances of a bigger and bigger difference between c(1) and i(n) also increase.

Carrying this over to the bond ladder, I agree with your calculation, except for your treatment of acquisition and disposition of the ladder. If inflation has been at all significant, the proceeds of the disposition of, say, the 10-year bond are likely to be smaller, and likely much smaller, than the initial acquisition of a 10-year bond. Granted, as the holding period of the ladder increases, the loss will be spread out over more and more years, and eventually the loss may not be so bad. Indeed, if the inflation increase is temporary, the problem goes away. But if we move to a "new" world of persistent higher inflation, there will still be some impact.

But the above critique, while interesting to me, isn't fatal to your argument. I find my second issue more concerning. This is your assumption that the investor will always be able to find a new bond with coupon equal to or greater than the previous year's inflation. To me, this assumption is far from obvious.

I don't have a study on this point, but anecdotally, when inflation took off after World War II and again in the 1970s, interest rates (or coupons) lagged, and lagged badly, for a while. Whether that was due to inertia in the markets (plausible) or whether there was financial repression by central banks (alleged by some writers), the coupons I could get on new bonds lagged inflation by several percentage points.

Again anecdotally, investors eventually got wise to accelerating inflation and began demanding instruments that more than covered inflation. Indeed, in my opinion they over-reacted. So yes, rolling over investments could work to recover your original principal (and a fair return). The over-recovery in later years made up for the under-recovery in earlier years. Note, however, that this worked only if you could roll over without too much loss of principal. That would have required instruments that matured in the (very) short term.

P.S. when you say you assumed a fixed yield curve in your analysis, I wasn't clear whether you meant both shape and absolute level, or just shape. If the former, that would explain why we are coming up with different conclusions.

Anyway, I want to think about your analysis some more. It does illustrate, very well, some other aspects of bonds.

George
The juice is worth the squeeze
longinvest
Veteran Contributor
Veteran Contributor
Posts: 3956
Joined: 10 Sep 2012 17:26
Location: QC

Re: Clippings 2017

Post by longinvest »

I'm not trying to say that nominal bond funds are as safe as RRBs in face of inflation. But, it's nice to understand their dynamics and to know that when yields don't adjust to higher inflation it means that bond prices are higher than they should be, leading to smaller losses when liquidating them at such higher prices to buy other investments with better prospects.

Yes, there's a possibility of a one-year inflation haircut, but that's far from the risk of accumulating of 10 years of inflation haircuts, as is often portrayed as a possibility when someone considers including a 10-year nominal bond in his portfolio.

The combination of 50% 7.6-year-duration total-market nominal bonds with 50% 15.4-year-duration total-market RRBs in the bond portion of my portfolio is an interesting one. Nominal bonds are vulnerable to a one-year inflation haircut. RRBs are indexed to the CPI and are thus immune to inflation, but they are quite sensitive to real-yield changes. So, all in all, RRBs can still provide me with an equivalent haircut due to a real-yield change, instead of a sudden surge of inflation. Different investments; different risks.

I discovered how bond funds work after developing a bond fund simulator (as a macro-less spreadsheet).

I don't know about the history of Canadian nominal bonds. But, when studying historical nominal US bond returns, I was intrigued to discover that yields remained low during the high-inflation 1940s era. Here's what I wrote about it:

SBBI Historical Bond Returns: are they inadequate? - Bogleheads.org
On Fri Aug 26, 2016 9:46 am, longinvest wrote: [...]

When I look at the entire 1871-2015 history of synthetic intermediate bond returns (using our simulator's Bond Fund 10-2), I only see a single period when intermediate bonds suffered a significant lag relative to inflation and didn't recover quickly: 1941 to 1948.

I don't know why bond investors didn't ask for yields that matched inflation (10-year government bond yields varied between 2% and 2.5%). Inflation was brutal over that period, specially in 1941-1942, and 1946-1947:

Code: Select all

1941	11.35%
1942	7.64%
1943	2.96%
1944	2.30%
1945	2.25%
1946	18.13%
1947	10.23%
I suspect that a wider look at the economic and political landscape (e.g. war) and a study of how inflation was calculated could reveal some hidden truths.

[...]
A few hours late, Barry Barnitz (the Bogleheads wiki admin and Bogleheads blog admin) provided the following links:

https://www.bogleheads.org/forum/viewto ... 7#p3030867
On Fri Aug 26, 2016 2:52 pm, Barry Barnitz wrote: Hi:
Please note that interest rates in the 1940's were subject to price controls and interest rate pegging. See --->

Interest Rate Controls: The United States in the 1940s on JSTOR
Before the Accord: US Monetary - Fiscal Policy 1945 -1951

regards,
So, it was no accident that investors didn't require higher yields; the government actually imposed controls on interest rates. Reading the above documents (just a quick reading) allowed me to discover (if I remember correctly) that the government was actually willing to buy back bonds at these low yields!

In the entire history of US rates since 1870, as reported by Prof. Shiller, and later in FRED data, I have not seen any other extended period of time when bond investors didn't ask for yields commensurate with inflation.

Requesting yields commensurate with inflation seems pretty intuitive to me. If Canadian inflation was to go up to 5% per year, we wouldn't expect an investor to be able to sell one of his bonds with a 2% YTM, right? We would expect bond buyers to wait for the price to drop enough so that the YTM reaches at least 5%. Wouldn't that be logical? That's how market pricing usually happens.

If this pricing doesn't happen, it means that either the bond market has completely frozen; YTM spreads have widened to 3% (e.g. ask YTM 2%, bid YTM 5%), and nobody crosses the line. No new money is invested into bonds and bond holders don't reduce their ask price to sell their bonds; instead, they hold all of their bonds until maturity. (Yeah, sure! Nobody ever needs money fast! That's something I would bet will never happen during my lifetime.) The other possibility is that some outside force (the government) is preventing the bond market from working freely. (This is what happened in the 1940s, according to the links provided by Barry Barnitz).

It's interesting to see that pricing worked normally in the 1970s. As inflation went up, yields went up too. This initially caused a loss of purchase power, for a 10-year bond ladder, but as bonds were now yielding more and more, returns eventually catched up, as implied by bond fund mathematics. My opinion is that nominal bonds behaved exactly as we want them to behave, in the 1970s-to-early-1980s high-inflation period. Bond market pricing remained relatively rational, when considering inflation.


An interesting collateral question is this. If the high-inflation/low-nominal-yields situation was ever to happen again in the future (with or without government meddling), what would happen with the real-yield of RRBs? Would it get significantly negative? [I'm thinking out loud:] If that happened, it would be a good time to sell one's highly overvalued RRBs. The tricky question would become: what to buy with the money? Of course, leaving it in cash would be worse than holding higher-yielding (in real terms) nominal bonds or RRBs. Maybe it would be a good time to buy real estate or stocks? In the 1940s, gold ownership was actually outlawed in the US, so neither (low-yield) cash nor gold were a refuge against inflation (even though they're often peddled as inflation refuge in the financial media).
Variable Percentage Withdrawal (finiki.org/wiki/VPW) | One-Fund Portfolio (VBAL in all accounts)
Park
Veteran Contributor
Veteran Contributor
Posts: 1724
Joined: 13 Aug 2010 20:51

Re: Clippings 2017

Post by Park »

longinvest wrote: 27 Dec 2017 10:21 I don't believe that simplistic measures (YTM for bonds, CAPE for stocks) can predict future long-term returns.

https://www.starcapital.de/fileadmin/us ... imling.pdf

How do you explain the above publication, which shows that CAPE and price book have some predictive ability, when it comes to future stock market returns?
longinvest
Veteran Contributor
Veteran Contributor
Posts: 3956
Joined: 10 Sep 2012 17:26
Location: QC

Re: Clippings 2017

Post by longinvest »

Park wrote: 28 Dec 2017 02:49
longinvest wrote: 27 Dec 2017 10:21 I don't believe that simplistic measures (YTM for bonds, CAPE for stocks) can predict future long-term returns.

https://www.starcapital.de/fileadmin/us ... imling.pdf

How do you explain the above publication, which shows that CAPE and price book have some predictive ability, when it comes to future stock market returns?
By digging long enough through historical data, one can "show" many things. Here's a nice cartoon illustrating the principle: https://xkcd.com/882/

I rely on logical reasoning which is independent from historical returns. I use historical returns to find counterexamples to claims (and prove them false); I don't use historical returns to predict the future.

Most of the financial literature about valuation metrics is based on analyzing historical returns to predict the future. Any writing doing that has no credibility whatsoever for me. A broken clock is right twice a day, that means that it's right more than 700 times every year. Books could be written about the partial accuracy of a broken clock. This doesn't make it anymore useful to get the current time.

CAPE (and other simplistic measures) weren't able to predict the demise of Enron or the raise of Apple. They're simply useless to predict the future, more or less like a broken clock. This doesn't mean that they won't be right from time to time, like a broken clock.

My posts above about nominal bond funds long-term returns explain how these long-term returns have little to do with the initial state of the market when a bond fund was bought. Yes, there's a small impact from the initial curve shape (relative to the final curve shape) and a small contribution from the initial yield of the longest rung, but the main driver of future long-term returns happens to be future yields of rungs bought in the future with reinvested coupons and reinvested maturing principals. Translated into the stock world, this is like saying that future stock long-term returns will be mostly dependent on future net earnings and future prices. It's kind of obvious.

Unfortunately, obvious principles can seem utterly hard to grasp by lots of interested parties in the financial world. "It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" -- Upton Sinclair
Last edited by longinvest on 28 Dec 2017 10:25, edited 2 times in total.
Variable Percentage Withdrawal (finiki.org/wiki/VPW) | One-Fund Portfolio (VBAL in all accounts)
Park
Veteran Contributor
Veteran Contributor
Posts: 1724
Joined: 13 Aug 2010 20:51

Re: Clippings 2017

Post by Park »

When I talk about CAPE, I'm not talking about individual stocks or bonds. I"m talking about national stock markets. Please see figure at the top of page 11. Historically when CAPE is greater than 60, real return over the next 10-15 years has been a maximum of about 2%. When CAPE is less than 10, it has been a minimum of 5%.
brucecohen
Veteran Contributor
Veteran Contributor
Posts: 13310
Joined: 20 Feb 2005 16:47

Re: Clippings 2017

Post by brucecohen »

NYT has an interesting article on how Sweden is embracing, not just coping with, the introduction of industrial robots. That includes unions as well as management.
DenisD
Veteran Contributor
Veteran Contributor
Posts: 4081
Joined: 19 Feb 2005 01:24
Location: Calgary

Re: Clippings 2017

Post by DenisD »

Don’t Demonize Buybacks by Larry Swedroe
Asness, Hazelkorn and Richardson concluded: “Aggregate share repurchase activity has not been at historical highs when measured properly, and when netted against debt issuance is almost a non-event, does not mechanically create earnings (EPS) growth, does not stifle aggregate investment activity, and has not been the primary cause for recent stock market strength. These myths should be discarded.”
User avatar
AltaRed
Veteran Contributor
Veteran Contributor
Posts: 33398
Joined: 05 Mar 2005 20:04
Location: Ogopogo Land

Re: Clippings 2017

Post by AltaRed »

Pretty much sums up my view of share repurchases too. That said, I worry about companies increasing debt too much, only to be handicapped when interest rates rise, and I worry about executive/Board perceptions of company shares being undervalued (versus overvalued) when they might have a rather distorted opinion of themselves.

That all said #2, I have a lot more misgivings on equity dilution to solve a balance sheet problem, or worse to fund stock options and DRIP programs. DRIP programs are fine as long as the company buys back enough shares to take care of dilution.
Imagefiniki, the Canadian financial wiki The go-to place to bolster your financial freedom
Locked