91% real loss in bonds - is this an accurate statement?
91% real loss in bonds - is this an accurate statement?
"Investors who bought Treasury bonds in 1946, when yields were around current levels, did not suffer a formal default.
But over the following 35 years they lost money in real terms at a rate of 2% a year. The cumulative real loss was 91%."
And if so, any idea on what the assumptions were?
October 19 edition of The Economist, 2nd paragraph
http://www.economist.com/news/finance-a ... ey-theres/
But over the following 35 years they lost money in real terms at a rate of 2% a year. The cumulative real loss was 91%."
And if so, any idea on what the assumptions were?
October 19 edition of The Economist, 2nd paragraph
http://www.economist.com/news/finance-a ... ey-theres/
70/30 AA for life, Global market cap equity. Rebalance if fixed income <25% or >35%. Weighted ER< .10%. 5% of annual portfolio balance SWR, Proportional (to AA) withdrawals.
Re: 91% real loss in bonds - is this an accurate statement?
Seeing as the Treasury does not issue 67 year bonds this statement seems to have a flaw.
Re: 91% real loss in bonds - is this an accurate statement?
They say 35 years but I had the same thought, we didn't have 35 year T-bonds to buy in 1946 did we?midareff wrote:Seeing as the Treasury does not issue 67 year bonds this statement seems to have a flaw.
70/30 AA for life, Global market cap equity. Rebalance if fixed income <25% or >35%. Weighted ER< .10%. 5% of annual portfolio balance SWR, Proportional (to AA) withdrawals.
Re: 91% real loss in bonds - is this an accurate statement?
The CPI got to 13.29% in 1979 and 12.52% in 1980. AAA Corporates were paying double digits from 1979 to 1984. Why would anyone with a brain have kept that low paying T paper? This is just a case of someone picking a time period and a set of numbers to cite meaningless babble.
- market timer
- Posts: 6535
- Joined: Tue Aug 21, 2007 1:42 am
Re: 91% real loss in bonds - is this an accurate statement?
I didn't read the article, but wonder if the journalist calculated the 91% loss by compounding 1.87% (rounds up to 2%) over 35 years to get 91% (1.0187^35 = 1.91). This would be an error, though. The correct calculation would be (1 - 1.87%)^35 = .517, i.e., a 48% loss.
Re: 91% real loss in bonds - is this an accurate statement?
I don't work with records prior to 1960 ... no finite reason, just don't. From 1960 through 2012 the CPI averaged 3.97% and AAA's 7.44%
What's different now is that from current levels interest rate's have little room to fall and much room to raise. We know that returns on bond funds will be subject to interest rate sensitivity commensurate with their duration going forward. It is suggested there is more reason to think it will negatively impact fund prices than other scenarios, and will result in muted bond fund returns over an uncertain, but perhaps extended upcoming time period.
For much other than that you need someone with a crystal ball much clearer than mine.
What's different now is that from current levels interest rate's have little room to fall and much room to raise. We know that returns on bond funds will be subject to interest rate sensitivity commensurate with their duration going forward. It is suggested there is more reason to think it will negatively impact fund prices than other scenarios, and will result in muted bond fund returns over an uncertain, but perhaps extended upcoming time period.
For much other than that you need someone with a crystal ball much clearer than mine.
Last edited by midareff on Fri Oct 25, 2013 11:27 am, edited 1 time in total.
-
- Posts: 13356
- Joined: Tue Mar 23, 2010 1:45 pm
- Location: Reading, MA
Re: 91% real loss in bonds - is this an accurate statement?
Nice catch, MT; that certainly must be it...
Attempted new signature...
Re: 91% real loss in bonds - is this an accurate statement?
Using Long Treasury Total returns (which implies rebalancing into the long index each year)....
$100 invested at the end of 1946 would have a portfolio of $205.736 at the end of 1981.
Inflation at 2% over that time would result in a dollar of goods costing 1.99989 by 1981.
Dividing the 205.74 by the inflated currency result in a real increase in value from $100 to $102.87. Not much but a positive real return.
$100 invested at the end of 1946 would have a portfolio of $205.736 at the end of 1981.
Inflation at 2% over that time would result in a dollar of goods costing 1.99989 by 1981.
Dividing the 205.74 by the inflated currency result in a real increase in value from $100 to $102.87. Not much but a positive real return.
-
- Posts: 25625
- Joined: Thu Apr 05, 2007 8:20 pm
- Location: New York
Re: 91% real loss in bonds - is this an accurate statement?
Certain investment funds such as trusts, endowments, pension funds, governmental entities are required to hold the highest rated paper. Generally speaking, government paper is the highest rated paper. No AAA Corporate and I'm sure there weren't too many of them (then or now) is safer than a US Treasury from a risk perspective. I'd take a lower paying US Treasury and save the risk taking for equity allocation.midareff wrote:The CPI got to 13.29% in 1979 and 12.52% in 1980. AAA Corporates were paying double digits from 1979 to 1984. Why would anyone with a brain have kept that low paying T paper? This is just a case of someone picking a time period and a set of numbers to cite meaningless babble.
"One should invest based on their need, ability and willingness to take risk - Larry Swedroe" Asking Portfolio Questions
- InvestorNewb
- Posts: 1663
- Joined: Mon Sep 03, 2012 11:27 am
Re: 91% real loss in bonds - is this an accurate statement?
Why are bonds regarded as a safe asset class if losses can be this great?
My Portfolio: VTI [US], VXUS [Int'l], VNQ [REIT], VCN [Canada] (largest to smallest)
Re: 91% real loss in bonds - is this an accurate statement?
I've read various numbers in the 50% real loss range bandied about from WWII era bond investments to the 1980's rate peak.
That's why I was very surprised to read this statement. I think the compounded loss calculation error mentioned by MT is what happened.
That's why I was very surprised to read this statement. I think the compounded loss calculation error mentioned by MT is what happened.
70/30 AA for life, Global market cap equity. Rebalance if fixed income <25% or >35%. Weighted ER< .10%. 5% of annual portfolio balance SWR, Proportional (to AA) withdrawals.
-
- Posts: 1093
- Joined: Tue Mar 26, 2013 12:13 pm
Re: 91% real loss in bonds - is this an accurate statement?
This seems off. Using the 1946 ten year bond rate (2.19%) and the recorded annual inflation rates I'm getting that you'd have almost 60% of your money left after 30 years with that dropping to ~40% after 35 years.
Amount Left By Year
Amount Left By Year
Code: Select all
Year Amount left
1946 93.890%
1947 82.426%
1948 77.555%
1949 80.184%
1950 80.897%
1951 76.278%
1952 76.499%
1953 77.563%
1954 78.718%
1955 80.757%
1956 81.314%
1957 80.412%
1958 79.921%
1959 81.112%
1960 81.510%
1961 82.480%
1962 83.461%
1963 84.204%
1964 84.953%
1965 85.455%
1966 84.848%
1967 84.076%
1968 82.386%
1969 79.659%
1970 76.863%
1971 75.164%
1972 74.405%
1973 71.421%
1974 65.129%
1975 60.629%
1976 58.440%
1977 55.921%
1978 52.896%
1979 48.077%
1980 42.640%
1981 39.181%
Re: 91% real loss in bonds - is this an accurate statement?
I think the author was challenging this notion. Really, it ends up being a defense of modern portfolio theory. Too much "safety", can be unsafe.InvestorNewb wrote:Why are bonds regarded as a safe asset class if losses can be this great?
"A new truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it."-MP
Re: 91% real loss in bonds - is this an accurate statement?
Bonds aren't a safe asset class all varieties of hazards contemplated.InvestorNewb wrote:Why are bonds regarded as a safe asset class if losses can be this great?
A bond is safe in the sense that one gets exactly the face value of the bond back at maturity, failing default (ie. not even that is perfectly safe). A bond is also safe in the sense that the promised interest payments are known exactly and will appear, again failing default.
So how many ways can you think of that this is not good enough? One start would be the condition in the statement that we want to talk about real value and not nominal value.
If what I am writing here might sound a little patronizing, that is not personal but rather a reaction to the meaningless use of words like "safe." It is far better to reason about things financial by talking about what actually happens, preferentially in quantitative terms, rather than to reason by label, such as "safe."
Re: 91% real loss in bonds - is this an accurate statement?
It is worth noting that today we have TIPS to help with the problem of inflation which played a part in this time period.
Unfortunately while the 91% is wrong, a ~50% loss in buying power was bad enough.
Unfortunately while the 91% is wrong, a ~50% loss in buying power was bad enough.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
Re: 91% real loss in bonds - is this an accurate statement?
While the dates were obviously chosen to show the worst case and maybe the numbers are off a bit, I think the point is still instructive. As others have noted, when evaluated from a real-return standpoint, buy-and-hold bond investors can and have experienced substantial negative real returns over very long holding periods.midareff wrote:This is just a case of someone picking a time period and a set of numbers to cite meaningless babble.
I think the 'bonds are for safety' camp sets up a false sense of security by implying that bonds aren't risky. While it's true that adding fixed income to an equity portfolio generally reduces its annual volatility, that's not the same as bonds being a safe investment. IMO, the main reason to embrace the riskiness of bonds and include them in an investment portfolio is because their risks have historically mixed well with stocks. Bonds offer a diversification story not a safety story.
IMO 'take your risk in equities' is also a bit misguided. Sure, there are worse investing mistakes one could make, but as far as I know there isn't a single academic basis for the 'take your risk on the equities side' mantra. Some feel credit risk is the same as equity risk, but the research is very mixed on this. Others say that credit risk isn't appropriately rewarded by the market, but we bogleheads know better than to fall for that story.
So it seems to me that a sensible approach is to set your stock/bond mix based on your ability and need to take risk and then just hold the market, for both stocks and bonds. Don't worry if your bonds drop in value from time to time, that's what risky investments do. No one knows if we're in for 40 years of negative real bond returns or if we're on the cusp of some great deflationary shock that will greatly reward bond holders as stocks tumble. Hold the market and stay the course.
Jim
(edited to fix typo)
Last edited by magellan on Fri Oct 25, 2013 4:04 pm, edited 1 time in total.
Re: 91% real loss in bonds - is this an accurate statement?
It is what Bill Bernstein refers to as Deep Risk and Shallow Risk.InvestorNewb wrote:Why are bonds regarded as a safe asset class if losses can be this great?
Stocks offer the risk of losing 50% of your nominal dollars overnight - but you eventually get it back. Shallow Risk.
An all bond portfolio has the potential to lose 90% of your purchasing power over decades - and you never get it back. Deep Risk.
It is a quirk of human nature that people fear shallow risk when the biggest threat to their prosperity is deep risk.
- nisiprius
- Advisory Board
- Posts: 52212
- Joined: Thu Jul 26, 2007 9:33 am
- Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry
Re: 91% real loss in bonds - is this an accurate statement?
THE 91% REAL LOSS FIGURE IS BOGUS. In the first place it is wrong, and in the second place it is for long-term bonds, not intermediate-term bonds.
Added. However, my own numbers are wrong. See the correction below.
The SBBI series include "intermediate-term government bonds" and "long-term government bonds." Figures are from my Ibbotson SBBI 2010 Classic Yearbook, table 5-2.
The "inflation adjusted series" of "indices of year-end cumulative wealth" are as follows:
Year-end 1945 + 35 years = Year-end 1980
Year-end 1945, long-term government bonds, 2.094
Year-end 1980, long-term government bonds, 1.054
Real loss, 49.7%
-->Notice how close this is to Market Timer's calculation and YttriumNitrate's estimate.<--
I do NOT think it is fair to excuse a misstatement of this magnitude by saying "Well, anyway, it was bad." Half a loaf is better than 1/11 of a loaf.
But it is worse than that. The scaremongers always talk about long-term bonds. I do not know whether this is an innocent error that reflect differences between what investment professionals track and what ordinary folks invest in, but most retirement savers are not putting the core of their fixed income in long term bonds, they are putting it in intermediate-term bonds. The figures for intermediate-term bonds are:
Year-end 1945, intermediate-term government bonds, 1.753
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 14.0%
Added. However, my own numbers are wrong. See the correction below.
The SBBI series include "intermediate-term government bonds" and "long-term government bonds." Figures are from my Ibbotson SBBI 2010 Classic Yearbook, table 5-2.
The "inflation adjusted series" of "indices of year-end cumulative wealth" are as follows:
Year-end 1945 + 35 years = Year-end 1980
Year-end 1945, long-term government bonds, 2.094
Year-end 1980, long-term government bonds, 1.054
Real loss, 49.7%
-->Notice how close this is to Market Timer's calculation and YttriumNitrate's estimate.<--
I do NOT think it is fair to excuse a misstatement of this magnitude by saying "Well, anyway, it was bad." Half a loaf is better than 1/11 of a loaf.
But it is worse than that. The scaremongers always talk about long-term bonds. I do not know whether this is an innocent error that reflect differences between what investment professionals track and what ordinary folks invest in, but most retirement savers are not putting the core of their fixed income in long term bonds, they are putting it in intermediate-term bonds. The figures for intermediate-term bonds are:
Year-end 1945, intermediate-term government bonds, 1.753
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 14.0%
Last edited by nisiprius on Fri Oct 25, 2013 8:01 pm, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
Re: 91% real loss in bonds - is this an accurate statement?
http://home.comcast.net/~rodec/finance/ ... nBonds.pdfnisiprius wrote:THE 91% REAL LOSS FIGURE IS BOGUS. In the first place it is wrong, and in the second place it is for long-term bonds, not intermediate-term bonds.
The SBBI series include "intermediate-term government bonds" and "long-term government bonds." Figures are from my Ibbotson SBBI 2010 Classic Yearbook, table 5-2.
The "inflation adjusted series" of "indices of year-end cumulative wealth" are as follows:
Year-end 1945 + 35 years = Year-end 1980
Year-end 1945, long-term government bonds, 2.094
Year-end 1980, long-term government bonds, 1.054
Real loss, 49.7%
-->Notice how close this is to Market Timer's calculation and YttriumNitrate's estimate.<--
I do NOT think it is fair to excuse a misstatement of this magnitude by saying "Well, anyway, it was bad." Half a loaf is better than 1/11 of a loaf.
But it is worse than that. The scaremongers always talk about long-term bonds. I do not know whether this is an innocent error that reflect differences between what investment professionals track and what ordinary folks invest in, but most retirement savers are not putting the core of their fixed income in long term bonds, they are putting it in intermediate-term bonds. The figures for intermediate-term bonds are:
Year-end 1945, intermediate-term government bonds, 1.753
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 14.0%
In the above we see that using 10-year US bonds, with returns reflecting both yield and estimated changes in NAV from Global Financial Data, and looking at 30-year horizons, the early 1980s show a worst case for GAGR of -1.7% real per year, or about a 40% loss in total purchasing power if you lump summed your bond purchase. Worse, if like most of us you started 30 years earlier and bought month by month your CAGR was -3.8% real.
Bonds really can stink. Don't kid yourself. However, it is slowly so perhaps easier to adjust to than the sudden loss that stocks can provide.
In part this comes from losing money not only to inflation but to rising interest rates. But again, here is a case where TIPS might have beaten nominal bonds and stocks if they were available.
Stocks by the way were not good in that period either, but were better. A mix of stocks and bond were the best, but even a lump sum 60/40 portfolio only returned 1.7% real CAGR.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
-
- Posts: 184
- Joined: Tue May 17, 2011 4:16 pm
Re: 91% real loss in bonds - is this an accurate statement?
Also, even if they had calced the loss correctly, it would still only be applicable for zero coupon bonds (which pay no interest).
For regular Treasuries, you would have received coupon payments over the years which could have been invested at higher interest rates when rates rose.
For regular Treasuries, you would have received coupon payments over the years which could have been invested at higher interest rates when rates rose.
Re: 91% real loss in bonds - is this an accurate statement?
In terms of possible risk:Rodc wrote:It is worth noting that today we have TIPS to help with the problem of inflation which played a part in this time period.
Unfortunately while the 91% is wrong, a ~50% loss in buying power was bad enough.
I heard that in the case of high inflation occurring in US, they will consult the Argentinian government inflation rate experts for help in determining the us official inflation rate for tips.
; )
Re: 91% real loss in bonds - is this an accurate statement?
Grt2bOutdoors wrote:Certain investment funds such as trusts, endowments, pension funds, governmental entities are required to hold the highest rated paper. Generally speaking, government paper is the highest rated paper. No AAA Corporate and I'm sure there weren't too many of them (then or now) is safer than a US Treasury from a risk perspective. I'd take a lower paying US Treasury and save the risk taking for equity allocation.midareff wrote:The CPI got to 13.29% in 1979 and 12.52% in 1980. AAA Corporates were paying double digits from 1979 to 1984. Why would anyone with a brain have kept that low paying T paper? This is just a case of someone picking a time period and a set of numbers to cite meaningless babble.
Thanks for the input Grt but the original post was talking about investors, implying individuals, not purchase restricted organizations. It's been well discussed that VG's ITIG (as an example) fund acts about 11% equity like in a crisis. That drop is generally fully recovered in 18-20 months range, or earlier. I don't see the need to take an ongoing monthly/annual loss of purchasing power, so I'll take that 11% risk as opposed to the alternatives.
Re: 91% real loss in bonds - is this an accurate statement?
Perfection is hard to come by, but if this keeps you up at night I hear you can buy gold foil from which inflation proof hats can be made.LH wrote:In terms of possible risk:Rodc wrote:It is worth noting that today we have TIPS to help with the problem of inflation which played a part in this time period.
Unfortunately while the 91% is wrong, a ~50% loss in buying power was bad enough.
I heard that in the case of high inflation occurring in US, they will consult the Argentinian government inflation rate experts for help in determining the us official inflation rate for tips.
; )
If that happens and even TIPS prove (near) worthless I will just have to hope at least my 60% in stocks survive.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
- Dale_G
- Posts: 3466
- Joined: Tue Feb 20, 2007 4:43 pm
- Location: Central Florida - on the grown up side of 85
Re: 91% real loss in bonds - is this an accurate statement?
And don't forget that the maximum ordinary tax rates over the period averaged about 70%. Lower for ordinary folks, but it is still not much fun to pay taxes on an investment that is losing buying power.
Tips will do better, but many will still pay taxes on appreciation due to inflation.
Dale
Tips will do better, but many will still pay taxes on appreciation due to inflation.
Dale
Volatility is my friend
- Barry Barnitz
- Wiki Admin
- Posts: 3353
- Joined: Mon Feb 19, 2007 9:42 pm
- Contact:
Re: 91% real loss in bonds - is this an accurate statement?
Hi:
See Developed market bonds - Bogleheads. Note that the historical downside real return periods in both US and UK markets both occurred during eras when individuals did not have tax-advantaged plans available for holding bonds. The results do not reflect the tax drag on nominal interest income.
regards,
See Developed market bonds - Bogleheads. Note that the historical downside real return periods in both US and UK markets both occurred during eras when individuals did not have tax-advantaged plans available for holding bonds. The results do not reflect the tax drag on nominal interest income.
regards,
Additional administrative tasks: Financial Page bogleheads.org. blog; finiki the Canadian wiki; The Bogle Center for Financial Literacy site; La Guía Bogleheads® España site.
- happyisland
- Posts: 915
- Joined: Thu Oct 03, 2013 1:36 pm
- Location: nos baranca tan stima
Re: 91% real loss in bonds - is this an accurate statement?
How would the bond portion of a typical 3-fund BH portfolio have performed during this time? I'm just curious how periodic rebalancing and dividend reinvestment would have affected things...
Re: 91% real loss in bonds - is this an accurate statement?
See my link above. It is only US and 10-year government bonds but will show give you a pretty good approximation to your question.happyisland wrote:How would the bond portion of a typical 3-fund BH portfolio have performed during this time? I'm just curious how periodic rebalancing and dividend reinvestment would have affected things...
Unfortunately stock were pretty stinky at the same time.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
Re: 91% real loss in bonds - is this an accurate statement?
Thank you for invariably sorting the "wheat from the chaff"nisiprius wrote:THE 91% REAL LOSS FIGURE IS BOGUS. In the first place it is wrong, and in the second place it is for long-term bonds, not intermediate-term bonds.
The SBBI series include "intermediate-term government bonds" and "long-term government bonds." Figures are from my Ibbotson SBBI 2010 Classic Yearbook, table 5-2.
The "inflation adjusted series" of "indices of year-end cumulative wealth" are as follows:
Year-end 1945 + 35 years = Year-end 1980
Year-end 1945, long-term government bonds, 2.094
Year-end 1980, long-term government bonds, 1.054
Real loss, 49.7%
-->Notice how close this is to Market Timer's calculation and YttriumNitrate's estimate.<--
I do NOT think it is fair to excuse a misstatement of this magnitude by saying "Well, anyway, it was bad." Half a loaf is better than 1/11 of a loaf.
But it is worse than that. The scaremongers always talk about long-term bonds. I do not know whether this is an innocent error that reflect differences between what investment professionals track and what ordinary folks invest in, but most retirement savers are not putting the core of their fixed income in long term bonds, they are putting it in intermediate-term bonds. The figures for intermediate-term bonds are:
Year-end 1945, intermediate-term government bonds, 1.753
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 14.0%
Re: 91% real loss in bonds - is this an accurate statement?
Because it took place over a 35-year period. Sort of like being run over by a glacier (actually I think glaciers move faster than that).InvestorNewb wrote:Why are bonds regarded as a safe asset class if losses can be this great?
-
- Posts: 1926
- Joined: Sat Dec 22, 2012 9:46 am
Re: 91% real loss in bonds - is this an accurate statement?
It bothers me to no end when real returns are used to stress a point vs nominal. Happens on the board all the time as well. I understand the idea but we don't live in an inflation adjusted world so it bends the numbers into terms contrary to our normal way of thinking.
- nisiprius
- Advisory Board
- Posts: 52212
- Joined: Thu Jul 26, 2007 9:33 am
- Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry
Re: 91% real loss in bonds - is this an accurate statement?
Folks, I screwed up. I read one year off and posted numbers for year-end 1946 instead of 1945. The correct numbers (I hope, this time) are:
The "inflation adjusted series" of "indices of year-end cumulative wealth" are as follows:
Year-end 1945 + 35 years = Year-end 1980
Year-end 1945, long-term government bonds, 2.477
Year-end 1980, long-term government bonds, 1.054
Real loss, 57.4%
Year-end 1945, intermediate-term government bonds, 2.051
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 26.5%
I still maintain that 57.4% is a far cry from 91%, and that the results intermediate-term bonds is more relevant to the kind of investing most Bogleheads do than the results for long-term bonds.
The "inflation adjusted series" of "indices of year-end cumulative wealth" are as follows:
Year-end 1945 + 35 years = Year-end 1980
Year-end 1945, long-term government bonds, 2.477
Year-end 1980, long-term government bonds, 1.054
Real loss, 57.4%
Year-end 1945, intermediate-term government bonds, 2.051
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 26.5%
I still maintain that 57.4% is a far cry from 91%, and that the results intermediate-term bonds is more relevant to the kind of investing most Bogleheads do than the results for long-term bonds.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
Re: 91% real loss in bonds - is this an accurate statement?
That gets you closer to the -40% I got for a slightly different period using 10-year government bonds. Given you are not correcting for changes in NAV due to increasing interest rates, I suspect you are under estimating the loss if using bond funds. Either way a loss over such a long holding period, not a transient loss such as generally seen in stocks, is not what most people think of when they think of the "safety of bonds".nisiprius wrote:Folks, I screwed up. I read one year off and posted numbers for year-end 1946 instead of 1945. The correct numbers (I hope, this time) are:
The "inflation adjusted series" of "indices of year-end cumulative wealth" are as follows:
Year-end 1945 + 35 years = Year-end 1980
Year-end 1945, long-term government bonds, 2.477
Year-end 1980, long-term government bonds, 1.054
Real loss, 57.4%
Year-end 1945, intermediate-term government bonds, 2.051
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 26.5%
I still maintain that 57.4% is a far cry from 91%, and that the results intermediate-term bonds is more relevant to the kind of investing most Bogleheads do than the results for long-term bonds.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
Re: 91% real loss in bonds - is this an accurate statement?
nisiprius originally wrote:The figures for intermediate-term bonds are:
Year-end 1945, intermediate-term government bonds, 1.753
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 14.0%
Nisi, using a completely different method I come up with figures similar to your original (and you say incorrect) 14.0% figure. I get a 13.4% real decrease for the 35 years from Jan 1 1946 to Jan 1 1981. I used the yields on the 5-Year Treasury Constant Maturity Rate (GS5) for January 1 1956, 1961, 1966, 1971, and 1976. Since the 5-year FRED data series doesn't go back before 1953, I used the 2.19% Jan 1 1946 rate for the first 10 years from 10 Year Treasury Rate by Year. The CPI figures are from CPI-U Since 1913.nisiprius in correction wrote: Year-end 1945, intermediate-term government bonds, 2.051
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 26.5%
Code: Select all
Yield Nominal CPI Annual Real Real
Jan 1 Growth Jan Change Yield Growth
----- ------ ----- ------ ----- ------
1946 2.19% 1.000 18.2 3.95% (1.69%) 1.000
1956 2.84% 1.242 26.8 2.14% 0.68% 0.843
1961 3.67% 1.429 29.8 1.31% 2.33% 0.872
1966 4.86% 1.711 31.8 4.59% 0.26% 0.979
1971 5.89% 2.169 39.8 6.91% (0.96%) 0.992
1976 7.46% 2.887 55.6 9.37% (1.74%) 0.945
1981 4.137 87.0 0.866
A 13.4% fall over 35 years is only 0.4% per year, well short of the 2% annual decline in the quote from the Economist. Besides not specifying term of the bonds, the article also doesn't specify whether the result is pretax or after tax. But if one assumes a 40% tax rate, $1 would decline to $0.494 in real after tax terms, which is equivalent to 2% per year.
Code: Select all
Yield Aftax Nominal CPI Annual Real Real
Jan 1 Yield Growth Jan Change Yield Growth
----- ----- ------ ----- ------ ----- ------
1946 2.19% 1.31% 1.000 18.2 3.95% (2.53%) 1.000
1956 2.84% 1.70% 1.139 26.8 2.14% (0.43%) 0.774
1961 3.67% 2.20% 1.240 29.8 1.31% 0.88% 0.757
1966 4.86% 2.92% 1.383 31.8 4.59% (1.60%) 0.791
1971 5.89% 3.53% 1.596 39.8 6.91% (3.16%) 0.730
1976 7.46% 4.48% 1.899 55.6 9.37% (4.47%) 0.622
1981 2.364 87.0 0.494
* As DetroitRed points out in this post, the actual growth would be somewhat higher since, with rising interest interest rates, coupons could be reinvested at more than the original yield.
Re: 91% real loss in bonds - is this an accurate statement?
FWIW the 40% loss shown in the Global Financial Data dataset using 10 year maturity is pretax. So the 1.7% real AGR loss over 30 years (about a 40% total loss in real dollars) would be worse because there were no tax advantaged accounts at the time.the article also doesn't specify whether the result is pretax or after tax.
Note well that the loss was far greater for someone making periodic purchases.
Agree this is a far cry from the 91%. But, I'm not sure if I had a compound loss in my bonds over all of my accumulation phase of anything like being discussed (14% to 40% or more due to periodic buys and/or taxes) that I would conclude current wisdom about the safety of bonds was particularly correct.
I haven't the slightest clue about the future other than another long bout of rising rates and inflation is not off the table.
This certainly could be repeated over the next 30 years.
My conclusion is that a very heavy nominal treasury portfolio may have more risk than many assume or corporates for that matter. So, I will keep a healthy dose of TIPS and stocks in addition to nominal treasury bonds. Might make sense to keep corp bonds as well though for now I take my company risk on the stock side (somewhat agnostic as to whether or not that makes complete sense).
Note: I'm using 10 year maturity, not sure what nisi is using, #cruncher is using 5-year, so there is some discrepancy in the datasets we each have available.
#cruncher, does the constant maturity dataset account for NAV changes due to rising rates? Or are you assuming you buy and hold and only turn over every 5 years, in which case your maturity is actually less than 5 years? Thanks.
Last edited by Rodc on Sat Oct 26, 2013 8:57 am, edited 2 times in total.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
- nisiprius
- Advisory Board
- Posts: 52212
- Joined: Thu Jul 26, 2007 9:33 am
- Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry
Re: 91% real loss in bonds - is this an accurate statement?
What can be said is that over a precisely-picked period of 35 years a hypothetical stay-the-course investor could have lost about half his real value in long-term bonds, or about a quarter of his real value for intermediate terms, and sure as hell wouldn't have lost any 90% of his real value either way. And this is important because it fits the general pattern of the financial press present valid concerns about bonds and amping them up fivefold. The "90% loss" meme is now out there. I'll bet a nickel even money that within the next few months we will see postings from people saying bonds are just as riskier than stocks because they once "lost 90%."#Cruncher wrote:Nisi, using a completely different method I come up with figures similar to your original (and you say incorrect) 14.0% figure.
It is alway endpoints, and I am always insecure about the "fencepost problem" because writers of are never very precise about what they used for endpoints. In the SBBI data, a year like "1946" always means year-end 1946, and I always need to think twice about that... At the start and end of the 35-40-year "lousy bond" period, things were changing rapidly from year to year, so endpoints really matter. And I don't think intermediate and long-term were perfectly synchronized, so the worst period for one isn't exactly the same as the worst period for the other. SBBI is explicit about how they calculate total return but I am not too sure how it correlates with the hypothetical experience of an investor holding Vanguard Total Bond (doubly hypothetical since neither the fund nor the Lehman Brothers Aggregate index existed!). "To the greatest extent possible, a one-bond portfolio with a term of approximately 20 years and a reasonably current coupon--whose returns did not reflect potential tax benefits, impaired negotiability, or special redemption or call privileges--was used each year. The bond was "held" for the calendar year and returns were calculated."
But this refreshes one of my concerns about financial data. Whenever you read the fine print, the neatly tabulated data always turns out to include a metric boatload of choices, assumptions, decisions, judgement, and good taste about what to select and what to calculate. Hopefully the people who do it aren't grinding an axe and are making decisions that other experts would endorse as sensible. But the problem is once tabulated and published, they are instantly enshrined as if they had the precision of astronomical observations, and people do calculations on them and give the results to three significant places--when the original data is for, quote, "a term of approximately 20 years and a reasonably current coupon!"
I've commented before on the curious choice of 1926 as a starting point for the CRSP data. It's not an unreasonable choice and I don't think it was done to grind an axe, but it was basically one guy's (Dr. James Lorie, probably) tasteful choice to include one full business cycle or something, whatever that means... and a shift in the starting point by two or three years either way results in nearly a 1% spread of values for "the historical return of the stock market" over an eight-decade time period!
Last edited by nisiprius on Sat Oct 26, 2013 9:06 am, edited 2 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
Re: 91% real loss in bonds - is this an accurate statement?
Agree completely.But this refreshes one of my concerns about financial data. Whenever you read the fine print, the neatly tabulated data always turns out to include a metric boatload of choices, assumptions, decisions, judgement, and good taste about what to select and what to calculate. Hopefully the people who do it aren't grinding an axe and are making decisions that other experts would endorse as sensible. But the problem is once tabulated and published, they are instantly enshrined as if they had the precision of astronomical observations, and people do calculations on them and give the results to three significant places--when the original data is for, quote, "a term of approximately 20 years and a reasonably current coupon!"
I've commented before on the curious choice of 1926 as a starting point for the CRSP data. It's not an unreasonable choice and I don't think it was done to grind an axe, but it was basically one guy's (Dr. James Lorie, probably) tasteful choice to include one full business cycle or something, whatever that means... and a shift in the starting point by two or three years either way results in nearly a 1% spread of values for "the historical return of the stock market" over an eight-decade time period!
I'll add that even minor effects from choices and other noise in the data when compounded for 30 years grow exponentially so that things like differences in the end result out 30 years may look large, but really might be meaningless differences.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
Re: 91% real loss in bonds - is this an accurate statement?
well we do live in an inflation adjusted world in that our dollars continually buy less (in general, sometimes more in depression etc). The world experience is continually adjusted to the present time. If you compare last years salary or last years 401k balance to this years nominally, you are not thinking correctly unless inflation is zero or near zero, over the past year, slowly, inflation if present has degraded the worth of money.MN Finance wrote:It bothers me to no end when real returns are used to stress a point vs nominal. Happens on the board all the time as well. I understand the idea but we don't live in an inflation adjusted world so it bends the numbers into terms contrary to our normal way of thinking.
Humans are programmed to think nominally, and it does not feel right really. That 2 percent raise one gets, if prices increase 2 percent, will still (especially initially) feel like a two percent increase, its REAL feeling to humans. But really its not a raise.
We really do not talk real enough here imo, and I have a hard time with it myself. (like looking at dividends, and saying, look I made money, or "breaking even" from the 2007/2008 drop..... both times, inflation adjusted, its false (or less than it appears), but it feels good anyway)
Re: 91% real loss in bonds - is this an accurate statement?
heheheh, "perfection" eh? Regardless, I bet if you can get it past the cap controls, the argentinians would gladly buy your gold foil hats from you : )Rodc wrote:Perfection is hard to come by, but if this keeps you up at night I hear you can buy gold foil from which inflation proof hats can be made.LH wrote:In terms of possible risk:Rodc wrote:It is worth noting that today we have TIPS to help with the problem of inflation which played a part in this time period.
Unfortunately while the 91% is wrong, a ~50% loss in buying power was bad enough.
I heard that in the case of high inflation occurring in US, they will consult the Argentinian government inflation rate experts for help in determining the us official inflation rate for tips.
; )
If that happens and even TIPS prove (near) worthless I will just have to hope at least my 60% in stocks survive.
TIPS are a good invesment, I am 50/50 tips/nominal but there is cpi calculation risk.
also, this type of situation is scary to contemplate I grant, but to devovle into "perfection", and "hat" type stuff, I posit is a mental block.
Argentinia 2001 is real, weimar real, depression real. "CPI calculation risk" is a risk. When things get bad, government stats tend to get bad in the real world historically, and currently in argentina.
I have read that current cpi if calculated by 1980s 1990s method, would be 6 percent, dunno if true or not, have not had time to dig into it, but interesting. New and improved = lower, just like chained cpi.
Last edited by LH on Sat Oct 26, 2013 12:12 pm, edited 3 times in total.
Re: 91% real loss in bonds - is this an accurate statement?
Nothing new.
A Pound used to be valued at a pound weight of silver. Henry VIII in 1526 however started to mix in copper into silver coins and devalued to 1/6th of former value. He became known as coppernose - as after some wear the nose on his image on silver coins revealed the copper beneath.
US mid 1930's, gold was compulsory purchased at the value it had been priced for 200 odd years, only to have the price hiked.
UK 1970's high inflation, high bond yields, high taxation (negative net real).
Other periods - low yields, high inflation.
In other cases falsely stated inflation (Argentina).
When Kings, Sovereign's or States see their treasury light - seizures in one form or another will occur. The major risk is that whatever you thought might have been a hedge - might turn out to be null and void when needed the most (pay insurance premium in good years to hedge the risk (lower returns/lost opportunity cost), only to see that insurance fail when the heat is on). Which isn't coincidental.
A Pound used to be valued at a pound weight of silver. Henry VIII in 1526 however started to mix in copper into silver coins and devalued to 1/6th of former value. He became known as coppernose - as after some wear the nose on his image on silver coins revealed the copper beneath.
US mid 1930's, gold was compulsory purchased at the value it had been priced for 200 odd years, only to have the price hiked.
UK 1970's high inflation, high bond yields, high taxation (negative net real).
Other periods - low yields, high inflation.
In other cases falsely stated inflation (Argentina).
When Kings, Sovereign's or States see their treasury light - seizures in one form or another will occur. The major risk is that whatever you thought might have been a hedge - might turn out to be null and void when needed the most (pay insurance premium in good years to hedge the risk (lower returns/lost opportunity cost), only to see that insurance fail when the heat is on). Which isn't coincidental.
Re: 91% real loss in bonds - is this an accurate statement?
Good catch, Rod! Yes, I was assuming buying and holding to maturity one 10-year bond followed by buying and holding to maturity five 5-year bonds. So you are right, the average maturity is less than 5 years. Actually it's about 3.2 years = 5 X 10/35 + 2.5 X 25/35. (I got so caught up in crunching the numbers that I didn't realize this.) So the holdings would more accurately be called "short term" than "intermediate term".Rodc wrote:#cruncher, ... are you assuming you buy and hold and only turn over every 5 years, in which case your maturity is actually less than 5 years?
I can get a little closer to intermediate by assuming three 10-year bonds followed by only one 5-year to fill out the 35 year period. The average maturity is then about 4.6 years = 5 X 30/35 + 2.5 X 5/35. This has a significant effect: the pretax real loss is then 22.2% (nominal 3.717 / 4.78 CPI change less 1) compared to the 13.4% I calculated originally. This is closer to nisiprius' corrected 26.5% figure.
Code: Select all
Yield Nominal CPI Annual Real Real
Jan 1 Growth Jan Change Yield Growth
----- ------ ----- ------ ----- ------
1946 2.19% 1.000 18.2 3.95% (1.69%) 1.000
1956 2.90% 1.242 26.8 1.73% 1.15% 0.843
1966 4.61% 1.653 31.8 5.75% (1.07%) 0.946
1976 7.46% 2.594 55.6 9.37% (1.74%) 0.849
1981 3.717 87.0 0.778
Re: 91% real loss in bonds - is this an accurate statement?
UK 1950's basic (lower) rate taxes got near to 50%, higher rate taxes were up at 80% levels.the pretax real loss
Basic (average saver/investor) net real rates are perhaps the better choice of comparison/figures.
Re: 91% real loss in bonds - is this an accurate statement?
Absolutely. Unfortunately while I have nominal bonds which would have helped in the depression (with deflation) and TIPS which would have helped in the 1980ish period (not sure off hand just when high inflation started and stopped), I have no hyper inflation bonds for Argentina in my tool box, but hopefully stocks will help as noted above.Argentinia 2001 is real, weimar real, depression real.
I'm not a believer in holding gold myself, but if in the unlikely (IMHO) event I might wish I did. Fortunately I do own property, maybe that will help.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
Re: 91% real loss in bonds - is this an accurate statement?
Thanks for the update.#Cruncher wrote:Good catch, Rod! Yes, I was assuming buying and holding to maturity one 10-year bond followed by buying and holding to maturity five 5-year bonds. So you are right, the average maturity is less than 5 years. Actually it's about 3.2 years = 5 X 10/35 + 2.5 X 25/35. (I got so caught up in crunching the numbers that I didn't realize this.) So the holdings would more accurately be called "short term" than "intermediate term".Rodc wrote:#cruncher, ... are you assuming you buy and hold and only turn over every 5 years, in which case your maturity is actually less than 5 years?
I can get a little closer to intermediate by assuming three 10-year bonds followed by only one 5-year to fill out the 35 year period. The average maturity is then about 4.6 years = 5 X 30/35 + 2.5 X 5/35. This has a significant effect: the pretax real loss is then 22.2% (nominal 3.717 / 4.78 CPI change less 1) compared to the 13.4% I calculated originally. This is closer to nisiprius' corrected 26.5% figure.Code: Select all
Yield Nominal CPI Annual Real Real Jan 1 Growth Jan Change Yield Growth ----- ------ ----- ------ ----- ------ 1946 2.19% 1.000 18.2 3.95% (1.69%) 1.000 1956 2.90% 1.242 26.8 1.73% 1.15% 0.843 1966 4.61% 1.653 31.8 5.75% (1.07%) 0.946 1976 7.46% 2.594 55.6 9.37% (1.74%) 0.849 1981 3.717 87.0 0.778
Given the amount of change that probably gets things reasonably in line with my results as my maturity is a constant 10 years.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
Re: 91% real loss in bonds - is this an accurate statement?
Sounds like a safer bet than TBM over the next 35 years.nisiprius wrote:The "90% loss" meme is now out there. I'll bet a nickel even money that within the next few months we will see postings from people saying bonds are just as riskier than stocks because they once "lost 90%."
"A new truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it."-MP
-
- Posts: 218
- Joined: Sun Mar 03, 2013 2:57 pm
Re: 91% real loss in bonds - is this an accurate statement?
Does your analysis include the impact of paying taxes on the nominal returns? I don't think IRAs even existed until the late 70s.nisiprius wrote:Folks, I screwed up. I read one year off and posted numbers for year-end 1946 instead of 1945. The correct numbers (I hope, this time) are:
The "inflation adjusted series" of "indices of year-end cumulative wealth" are as follows:
Year-end 1945 + 35 years = Year-end 1980
Year-end 1945, long-term government bonds, 2.477
Year-end 1980, long-term government bonds, 1.054
Real loss, 57.4%
Year-end 1945, intermediate-term government bonds, 2.051
Year-end 1980, intermediate-term government bonds, 1.508
Real loss, 26.5%
I still maintain that 57.4% is a far cry from 91%, and that the results intermediate-term bonds is more relevant to the kind of investing most Bogleheads do than the results for long-term bonds.
-
- Posts: 218
- Joined: Sun Mar 03, 2013 2:57 pm
Re: 91% real loss in bonds - is this an accurate statement?
The depression was a deflationary period, so I don't know why you're bringing it up. If you really want to hedge against a "weimar" or to a lesser extent Argentina type situation, your investment plan probably needs to include a foreign passport and a one way plane ticket.LH wrote: Argentinia 2001 is real, weimar real, depression real. "CPI calculation risk" is a risk. When things get bad, government stats tend to get bad in the real world historically, and currently in argentina.
Re: 91% real loss in bonds - is this an accurate statement?
There is no way to hedge or insure against every possible risk, but a diversified portfolio comes as close as we can.
Bonds are in a portfolio for income and to decrease volatility in a portfolio. But as we have seen, it is possible for bonds to lose purchasing power over some time periods. If the author had picked 1982-2012, the story for bonds particularly long term bonds would have been different. It would have been a glowing article about how bonds beat inflation.
I would also point out that stocks can have negative real rates of return depending on the time periods you pick.
We know that over long time periods (since 1926) that both stocks and bonds have beat inflation. Stocks by a wider margin than bonds. It is that you might have to wait a while to get your inflation adjustment.
Bonds are in a portfolio for income and to decrease volatility in a portfolio. But as we have seen, it is possible for bonds to lose purchasing power over some time periods. If the author had picked 1982-2012, the story for bonds particularly long term bonds would have been different. It would have been a glowing article about how bonds beat inflation.
I would also point out that stocks can have negative real rates of return depending on the time periods you pick.
We know that over long time periods (since 1926) that both stocks and bonds have beat inflation. Stocks by a wider margin than bonds. It is that you might have to wait a while to get your inflation adjustment.
A fool and his money are good for business.
-
- Posts: 49031
- Joined: Fri May 11, 2007 11:07 am
Re: 91% real loss in bonds - is this an accurate statement?
Be warned that we have far too few truly independent 30 year periods (5 say at max for US data) to really tell us if stocks and bonds outperform inflation, or outperform in all possible economic circumstances.nedsaid wrote:There is no way to hedge or insure against every possible risk, but a diversified portfolio comes as close as we can.
Bonds are in a portfolio for income and to decrease volatility in a portfolio. But as we have seen, it is possible for bonds to lose purchasing power over some time periods. If the author had picked 1982-2012, the story for bonds particularly long term bonds would have been different. It would have been a glowing article about how bonds beat inflation.
I would also point out that stocks can have negative real rates of return depending on the time periods you pick.
We know that over long time periods (since 1926) that both stocks and bonds have beat inflation. Stocks by a wider margin than bonds. It is that you might have to wait a while to get your inflation adjustment.
TIPS by their construction will match CPI (ditto ibonds). But even there there are tax issues.
Re: 91% real loss in bonds - is this an accurate statement?
I thought it might be interesting to expand the two calculations I made above to see how four different fixed income durations would have fared over the 35 year period of generally rising interest rates and high inflation from Jan 1 1946 to Jan 1 1981.
Conclusions:
- Ultra short term: buy and hold a 3-month T bill starting 1/1/1946 and rolling it over every 3 months. Source: FRED 3-Month Treasury Bill: Secondary Market Rate (TB3MS).
- Short term: buy a 10-year T note 1/1/1946, hold for 10 years; and then buy and hold until maturity five 5-year T notes starting 1/1/1956. Source: my post # 1 above.
- Intermediate term: buy and hold to maturity three 10-year T notes; and then buy and hold a 5-year T note on 1/1/1976. Source: my post # 2 above.
- Intermediate/long term: buy a 10-year T note 1/1/1946; sell it 1/1/1947 at market price and reinvest in a new 10-year T note. Repeat every year through 1/1/1980. Sell the last one a year later 1/1/1981 at the market price. * Source: 10 Year Treasury Rate by Year.
Code: Select all
---------- Nominal Return ---------- ----------- Real Return -----------
Pretax After 40% tax Pretax After 40% tax
Average ---------------- ---------------- ---------------- ----------------
Maturity 35 Years Annual 35 Years Annual 35 Years Annual 35 Years Annual
-------- -------- ------ -------- ------ -------- ------ -------- ------
0.1 286.6% 3.94% 125.6% 2.35% -19.1% -0.60% -52.8% -2.12% 1: 3 mo T Bills
3.2 313.7% 4.14% 136.4% 2.49% -13.4% -0.41% -50.6% -1.99% 2: 1 X 10Yr - 5 X 5Yr
4.6 271.7% 3.82% 121.4% 2.30% -22.2% -0.72% -53.7% -2.17% 3: 3 X 10Yr - 1 X 5Yr
-26.5% -0.88% Nisi Intermediate
9.5 166.9% 2.85% 82.1% 1.73% -44.2% -1.65% -61.9% -2.72% 4: 10Yr Held 1 Yr
-57.4% -2.41% Nisi Long
- None of the approaches provided a nominal return large enough to exceed the average 4.57% annual increase in the CPI from 18.2 in Jan 1946 to 87.0 in Jan 1981. That is, they all had a negative pretax real return. Source: CPI since 1913.
- Even 3-month Treasury bills, which are often touted as a good inflation hedge, declined an average 0.60% per year before taxes in real terms.
- The best return of the four was # 2 which had an average maturity of about 3.2 years and declined an average 0.41% per year before taxes in real terms.
- As might be expected since interest rates were generally rising -- except for the 3-month T bills -- the longer the average maturity, the worse the return.
Re: 91% real loss in bonds - is this an accurate statement?
Or the investor could have held an equity-heavy portfolio. Here is a table from Bernstein's book on deep risk giving local equity returns (not including dividends) during some of the worst inflationary periods from the 20th century. If dividends were included, equity investors would obviously have done even better. Those holding bond-heavy and cash-heavy portfolios would have suffered catastrophic losses.lostInFinance wrote: If you really want to hedge against a "weimar" or to a lesser extent Argentina type situation, your investment plan probably needs to include a foreign passport and a one way plane ticket.