Crestmont's Perspective: Beware The Assumptions!
Posted: Wed Jul 30, 2003 11:47 am
A. Models are only as good as their assumptions.
B. Valid assumptions are not always good ones.
C. A perfect model with perfect assumptions is rarely valid over the long term.
In any model used to project future savings, whether SWR or other financial planning model, the assumptions are as critical as the structure of the model. Valuation does matter toward future returns in the financial markets. For what it's worth, valuation always affects future returns in any financial instrument or model (by definition).
There has not been an extended period over the past century or more when history can be said to fully repeat itself. Said another way: if an analyst had used historical data before 1950 or before 1975 or before 2000 to estimate what could have been expected over the subsequent 25 years or so, the forecasted scenarios would not have happen exactly as expected. Several key considerations to consider about the using the historical data since the late 1800's as the full analysis set:
1. Prior to the 1960's, interest rates and inflation were not aligned and were inconsistent - after 1960 or so, the relationship is much more consistent (as one would expect fundamentally). [see http://www.crestmontresearch.com/pdfs/i ... onship.pdf]
2. P/E ratios relate to inflation, not to interest rates. This will explain some of the perceived anomalies over the past century. [see http://www.crestmontresearch.com/pdfs/S ... &%20PE.pdf]
3. Stock market returns tend to occur in cycles: P/Es rise when inflation heads toward price stability; EPS (earnings) are fairly constant with the economy over longer periods; therefore stock returns are dependant upon positive trends in P/Es (i.e. inflation trending toward 1%) and suffer when P/Es fall.
4. We are currently in a period of (a) high P/E valuations. (b) low interest rates, and (c) low inflation - significant "Vulnerability" for financial markets and securities. If inflation declines much further (deflation), P/E's will fall. If inflation rises, P/E's will fall as well. The prospects for significant gains in the stock market this decade are dismal (P/Es are likely to decline slightly as EPS increases, with a net result being a very choppy and relatively flat market).
5. CLOSELY NOTE: When using historical data about the stock market and cash returns (especially during the early 1900's), note the substantial dividend yield and cash returns (the commercial paper rate is used in some popular models using historical data). Today, dividend yields and cash yields are significantly lower than the early part of the 20th century. High dividends and cash yields in historical data sets can distort the forecast results from financial planning models (including SWR). Particularly for SWR models, especially in the earlier parts of the 1900's, this enables the hypothetical investor in SWR models to endure large market swings. Without the same level of current cash flow, the SWR models have significantly different results.
6. BIG PICTURE: SWR seems to be about how much a retiree can withdraw from savings without running out of principal over a prescribed long term period. Other retirement planning models help people project the nest egg that they'll have going into retirement. If stocks are destined for mediocre returns (or maybe none over the next decade or so) and other sources of income imbedded in the historical data (i.e. dividends, interest, etc.) are substantially less today, there could be significant implications for investors.
7. Transaction Costs: the assumptions in some SWR and financial planning models assume index returns - without regard for transaction costs, asset management fees, commissions, bid/ask spreads, taxes, etc. There costs can be material - bake them in as well - and total transaction costs often run 1% to 3% annually or more.
8. Techniques: seek advice from an independent financial advisor about techniques to deal with non-trending markets. The stock and bond bull markets of the 1980's and 1990's were trending markets and benefited from "buy-and-hold" and "buy-on-dips" strategies. The next phase could be another secular bear cycle (see http://www.crestmontresearch.com/pdfs/S ... 0Chart.pdf). Note the difference in secular bull and secular bear periods. Ask your advisor about the effect of more frequent rebalancing, covered call option writing, higher yielding preferred stocks and other securities with higher current return, TIPS, hedge funds, and other risk-controlled and value-added investment strategies.
Please consider reviewing our research at www.CrestmontResearch.com relating to stock market cycles, long-term returns, and interest rate cycles. For each of the charts with links listed above, you'll find a brief description within the Stock Market and Interest Rate sections. We welcome your insights and perspectives to further the research. Our analysis is not above challenge: I welcome it. It has not gotten to were it is today without challenge. All along, I've hoped for rational insights that would support strong upside potential in the financial markets. To date, the results instead have been rather sobering. Our research is not intended to generate predictions; rather it is intended to be provocative and enlightening. Hope it's helpful; there's no cost and it's free of banner ads. I welcome your comments, questions, and persepctives either in reponse to this posting or through our website.
Unlike Bill Murray in "˜Groundhog Day', we don't get another chance to replay our retirement scenario. The risk for a retiree or aspiring retiree of an error in their return assumptions is serious: you only find out that the assumptions were insufficient when it's too late to change the course.
Ed Easterling
Crestmont Research
(A similar version of this composition was first posted in the SWR Research Group board; it's been posted here at the suggestion of "peteyperson."
B. Valid assumptions are not always good ones.
C. A perfect model with perfect assumptions is rarely valid over the long term.
In any model used to project future savings, whether SWR or other financial planning model, the assumptions are as critical as the structure of the model. Valuation does matter toward future returns in the financial markets. For what it's worth, valuation always affects future returns in any financial instrument or model (by definition).
There has not been an extended period over the past century or more when history can be said to fully repeat itself. Said another way: if an analyst had used historical data before 1950 or before 1975 or before 2000 to estimate what could have been expected over the subsequent 25 years or so, the forecasted scenarios would not have happen exactly as expected. Several key considerations to consider about the using the historical data since the late 1800's as the full analysis set:
1. Prior to the 1960's, interest rates and inflation were not aligned and were inconsistent - after 1960 or so, the relationship is much more consistent (as one would expect fundamentally). [see http://www.crestmontresearch.com/pdfs/i ... onship.pdf]
2. P/E ratios relate to inflation, not to interest rates. This will explain some of the perceived anomalies over the past century. [see http://www.crestmontresearch.com/pdfs/S ... &%20PE.pdf]
3. Stock market returns tend to occur in cycles: P/Es rise when inflation heads toward price stability; EPS (earnings) are fairly constant with the economy over longer periods; therefore stock returns are dependant upon positive trends in P/Es (i.e. inflation trending toward 1%) and suffer when P/Es fall.
4. We are currently in a period of (a) high P/E valuations. (b) low interest rates, and (c) low inflation - significant "Vulnerability" for financial markets and securities. If inflation declines much further (deflation), P/E's will fall. If inflation rises, P/E's will fall as well. The prospects for significant gains in the stock market this decade are dismal (P/Es are likely to decline slightly as EPS increases, with a net result being a very choppy and relatively flat market).
5. CLOSELY NOTE: When using historical data about the stock market and cash returns (especially during the early 1900's), note the substantial dividend yield and cash returns (the commercial paper rate is used in some popular models using historical data). Today, dividend yields and cash yields are significantly lower than the early part of the 20th century. High dividends and cash yields in historical data sets can distort the forecast results from financial planning models (including SWR). Particularly for SWR models, especially in the earlier parts of the 1900's, this enables the hypothetical investor in SWR models to endure large market swings. Without the same level of current cash flow, the SWR models have significantly different results.
6. BIG PICTURE: SWR seems to be about how much a retiree can withdraw from savings without running out of principal over a prescribed long term period. Other retirement planning models help people project the nest egg that they'll have going into retirement. If stocks are destined for mediocre returns (or maybe none over the next decade or so) and other sources of income imbedded in the historical data (i.e. dividends, interest, etc.) are substantially less today, there could be significant implications for investors.
7. Transaction Costs: the assumptions in some SWR and financial planning models assume index returns - without regard for transaction costs, asset management fees, commissions, bid/ask spreads, taxes, etc. There costs can be material - bake them in as well - and total transaction costs often run 1% to 3% annually or more.
8. Techniques: seek advice from an independent financial advisor about techniques to deal with non-trending markets. The stock and bond bull markets of the 1980's and 1990's were trending markets and benefited from "buy-and-hold" and "buy-on-dips" strategies. The next phase could be another secular bear cycle (see http://www.crestmontresearch.com/pdfs/S ... 0Chart.pdf). Note the difference in secular bull and secular bear periods. Ask your advisor about the effect of more frequent rebalancing, covered call option writing, higher yielding preferred stocks and other securities with higher current return, TIPS, hedge funds, and other risk-controlled and value-added investment strategies.
Please consider reviewing our research at www.CrestmontResearch.com relating to stock market cycles, long-term returns, and interest rate cycles. For each of the charts with links listed above, you'll find a brief description within the Stock Market and Interest Rate sections. We welcome your insights and perspectives to further the research. Our analysis is not above challenge: I welcome it. It has not gotten to were it is today without challenge. All along, I've hoped for rational insights that would support strong upside potential in the financial markets. To date, the results instead have been rather sobering. Our research is not intended to generate predictions; rather it is intended to be provocative and enlightening. Hope it's helpful; there's no cost and it's free of banner ads. I welcome your comments, questions, and persepctives either in reponse to this posting or through our website.
Unlike Bill Murray in "˜Groundhog Day', we don't get another chance to replay our retirement scenario. The risk for a retiree or aspiring retiree of an error in their return assumptions is serious: you only find out that the assumptions were insufficient when it's too late to change the course.
Ed Easterling
Crestmont Research
(A similar version of this composition was first posted in the SWR Research Group board; it's been posted here at the suggestion of "peteyperson."