1989issue C071-5
When long-wave dominant cycles cannot be disproved
Archive accounts of a multi-decade prosperity long-wave show mixed inputs, disputed troughs, and stage-dating that can be moved after the fact. Editorial reading: a dominant-cycle check is out-of-sample only when the series, cycle-length, and trough date are locked first.
- Editorial test: lock the measured series, the cycle-length, and the trough date before looking ahead.
- The long-wave construct combines different variables and weights, so it is not a single-series dominant cycle.
- Out-of-sample-dating from locked 50-year, 54-year, and 60-year spans did not produce the implied later troughs.
- After 1979-82 failed to resemble 1929-32, stretched stage-dating left an unfalsifiable-scenario rather than a scientific theory.
A multi-decade prosperity long-wave
Long-wave advocates typically describe a prosperity cycle averaging 54 years, with a trough near the Depression, a peak in the early 1970s, and an expected severe downturn in the 1990s. In archive terms a long-wave is a hypothesized multi-decade swing in general prosperity assembled from several economic series rather than one fixed observable.
A dominant-cycle is a more-or-less regular swing from one extreme at least partway back that a quantitative baseline treats as the main repeating interval in an ordered series. The long-wave story is often read as if it supplied that interval. The archive record shows how the claim was assembled and then adjusted after later years.
Not one agreed series
The construct does not track one agreed series. Different writers combine different variables and weights, which already separates it from simple single-series cycles.
Only about three prior supposed turns are cited, the current trough date and expected length remain disputed, and successive highs and lows are described as trending higher.
Later statistical review of the original price data was reported as not supporting the long-wave reading. Remaining computer-based support was framed as a 45-to-60-year pattern in combined industrial activity rather than prices.
Subjective stage-dating
The stage chart used to date each wave is described as subjective. Endpoints are vague, many recessions are omitted, a wartime interval stands in for a recession, and an exact 54-year length is not shown. Stage-dating here means assigning named phases such as peak war, primary recession, plateau, and secondary recession to calendar intervals.
Out-of-sample-dating after a locked trough
Out-of-sample-dating adds a locked interval to a chosen historical trough and checks whether the implied later trough actually appears. Adding 54 years to a 1932 stock-price trough implied a major low in 1986, a date that instead fell inside a bull market. A 50-year variant pointed to 1982, while a 60-year span from 1941 implied a 2001 economic trough and, with a nine-year lead of stock lows, a 1992 market trough.
A 50-year overlay that expected a 1979 crash and a following depression did not produce that sequence. The period was later described as weak prices and a 1981-82 recession, after which the economy and market expanded.
When the path can still be fitted
After 1979-82 failed to resemble 1929-32, stage labels were stretched so that 1974-82 became the primary-recession analogue. Later non-events could still be absorbed by moving the start or declaring that the wave had already bottomed.
Because practically any later path can be fitted to one or another long-wave scenario, the framework is described as unfalsifiable and therefore not a scientific theory. That is an unfalsifiable-scenario: a narrative that can absorb any later path by moving the start date, stretching the period, or renaming stages.
Editorial test after a miss
Editorial reading: once the series, the cycle-length, and the trough date are locked, stretching the period, swapping inputs, or renaming stages does not preserve the same dominant-cycle model. It records that the out-of-sample check failed.
All readings on this track · 31 readings
- 1982Cycle phase windows for chart signal filters
- 1987Constructing a cycle-scaled trend oscillator
- 1987Constructing a dominant-cycle grid from marked lows
- 1988Cycle lead from staggered exponential averages
- 1988Auditing the forty-month stock-price cycle
- 1989When long-wave dominant cycles cannot be disproved
- 1991Half-cycle average plot shift versus cycle attenuation
- 1991Half-cycle average contact as an amplitude-ratio test
- 1993Building a restoring-pull indicator from cycle frequency and volume
- 1995Regime filters for a dominant long wave
- 1995A cycle-tuned lead filter from bounded oscillators
- 1998Testable cycle rules instead of fear and greed
- 1999Nested Euro cycle timing as one checkable procedure
- 2002Constructing an instantaneous trendline from a dominant cycle
- 2002Half-cycle center of gravity oscillator from moving-average balance
- 2004Testing a locked forty-week cycle with a hold-or-sit-out rule
- 2005Nested timing bands for dominant-cycle confirmation
- 2005Dominant-cycle baselines versus policy-news narratives
- 2006Pairing a dominant-cycle horizon with trend and oscillators
- 2006A dominant-cycle split into a trend filter and residual Relative Strength Index
- 2007Construct a momentum difference from the dominant cycle
- 2007Naive dominant-cycle rules fail without crowd tests
- 2012Constructing a dominant-cycle forecast as a timing window
- 2012Open-parameter construction of dominant-cycle baselines
- 2013Using a second-term election to check a predeclared dominant-cycle forecast
- 2014Constructing a dominant-cycle forecast baseline
- 2014Quotient transform as an early-onset trend filter
- 2014Construct a trough-to-trough cycle map with the Detrended Price Oscillator
- 2015Dominant-cycle alignment before an earnings catalyst
- 2017Causal reverse exponential average for cycle and trend
- 2020Constructing a cycle-plus-trend oscillator from a one-wavelength chord