Summary
- Each bar shows one market decline episode from its pre-crash peak through recovery. The red segment is the decline phase (peak to trough); the teal segment is the recovery phase (trough back to the prior peak level).
- The chart defaults to real (CPI-adjusted) returns ranked by longest recovery. The 1973 episode is the starkest example: nominal recovery from the January 1973 peak took roughly 3.5 years, but in real terms the underwater period stretched approximately 12 years.
- In real terms, the 2000 dot-com crash and the 2007 Subprime Crisis appear as one continuous 12-year-9-month episode. Switching to nominal returns separates them into two distinct rows.
- For an investor drawing income during a bear market, the length of the underwater period is more consequential than the eventual recovery date.
What these metrics measure
Each row in the chart represents one decline episode, defined by three points in time:
Peak is the all-time high of the cumulative total return index just before the decline began. Each episode starts here. The row label shows the peak month and year.
Trough is the lowest point reached before the market turned back up. The red bar covers the period from the peak to the trough. The drawdown percentage shown below the trough dot is the total decline from peak to trough in the selected return series (nominal or real).
Recovery is the first month the index returned to or exceeded its prior peak level. The teal bar covers the period from the trough to that recovery point. The total time shown below the recovery date covers peak to recovery (the full underwater period).
This peak-to-trough methodology means episodes are defined by the actual market cycle, not the calendar year. A decline that began in late 2007 is labeled October 2007, not 2008, even though the steepest losses occurred in 2008. It also means two separate nominal crashes can merge into one episode in real terms if the first crash never fully recovered before the second began.
What the recovery data shows
The chart offers three views, selectable at the top:
- Longest recoveries (default): episodes ranked by how long the underwater period lasted, from longest to shortest. This view highlights the cases most consequential for a withdrawing retiree.
- Deepest drawdowns: episodes ranked by peak-to-trough decline magnitude. This shows how severe each decline was at its worst point.
- Chronological: all episodes in reverse time order, useful for tracing the full historical sequence.
The Nominal / Real toggle switches between two frames of reference. Nominal measures total return including dividends reinvested, without adjusting for purchasing power. Real adjusts using the CPI series from Robert Shiller's dataset, showing what an investor's purchasing power actually experienced. Because the two series can diverge substantially over multi-year periods, the same decline can look very different depending on which frame is used.
The drop threshold controls how large a peak-to-trough decline must be to appear. The default of 10% includes moderate pullbacks and shows the full breadth of historical episodes. Raising the threshold to 20% or more filters to the major bear markets most commonly discussed in the financial press.
Data: Robert J. Shiller, shillerdata.com. Total return series including dividends reinvested; real returns CPI-adjusted using Shiller's CPI series. Past performance does not indicate future results.
The 2000s: when two crashes become one
In nominal terms, the 2000s produced two distinct bear markets separated by a partial recovery. In real purchasing-power terms, they were one unbroken episode spanning nearly 13 years.
| Dot-com (nominal) | Subprime (nominal) | Combined (real) | |
|---|---|---|---|
| Peak | Aug 2000 | Oct 2007 | Aug 2000 |
| Trough | Feb 2003 | Mar 2009 | Mar 2009 |
| Drawdown | -41.6% | -49.0% | -51.8% |
| Recovery | Oct 2006 | Aug 2012 | May 2013 |
| Underwater | 6 yr 2 mo | 4 yr 10 mo | 12 yr 9 mo |
In real terms these two rows collapse into one. By October 2007, the inflation-adjusted index had not yet recovered its August 2000 real peak. The S&P 500 was still underwater from the dot-com crash when the financial crisis hit. The peak-to-trough algorithm treats the entire period as a single episode: August 2000 peak, March 2009 trough (-51.8% real), May 2013 recovery — 12 years and 9 months underwater.
Switching the chart to nominal returns separates them into two distinct rows. Staying on real returns shows what an investor measuring their experience in purchasing-power terms actually lived through.
The 1973 bear market: when inflation extended a recovery by nearly a decade
The January 1973 episode is the starkest illustration of how inflation can quietly extend the real underwater period far beyond what the nominal chart shows.
| Nominal | Real (CPI-adj.) | |
|---|---|---|
| Peak | Jan 1973 | Jan 1973 |
| Trough | Dec 1974 | Dec 1974 |
| Drawdown | -39.2% | deeper |
| Recovery | Jul 1976 | Jan 1985 |
| Underwater period | ~3.5 years | ~12 years |
On a nominal basis the episode looks manageable: a sharp decline followed by recovery within 3.5 years. In real terms the picture is starkly different. Consumer prices rose roughly 9-12% per year through the mid and late 1970s. Each year of nominal price recovery was partially eroded by inflation, meaning the real purchasing power of the index remained below its January 1973 level long after the nominal chart showed "recovered."
Switching the chart to real returns makes this visible: the teal bar stretches almost a decade longer. The gap between the two bars represents purchasing power quietly consumed by inflation during a period when the nominal recovery appeared complete.
What the underwater period means for retirement planning
The concept of the underwater period is most consequential for investors who are drawing income from a portfolio during the bear market and its aftermath. An accumulating investor who stays the course during a multi-year underwater period experiences the drawdown as a paper loss but does not crystallize it through sales. A withdrawing investor is selling assets at depressed prices to fund living expenses, which reduces the portfolio's ability to participate fully in the eventual recovery.
The difference between a three-year and a ten-year underwater period is not just an emotional one: it is a compounding disadvantage that accumulates with each year of withdrawals from a reduced base. A portfolio that begins a 30-year retirement in the year before a major decline faces materially different long-run outcomes depending on whether the underwater period lasts two years or eight.
This is why retirement income planning does not simply rely on the observation that markets eventually recover. Coordinating withdrawal sources, maintaining lower-volatility reserves, and structuring Roth conversions and tax-efficient withdrawals to reduce forced sales during depressed markets are strategies designed precisely to manage the underwater period problem, not just the eventual recovery.
The real return view of this chart adds another dimension: in high-inflation periods, the real underwater period can be substantially longer than the nominal one. A retirement plan built around nominal recovery milestones may underestimate how long purchasing power is actually impaired.