In the last two recessions share prices hit bottom months before unemployment peaked, and jobs took years longer than markets to come back.

The pattern from the last two recessions is consistent enough to plan around. Share prices fall first and recover first; jobs are lost later and come back last. In the United States the S&P 500 fell 57% from its October 2007 peak to its trough in March 2009, and unemployment, 5% in December 2007, kept climbing for seven months after that market low, to 10% in October 2009¹. Payroll employment did not get back to its January 2008 level until May 2014². A household needs two things from that history: enough cash to outlast a long job search, and a plan that does not depend on selling shares when they are cheapest.

How far unemployment rose, country by country

The two recessions were different animals. 2008–09 was a financial crisis that ground on for years; 2020 was a sudden shutdown, followed by an unusually fast rebound in most places. The table shows the rate before each, its peak, and the latest reading as of early October 2026. India is missing on purpose: its Periodic Labour Force Survey began only in 2017–18 and moved to monthly releases in 2025, so there is no comparable series covering 2008, and this article does not borrow one from elsewhere⁵.

Three things stand out. The euro area’s 2008 recession never really ended for its labour market: unemployment kept rising into 2013, through the sovereign-debt crisis that followed¹⁰. The 2020 spike was enormous in North America and muted in the UK and continental Europe, where furlough and short-time work schemes kept people on payrolls while they were idle⁶¹⁰. And the UK’s peak arrived in September to November 2011, long after output had started growing again. That slow tail is the part households underestimate⁶.

How long a job search takes, in good years and bad

The question people ask after a layoff is not the national rate but their own odds of being back at work in three or six months. The US Bureau of Labor Statistics publishes the answer every month. In an ordinary labour market the median spell of unemployment is around two to two and a half months; at the worst of the last crisis it was 25.2 weeks, close to six months, in June 2010⁴. The average was higher still, peaking at 40.7 weeks in July 2011, because a minority of very long searches pulls the mean up¹¹. In April 2010, 45.5% of unemployed Americans had been looking for more than 27 weeks¹².

Britain shows the same long tail today. In May to July 2026, of 1,778,000 unemployed people, 973,000 had been looking for up to six months, 342,000 for six to twelve months and 463,000 for more than a year¹³. That is about 26% unemployed for over a year and about 45% for over six months — in a labour market with no recession declared¹³. Long-term unemployment is not a recession-only risk; it is what happens to a large minority of people every year.

Markets fall first and recover first

Share prices are a guess about future profits, so they move on expectations; unemployment moves on what firms actually do months later. The table lines up the two US episodes. In both, the market low came before the recession ended and before unemployment peaked, and jobs were the last thing to recover¹¹⁴².

Measured from the 2007 peak, the market low came after 17 months, the recession ended at 20, unemployment peaked at 24 and jobs were regained only after 79. In 2020 the same sequence ran in fast-forward: the low after one month, unemployment’s peak after two, a new market high after six and jobs after 28¹¹⁵². On the OECD’s monthly-average US share index, the 2007–09 fall was 53% and the index did not regain its October 2007 level until January 2014¹⁶.

Markets also fall without recessions. The S&P 500 dropped 25.4% between 3 January and 12 October 2022, as interest rates rose, while US unemployment never went above 4.0% that year; the index did not close above its January 2022 level until 19 January 2024¹⁵³. And recovery has no timetable: Japan’s Nikkei 225 took more than 34 years to close above its 1989 peak¹⁷. The US record is the friendliest in the data, which is one reason to own more than one market (global diversification over 125 years).

Which jobs go first

Recessions are not evenly spread. Each one hits the industries closest to its cause, and the households working in them take most of the loss.

The technology layoffs of 2022 onwards are the clearest case of an industry downturn running without a general recession. Official US data show employment in the information sector falling from 3,115,000 in November 2022 to 2,739,000 in September 2026, a loss of 376,000 jobs, while total payrolls kept growing²¹². Layoffs.fyi, the tracker most often quoted in the press, has counted publicly reported layoffs at technology companies worldwide since 2020²³. It is useful as a running signal, but it counts announcements, not people out of work: some cuts are planned and never happen, some are overseas, and none of it nets off hiring. For the size of the damage, use the statistics offices.

Where the labour market stands in October 2026

The NBER, which dates American recessions, had not called one as of early October 2026; its last recorded downturn ended in April 2020¹⁴. But the direction has turned. US unemployment was 4.2% in September, after touching 4.5% in November 2025³, and payrolls grew by only about 29,000 in September on the first estimate². The median job search, 11.5 weeks, is longer than at any point in 2022 or 2023⁴. British unemployment was 4.9% in May to July 2026, against 3.6% in mid-2022⁶. Canada’s was 6.4% in August⁷, the euro area’s 6.4%¹⁰, and Australia’s rose 0.4 points over the year to 4.6% in August⁹ — before the Reserve Bank of Australia raised its cash rate to 4.60% at the end of September²⁴. Higher borrowing costs and softer hiring together are the conditions in which a household’s buffer matters most.

Four households, one question: how long could we last?

The figures below are illustrative households, chosen to look like readers rather than averages. The question each asks is the same: if one income stopped tomorrow, how many months could we pay our essential bills?

Dan and Lauren, 38 and 36, Columbus, Ohio. Together they take home $9,200 a month, $5,000 of it Dan’s. Their essential spending — mortgage, food, insurance, utilities, childcare, minimum debt payments — is $6,100, and they keep $24,000 in a high-yield savings account. That is 3.9 months of essentials, which sounds thin. But if Dan lost his job, Lauren’s $4,200 would still arrive, leaving a monthly gap of $1,900 that $24,000 covers for 12.6 months, longer than the average US search at its 2011 peak of 40.7 weeks¹¹. A second income is the strongest emergency fund most couples have.

Chris, 44, a Leeds engineering manager and his family’s only earner. Essentials are £3,400 a month and the family has £12,000 in easy-access savings, 3.5 months’ worth. With about 26% of Britain’s unemployed out of work for more than a year¹³, his target is six months — £20,400, so £8,400 more. Beyond that, he and his wife have written down which costs go first if a search runs past six months: the second car, the holiday fund, and the overpayments on the mortgage.

Pooja, 31, a product manager in Bengaluru. She takes home ₹2.1 lakh a month, spends ₹85,000 on essentials and keeps ₹4.5 lakh in a sweep-in deposit: 5.3 months. Technology is the sector where cuts have run longest, with US information-sector employment down 12.1% from its 2022 peak²¹, so she aims for nine months, ₹7.65 lakh, and closes the ₹3.15 lakh gap over the next year before raising her SIPs. Her provident fund is not part of the buffer: it is for later, and drawing on it early costs her future self.

Wen Jie and Hui Lin Lim, 52 and 50, Singapore. Both work; their portfolio is S$900,000, of which S$630,000 is in global equity funds. A fall like the 33.9% of February–March 2020¹⁵ would take about S$214,000 off the equities in a month. Their cash covers a year of essentials, so nothing has to be sold. Had they sold at the March 2020 low, they would have locked in that loss five months before the US index set a new high¹⁵. The cash buffer is what makes not selling possible.

Why selling in a downturn costs so much

The best measurement of what investors actually do in volatile markets is Morningstar’s annual Mind the Gap study, which compares a fund’s return with the return the average dollar invested in it earned, given when money went in and out. Over the ten years to the end of 2025 the average dollar in US mutual funds and ETFs earned 8.7% a year against the funds’ own 9.9%, a gap of about 1.2 points a year, around 12% of the return²⁵. The gap was about 0.4 points in the least volatile fifth of funds and more than 2 points in the most volatile fifth, the ones most likely to provoke a trade²⁵. Morningstar is careful to say that not all of the gap is bad behaviour; regular monthly investing opens one too. Its own reading of the volatility pattern is that big swings push investors into trading at the wrong moments.

Morgan Housel’s *The Psychology of Money* makes the point that matters for a recession: the goal is to be able to stay invested through the bad years, which means building room for error before they arrive²⁶. Jeremy Siegel’s long record in *Stocks for the Long Run* shows shares rewarding patience over decades, while warning that the short-run path is violent²⁷. Neither book says markets always recover quickly. Both say that a household forced to sell at the bottom turns a temporary fall into a permanent loss.

There is a real exception. Somebody already living off their portfolio cannot simply wait, because every withdrawal during a fall sells low (sequence-of-returns risk). The answer there is a runway of cash and short bonds for the first years of spending, so that shares are not sold into a crash — the idea behind the bucket strategy.

1. Work out essential monthly spending: what still has to be paid if an income stops.
2. Multiply by the months you need — about six for a single earner in a normal field, nine or more for one income in a cyclical industry or a household with dependants.
3. Hold that amount in deposits you can reach within days, in the currency you would spend, and mark it as separate from investments.
4. Write down now what you would cut, and in what order, if a search ran past the buffer.
5. Decide in advance that a market fall is not a reason to sell, and keep contributions going if your job is safe.
6. Refill the buffer before resuming extra investing once it has been used.

For the detail of sizing and holding the buffer, see how big an emergency fund should be; for the first weeks after a layoff, what to do after a job loss.

Testing your own plan against a bad year