Recession Odds Explained: How to Read Them and Test Your Calls

Recession odds are the probability that an economy enters a recession within a stated window, as estimated by prediction markets, economists or banks. In 2026 the figures have ranged widely, from low single digits to around 30 per cent, and that is not a contradiction. Each number answers a slightly different question, at a different date, under different rules. Read the definition and the deadline first, and the number starts to make sense.

This guide explains where recession odds come from, how the 2026 picture has moved, what research says about forecast accuracy, and how to make and score your own economic calls on Vatic.

What recession odds actually measure

In a prediction market, the price of a contract is read as a probability. A share trading at 0.28 implies a 28 per cent chance of the event, as one analysis of 2026 recession markets put it in plain terms. Source

The catch is the contract wording. Recession markets usually settle on official data, and the two main US contracts do not word it identically. Polymarket's contract on a US recession by the end of 2026 resolves yes if two consecutive quarters of negative GDP growth appear between Q2 2025 and Q4 2026, or if the NBER formally calls a recession in that window. Source Kalshi's version pays on two straight quarters of negative GDP growth. Source

Two contracts with similar headlines can therefore price differently, and a price also reflects fees and market rules. Source

Where the numbers come from

There are three main sources of recession odds, and they measure different things.

Prediction markets

Markets aggregate the views of people who have chosen to trade. The price moves as new information arrives, such as a jobs report or a GDP release. Forbes reported that Polymarket's year-end recession contract sat at 30 per cent in mid-April 2026, with a little over $1.3 million traded on it since September 2025. Source

Economist surveys

Surveys ask professionals for their own probability. Bankrate's first-quarter 2026 survey was conducted between 23 March and 1 April and put the chance of a recession in the next year at 34 per cent, up from 28 per cent at the end of 2025. Source

The Wall Street Journal's October 2026 survey was published on 11 October and collected answers from 71 academic, financial and business economists between 2 and 7 October. The average recession probability was 24 per cent, but individual answers ranged from 1 per cent to 80 per cent. Source That spread is a useful reminder that an average can hide a lot of disagreement.

Banks and research desks

Banks publish their own outlooks. RSM US, for example, lowered its 12-month recession probability to 30 per cent from 40 per cent in its 2026 outlook. Source Forecasts like these are updated when new data arrives, so check the publication date before comparing them.

How the 2026 picture moved

The 2026 story is a good example of why dates matter. Prediction market odds rose in the spring and then fell sharply.

• Early 2026: CryptoBriefing reported that Polymarket odds had been around 30 to 35 per cent in spring 2026. Source
• March spike: Motley Fool reported Polymarket at 37 per cent on 8 March, when oil was above $100 a barrel, compared with 21 per cent just before the war began on 27 February. Source
• Summer and late August: The same CryptoBriefing piece says the odds had dropped to single digits by late August 2026, with unemployment holding in a 4.1 to 4.3 per cent range. Source A July snapshot from another tracker showed a 10 per cent Yes price on the same question. Source

Notice what this shows. The same question moved by roughly 25 percentage points in a few months, and the markets were reacting to oil, labour data and rate expectations. Any single number you see online is a snapshot, so always ask when it was taken.

How good are these forecasts?

The research is encouraging, but it is not a blank cheque.

• Elections: A study of Iowa Electronic Markets compared market prices with 964 polls. Election-eve forecasts had average absolute errors of 1.33 percentage points, and markets significantly outperformed polls when forecasting more than 100 days ahead. Source
• Sports: Across thousands of American football games, betting markets had only a tiny edge over opinion polls, and very simple statistical models came close to the markets. Source That is a lesson for fight and football calls too: the crowd is good, but not magic.
• Expert crowds: A 2023 study found that small, elite crowds produced more accurate aggregate forecasts than non-elite crowds, and that prediction markets and prediction polls were roughly tied. Source
• Corporate tests: One MIT study of a company's internal markets found that short-term prediction markets were no more accurate than conventional forecasting methods. Source

Accuracy also degrades the further away the outcome sits. Source A recession call for year-end is harder than a call on a fight tonight, and a thin market can move on very little money. One 2026 snapshot showed about $1.65 million in total volume but only about $25,000 in active liquidity. Source

How to read a recession call

Use this checklist before you trust any figure:

• Definition: Does it mean two negative quarters of GDP, an official NBER call, or a survey question about the next 12 months?
• Deadline: Is the window the rest of 2026, the next year, or something else?
• Date: When was the number taken? A figure from spring can be out of date by autumn.
• Source type: Is it a market price, a survey average or a single bank's view?
• Liquidity: Is the market deep enough that one trader cannot move it?

Remember too that a 30 per cent chance is not a forecast that nothing will happen. Even a 34 per cent survey figure still leaves the odds in favour of no recession, as Bankrate's own commentary noted. Source Low-probability events happen, and high-probability ones fail. Judge a forecaster over many calls, not one.

Practise your own calls on Vatic

You do not need a trading account to practise this skill. Vatic is free to play, forever. Vpoints have no cash value, and they can't be bought, sold or cashed out. You make your call on Vatic markets, and correct calls earn Vpoints and VXP toward levels, ranks and the leaderboard.

Pick questions with clear deadlines. Two examples from Vatic's live markets show how the discipline carries over:

• Fights: Check the event date and the exact result you are predicting before you make a call on who will win at UFC 335, Joe Pyfer or Bo Nickal. The same logic applies to Alex Pereira or Sergei Pavlovich.
• Crypto: Short-horizon calls are the fastest way to check your reading of the crowd. Ranked mode runs 5 and 15 minute Up/Down rounds on BTC, ETH, SOL and XRP, priced from the live Coinbase feed. Recession calls take months, but a Ranked round tells you within minutes whether you were right.

If you want to compare notes, you can start a league with friends and see who reads deadlines most carefully.

This article is not financial advice. It describes what markets, surveys and data showed at the dates given and does not recommend buying, selling or holding any asset.

FAQ

What are recession odds?

Recession odds are the estimated probability that an economy enters a recession within a set window. Prediction markets express them as contract prices, while economists give them as survey or model probabilities.

Why do recession odds differ so much between sources?

Each source uses a different definition, deadline and method, and the figures are taken on different dates. A market price can also shift quickly with news, while a survey is a single moment.

Is a 30 per cent recession chance high?

It means the event is plausible but still less likely than not. A 30 per cent outcome happens often, so one result does not prove a forecast right or wrong.

Can I test my own recession reading for free?

You can make calls on economic-style questions on Vatic for free and track your results. Vpoints have no cash value, and Vatic is 18+ and not available in Spain.

Researched and written with AI assistance by the Vatic Desk, from the sources listed above. Vatic is a free prediction game for adults (18+). Vpoints have no cash value and can't be bought or cashed out. Nothing here is financial or betting advice.