What the Invest index measures

What the index is for

The Invest index answers one question: are markets under more stress than usual? It does not predict prices. It reports where today sits against the recent past, on a scale where 50 is an ordinary day, 0 is the calmest and 100 the most strained. How to read a 0–100 reading explains the scale itself.

It is built from seven series. Four come from the United States, because that is where the most widely used, free and long-running public market data is published, and US conditions tend to reach European savings and loans. Three describe the euro’s exchange rates and the price of bitcoin in euros.

• Financial stress (St. Louis Fed index): 25% of the weight.
• US stock drawdown (SPY): 15%.
• Euro against the dollar, size of daily moves: 15%.
• Bitcoin in euros, size of daily moves: 15%.
• Oil price swings (WTI crude): 10%.
• Yield curve (US 10-year minus 2-year): 10%.
• Euro against six other European currencies, size of daily moves: 10%.

How a series becomes a score

Each series is first turned into a number where higher means more unusual pressure. The site uses four such transformations for this index:

• Level: the value as published (financial stress).
• Inverse: the value with its sign flipped, so a lower yield-curve spread gives a higher number.
• Volatility: how large the daily moves were over the last 30 prices. Technically, the standard deviation of the last 30 daily log returns times 100, roughly a typical day’s move in per cent.
• Drawdown: how far, in per cent, the latest price is below the highest price of the last 90 observations.

The result is then ranked against up to 365 earlier observations of the same series, and the rank becomes the 0 to 100 score. A series needs 20 earlier observations before it gets a score at all. The index is the weighted average of the seven scores. The full method is on How it works.

Because the ranking counts observations, not calendar days, ‘the past year’ is approximate. Markets do not trade at weekends, so 365 trading days is about 17 months. A weekly series covers more than two years. Each section below gives the real figure.

Financial stress: the St. Louis Fed index

The Federal Reserve Bank of St. Louis publishes a weekly Financial Stress Index, now in its fourth version (STLFSI4). It combines 18 weekly market series: seven interest rates, six yield spreads and five other indicators^[1]. It is designed so that zero is an average level of stress; values above zero mean more stress than average and values below zero mean less^[1][2]. The index was created in 2010, and version 4, introduced in November 2022, changed how two of its spreads are measured, using a forward-looking 90-day SOFR rate in place of the earlier backward-looking one^[2].

Why it matters to an ordinary person: the index tracks borrowing costs and the gaps between safe and less safe lending. When those gaps widen, banks and companies pay more to borrow and lend more cautiously. Much of the world’s finance is priced in US markets, so strain there can feed through to European banks, and from there to the rates on new loans and mortgages. Many home loans in the Baltic states have a floating rate linked to Euribor, which responds to similar conditions on the euro side.

How the site uses it: the value is scored as it is (the ‘level’ transformation) and carries 25% of the index, the largest single share, because it is already a summary of many markets built by a central bank. The site fetches about 800 days of weekly readings, roughly 114 weeks, so each new week is ranked against about two years of earlier weeks rather than one.

Known weaknesses: it is weekly, and each reading covers the week ending on a Friday^[1], published several days later. The site treats any indicator older than 10 days as out of date, so for a day or two before each weekly release this indicator can drop out and the index runs on the other six. It needs a free FRED API key; without one it shows as ‘not connected’.

The yield curve

The series used is the yield on 10-year US Treasury bonds minus the yield on 2-year Treasury notes, published daily in per cent^[3]. Normally longer loans pay more, so the figure is positive. When it falls below zero the curve is said to be ‘inverted’: investors accept less for lending for ten years than for two, which usually means they expect rates, and growth, to fall.

Why it matters: an inverted curve has a long record as an early signal of US recessions. A 1996 study by Arturo Estrella and Frederic Mishkin at the Federal Reserve Bank of New York found that the spread between 10-year bonds and 3-month bills did better than other financial and economic indicators at predicting recessions two to six quarters ahead^[4]. The New York Fed still publishes a model that turns the slope of the curve into a probability of a US recession twelve months ahead^[5].

How the site uses it: the spread is scored with its sign flipped (‘inverse’), so a flatter or more inverted curve scores higher. It carries 10%. About 800 days of daily data are fetched, so each day is ranked against about 17 months of trading days.

Known weaknesses: it is a signal, not a forecast. The lead time between inversion and recession has varied, and the research above uses the 10-year minus 3-month spread, not the 10-year minus 2-year spread used here. The score is a rank against the recent past, so a high score means the curve is flatter than it has been lately, not necessarily that it is inverted. It describes US rates only; there is no euro-area curve in the index.

Oil price swings

The series is the daily spot price of West Texas Intermediate crude oil at Cushing, Oklahoma, in dollars per barrel, published by the US Energy Information Administration and republished by FRED^[6]. WTI is one of the two main oil benchmarks; Europe’s crude is mostly priced off the other, Brent, but the two usually move together.

Why it matters: oil feeds into petrol and diesel prices, heating, transport and the cost of almost everything that is shipped. Sharp swings in either direction make household and business budgets harder to plan.

How the site uses it: the ‘volatility’ transformation, so it measures how large the daily moves were, not whether the price is high or low. A sharp rise counts the same as a sharp fall. The last 30 prices are about six weeks of trading days. It carries 10%, and each day is ranked against about 17 months of trading days.

Known weaknesses: FRED’s copy of this series is refreshed about once a week, so the latest price can be several days old^[6]. It is a US benchmark. It says nothing about natural gas or electricity prices, which matter more to many European energy bills.

US stocks below their recent high

The series is the daily closing price of SPY, an exchange-traded fund that tracks the S&P 500 index of large US companies. The site gets it from Alpha Vantage, whose daily series is raw, as-traded prices, not adjusted for dividends^[7].

Why it matters: US shares are a large part of the global index funds held in many European pension and savings products. A fall in US stocks is one of the most direct ways market stress reaches ordinary savers.

How the site uses it: the ‘drawdown’ transformation, the percentage by which the latest close is below the highest close of the last 90 trading days. It carries 15%.

Known weaknesses: the history is short. Alpha Vantage’s free tier returns the latest 100 data points; the full history needs a paid key^[7]. That means about 100 trading days, so each day is ranked against roughly four to five months, not a year, and early in that window the 90-day high is based on fewer days. Because the prices are not adjusted, the small drop when the fund pays a dividend shows up as a small drawdown. It needs an Alpha Vantage API key.

The euro against the dollar and its neighbours

Both currency indicators use the European Central Bank’s euro reference rates. The ECB sets them through a daily procedure among central banks, aiming to reflect the market at about 14:15 CET, and publishes them around 16:00 CET on each working day except TARGET closing days^[8]. The ECB states that they are for information only and should not be used for transactions^[8]. That is exactly how this site uses them. The site reads them through Frankfurter, a free service that republishes ECB rates.

Euro swings looks at the euro against the US dollar. Why it matters: oil, many raw materials and much electronics are priced in dollars, so a large move in the euro changes what people in the euro area pay for them, and changes the euro value of any savings held in US investments.

European currency swings looks at the euro against six other European currencies: the British pound, Polish złoty, Swedish krona, Swiss franc, Czech koruna and Norwegian krone. They include Poland and Sweden, neighbours across the Baltic Sea. Why it matters: these rates affect cross-border shopping, work and travel, and the prices of goods traded within Europe.

How the site uses them: both use the ‘volatility’ transformation over the last 30 published rates, about six weeks. For the basket, each currency’s rate is divided by its value on the first day of the data the site fetches, and the six results are averaged, so every currency counts the same however it is quoted. Euro swings carries 15% and European currency swings 10%. About 800 days of rates are fetched, so each day is ranked against about 17 months of working days.

Known weaknesses: there is one rate per working day, so moves during the rest of the day are not seen. Averaging six currencies smooths the basket: if two move strongly in opposite directions, they partly cancel. Like the other swing measures, it treats a rise and a fall alike.

Bitcoin swings

The series is the daily bitcoin price in euros from CoinGecko. CoinGecko’s price is a volume-weighted average across many exchanges and trading pairs, with prices far from the rest removed as outliers before averaging^[9].

Why it matters: bitcoin trades every day of the year and reacts quickly to changes in investors’ appetite for risk. It is also held directly by some households.

How the site uses it: the ‘volatility’ transformation over the last 30 daily prices, which for bitcoin is 30 calendar days. It carries 15%. CoinGecko’s free service returns one year of daily prices, so after the first 30 days are used to start the calculation, each day is ranked against about 11 months.

Known weaknesses: bitcoin moves much more than the other series even on an ordinary day. A score of 50 here still means daily moves larger than the other series usually show; the score only says whether this month is calmer or rougher than bitcoin’s own recent past. It measures one asset, not the economy.

Why these weights

The weights add up to 100%. The St. Louis Fed index has the largest share because it already summarises 18 market measures. Price swings (oil, the two currency indicators and bitcoin) together carry half the weight, because they react within days and can be measured from free daily data. The stock drawdown and the yield curve add the two most widely watched signals of US market strain.

The currency and bitcoin indicators need no API key, and together they carry 40%, which is the minimum share of weight the site needs before it reports a full reading. If the FRED and Alpha Vantage keys were both missing, the index would be built from currency and bitcoin swings alone. When indicators are missing, the page says ‘not connected’ next to them and the reliability figure falls; the sources page lists what each one needs.

What it leaves out

• Euro interest rates: there is no Euribor, ECB policy rate or euro-area bond spread in the index.
• Inflation, wages, unemployment and house prices, which change slowly and are published monthly or quarterly.
• European and Baltic stock markets. Only US large companies are covered.
• Gas and electricity prices, which drive many European energy bills more directly than crude oil.
• Direction. Most indicators measure the size of moves, not whether prices rose or fell.
• Your own situation. A Quiet reading does not make any investment safe, and a High reading does not mean anyone will lose money.

If a reading is high and you are unsure what, if anything, to do, What to do when a reading is high covers it. For decisions about your money, speak to a regulated adviser.