Strong on this measure

How the score is calculated, where the data comes from and where it falls short. Nothing here changes the figures themselves.

1. We look at 12 measures of the economy, such as growth, unemployment, fish exports, inflation, energy prices and the króna.
2. Each measure gets points from −100 to +100. For example, 6% GDP growth gives +100, 2% gives 0 and −2% gives −100.
3. The most important measures count the most: growth 15%, the króna (trade-weighted) 12%, unemployment 10% and so on.
4. The points are multiplied by their weight and added up. The result is the score, and the breakdown shows what each one adds.
5. The score then falls into one of five bands, from ⛈️ Very weak (−100 to −50) to ☀️ Very strong (+50 to +100).

More about the method

What does the score measure? A blend of three things: activity (growth, jobs, tourism, fisheries), household conditions (unemployment, real wages, the loan example, energy prices) and price stability (inflation and the króna). It is not a pure business-cycle measure, and "strong" and "weak" describe the model's result, not an official assessment.

The zero point: each measure scores 0 at a chosen benchmark (such as 2% growth, 4.5% unemployment, 2.5% inflation). The benchmarks are not calibrated to the average of any period, so 0 does not mean "a normal year".

Missing data: when a measure is missing for a period it is left out and the others are scaled up in proportion. Forecast years use only the measures the forecast covers, and the first years lack tourism data. That is why the page shows data coverage, effective weights and comparisons on the same measures.

Why so much weight on jobs? The labour market counts for 26% in total: the unemployment rate (10%) shows how much slack there is, its change (8%) catches turning points early, since unemployment jumps at the start of every Icelandic downturn, and job growth (8%) shows whether the economy is adding work. The level and its change overlap: when both point the same way, the change counts gradually less the clearer the level is, down to half, so the labour market cannot decide the headline on its own. The discount grows smoothly, so a tiny move in unemployment never makes the score jump.

Inflation: the score uses CPI without housing, the warning light headline CPI. The central bank's 2.5% target refers to headline CPI; using the same benchmark for CPI without housing is the model's choice. Housing was left out to avoid counting rising house prices twice, since they already enter through house prices and the loan example. The overlap is only partial: the CPI housing component measures the cost of living in a home (including rent and imputed rent), while house prices measure asset prices. Without it, rising housing costs for tenants and owners drop out of the score's inflation measure.

"How does that year look with today's data?" is a different question from "What could we have known then?" The time machine answers the first. At the time, figures came out weeks or months later and were often revised afterwards, GDP growth sometimes by a percentage point or more, so a year may look clearer now than it did then.

The currency switch only changes amounts: króna figures are converted at that year's average rate (now: the latest ECB reference rate). The score, percentages and returns do not change.

Energy prices: the 12-month change in the price of electricity and heating (VIS01308, CP045) and fuel (CP0722), measured to the same month as the headline consumer price index (VIS01000). Each is adjusted for inflation, (1 + change) / (1 + inflation) − 1, and the two are weighted 46/54 by their CPI weights (3.07% and 3.61%). The Life in Iceland page shows both figures: the nominal change and the change relative to inflation.

Assumptions open to debate: a stronger króna gives positive points (more purchasing power, less imported inflation) even though it can squeeze exporters, and rising real house prices give positive points (stronger balance sheets, activity) even though they make first purchases harder. Never letting these give positive points was tested (calculated 30 Oct 2026): the score for the last 12 months would move from −25 to −26, while strong years fall, e.g. 2016 from +58 to +46 and 2005 from +38 to +26. The ranking of years changes little, so the model has not been changed; a steep rise in real house prices does turn on a warning light.

Assets: global stocks are the iShares MSCI World ETF (URTH) and US stocks the SPDR S&P 500 ETF (SPY), both with prices adjusted for dividends (adjusted close from Yahoo Finance); gold is PAX Gold (PAXG) on Kraken, with monthly data from October 2019, and LBMA annual average prices before 2020; Bitcoin is from Kraken. OMXI15 is a price index without dividends, shown from October 2016, and is not linked to earlier indices (ICEX-15, OMXI8, OMXI6).

"Now" compares the latest 12 months with the 12 months before. The 🔮 forecast years use Statistics Iceland's economic forecast. The warning lights are shown apart from the score because they flag risks the score does not capture, such as borrowing from abroad or a housing bubble. Bitcoin, gold, bonds, stocks, domains, catch and food prices are just for context and do not affect the score.

Kibba's words for changes follow fixed bands: under 0.5% is "about the same", 0.5–2% "a little", 2–5% "noticeably" and 5% or more "a lot", up or down alike.

The monthly view

The monthly scale scores every month from the 12 months up to it compared with the 12 months before. GDP and house prices are quarterly, so they use the latest quarter; fish exports use the latest full year. Króna strength uses the BIS trade-weighted index, which also covers 2009–2018 when the ECB did not quote the króna.