The Economic Atlas
Every source, every update cadence, every conversion, and every known limitation, in one place, rather than scattered across 62 individual pages. For the series behind every indicator of each economy, see Methodology by economy.
This site deliberately shows official statistics with no interpretation, no forecasts, and no AI-generated commentary on what the numbers mean. That's not an oversight, it's the founding constraint the whole site is built around. If you want someone's opinion on what a number means, that's everywhere else on the internet; this is built to be the place that just shows you the number, sourced precisely, and lets you draw your own conclusion.
Every figure on this site traces back to an official national statistics agency or a recognised international body, never a news article, a blog, or a secondary aggregator. The primary sources in use across the site: national statistical offices (including ONS, MoSPI and Japan's Statistics Bureau, wherever a country's own office publishes a series we can fetch directly), FRED (Federal Reserve Bank of St. Louis), OECD, World Bank, Eurostat, IMF (IFS and WEO), BIS, UN Comtrade, the Bank of England, and the European Central Bank. Every individual metric on every country page has its own ⓘ info button showing the exact agency and series code behind that specific number, this page describes the system as a whole; the per-metric ⓘ is the precise citation for any single figure.
Every indicator page states whether its series is seasonally adjusted, not seasonally adjusted, or not applicable, shown under "Seasonal adjustment" in the "How this series is defined" panel. This is determined, never inferred from the numbers themselves, in a fixed order: first, whether the source citation states it in words; then, whether the series is an interest rate or bond yield, which is not applicable, as these are point-in-time prices rather than flows with a season to remove; then, whether it's annual data, also not applicable, since there's no within-year pattern to strip out; then, whether the series identifier itself states it, for publisher families with a documented, publicly stated naming convention (the IMF's International Financial Statistics, the OECD's Main Economic Indicators, and Eurostat's national accounts all encode this in their own series codes); and finally, for anything the above doesn't settle, checked directly against that specific series' own publisher documentation or the site's own fetch query for it. Where none of this settles it, the page says so plainly rather than guessing.
Most series are checked throughout the day by an automated pipeline that queries each source directly. Trade-partner data (the country-to-country trade breakdowns) is checked once a day, since the underlying UN Comtrade figures change only occasionally. Most official statistics are published monthly, quarterly or annually, so frequent checking does not mean the numbers change often: it means a new official reading appears here soon after its source publishes it.
Each series shows a small freshness indicator (green or orange) next to it. Green means the series is within its normal publication window for that source and frequency. Orange means it's gone significantly past that window, which usually means the source hasn't published its next release yet, but occasionally means the automated fetch for that specific series needs attention. Either way, orange is a genuine signal, not decoration.
The Dollarise toggle converts a country's own-currency figures into US dollars. Critically, this does not take today's exchange rate and apply it retroactively across history, each historical data point is converted at its own contemporaneous exchange rate. For most currencies that rate history is monthly. For a handful (Argentina, Chile, Colombia, Indonesia, Israel, Morocco, Poland and Turkey) only an annual rate history is available, so each point is converted at its own year's rate rather than its own month's. Which resolution applies is stated in the info panel for every metric on those countries' pages. A reading from 2015 is converted at 2015's rate, not today's. Where a currency's exchange-rate history doesn't extend back as far as the underlying economic data (for example, before a currency existed in its current form), the earliest available rate is used for anything before that point, and this is disclosed on the relevant chart rather than silently applied. Where a currency had no market rate for part of its history (Vietnam before 1989, Egypt before 1991, Nigeria before 1999, when official rates were fixed), earlier figures are left out of the dollar view altogether, and the note under the toggles says from which year dollar figures start. Where a currency was replaced by a new one (Chile in 1975, Israel in 1985, Argentina in 1992, Poland in 1995), the dollar view starts with the current currency. Turkey's 2005 redenomination is the exception: its earlier figures are already restated in new lira, so its dollar view runs from 1995.
Dollarise uses market exchange rates, so a figure converted into dollars moves for three reasons: output changed, prices changed, or the currency moved against the dollar. That makes it the right measure of what an economy is worth in dollars, and the wrong one for comparing what people in different countries can actually buy. For international comparisons of real output or living standards, use a purchasing power parity (PPP) measure, which the Atlas does not currently carry.
The Make It Real toggle adjusts nominal figures for inflation, so you're looking at real (inflation-adjusted) terms rather than raw current-price numbers. For countries experiencing very high inflation (Turkey and Argentina, currently), combining Make It Real with Dollarise at the same time can produce a result that's technically calculated correctly but economically difficult to interpret meaningfully, both pages carry an explicit warning when both toggles are active together for exactly this reason. When both toggles are on, real GDP is converted at a single exchange rate, the average rate of its latest year, rather than at each year's own rate. Real GDP is measured in constant local prices, so converting it at a rate that moves year to year would mix a currency's fall into what should be a measure of output: a devaluation would look like the economy shrinking. Using one rate keeps real growth in dollars identical to the published real growth. The World Bank publishes its own real GDP in constant 2015 dollars (NY.GDP.MKTP.KD) for every economy here except Taiwan. The site does not use it for the chart, because it is built from an older vintage of each country's national accounts and, for some countries, on a fiscal rather than calendar year, so its growth can differ from the national statistics office's own figure shown on the same page. It is used instead as an independent check: any recent year where the two disagree by more than 2 percentage points is investigated.
This site doesn't claim complete coverage for every country on every metric. Where a genuine gap exists, a metric with no reliable live source yet, or a country still on a fallback data source rather than a directly-fetched one, that's disclosed on the page itself (as a "known gap" note or an illustrative-fallback label), not hidden. Two specific, current examples: a few countries' inflation figures fall back to an annual series from the World Bank or the IMF when the monthly source is unavailable, which is why the period under those figures shows a year rather than a month; and any metric using a derived calculation (for example, a trade balance computed as exports minus imports rather than published directly as a single series) is labelled as derived rather than presented as if it came from one direct source.
China's statistics bureau and central bank cannot be queried from outside China, so every China series comes from an international body that republishes the official figures: GDP, growth and inflation from the OECD, the interest rate from the Bank for International Settlements, goods trade from the IMF, and government finances, unemployment, the current account and foreign investment from the IMF and the World Bank. Three consequences are disclosed on the page. The interest rate shown is the one-year loan prime rate, the benchmark lending rate, not the seven-day reverse repo rate the People's Bank of China has used for its operations since 2024. Before August 2019 that series is the one-year benchmark lending rate, which the loan prime rate replaced. Unemployment is the International Labour Organization's modelled annual estimate, not the bureau's monthly surveyed urban rate. Month-on-month inflation is not shown, because no reachable source publishes the bureau's own monthly rate and a rate rebuilt from a rounded index does not match it.
Malaysia's figures come from the publishers themselves wherever they publish openly: GDP, inflation, the labour force survey and goods trade from the Department of Statistics Malaysia's open data service, and the overnight policy rate from the Bank for International Settlements' copy of Bank Negara Malaysia's series, checked against Bank Negara's own. The policy rate is shown from May 2004, when the overnight policy rate was introduced, rather than joined to the earlier rates the series carries. The monthly labour series start in January 2011, after a re-basing of the survey. Quarter-on-quarter GDP growth is the seasonally adjusted figure DOSM headlines. The government balance is the IMF's general government measure, which differs from the federal government deficit the Ministry of Finance publishes.
New Zealand's consumers price index and labour force survey are quarterly, so its inflation and unemployment figures update four times a year rather than monthly; the inflation rates are worked out to one decimal place from Stats NZ's all-items index as republished by the OECD, and checked against the Bank for International Settlements' independent copy. GDP, the 10-year yield and business confidence also come through the OECD. The official cash rate is the Bank for International Settlements' copy of the Reserve Bank's series, from April 1999, with its daily series used for the months its monthly series has not reached yet. Goods trade is the IMF's, converted from US dollars at each month's rate. Annual GDP growth compares a quarter with the same quarter a year earlier, the measure used for every country here; Stats NZ also publishes a twelve-month total comparison, which moves more slowly.
Belgium uses the euro and has no interest rate of its own, so the rate shown is the European Central Bank's deposit facility rate, the same series as every euro area page. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from Statbel's national CPI used for wage indexation; Belgium's HICP swings sharply around the January and July sales, and its jump from 1.4% to 4.7% in August 2021 was temporary, because the 2020 sales had moved from July to August. Goods trade is Eurostat's total with the rest of the world: Belgium is a transit hub, so goods passing through the port of Antwerp-Bruges count in both directions, and figures on other concepts, such as the IMF's, can differ by a few billion a month. Debt and the deficit are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Portugal uses the euro and has no interest rate of its own, so the rate shown is the European Central Bank's deposit facility rate, the same series as every euro area page. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from INE's national CPI. The OECD's employment and participation rates both step down together in the first quarter of 2009 while the unemployment rate they imply moves normally, the sign of a break in the population figures behind them rather than a fall in jobs, so they are shown from 2009; the unemployment rate itself, from Eurostat, keeps its full history. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's; 2025 was a small surplus.
Finland uses the euro and has no interest rate of its own, so the rate shown is the European Central Bank's deposit facility rate, the same series as every euro area page. Inflation is the EU's harmonised index (HICP); Statistics Finland's national CPI usually differs by a few tenths because it also counts owner-occupied housing and interest on consumer loans. Unemployment is Eurostat's seasonally adjusted rate, which can differ by several tenths from the unadjusted monthly figure or the trend that Finnish news usually quotes. GDP growth shows the latest revised figure; Statistics Finland's flash estimates are often revised. The 10-year bond yield starts in March 1990 because the source series has no value for February 1990. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Greece adopted the euro in 2001, so the European Central Bank's deposit facility rate is shown from then; before 2001 the Bank of Greece set rates. Inflation is the EU's harmonised index (HICP), whose weights include spending in Greece by visitors; ELSTAT's national CPI covers resident households only and can differ by a few tenths. The 10-year bond yield has no value for July 2015, when Greek markets were closed during the bank holiday and capital controls, and the chart joins June directly to August and does not estimate July. The labour force survey changed method in 2021 under a new EU regulation. Eurostat publishes Greek data under the code EL rather than GR. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Czechia is an EU member outside the euro area, so every currency figure is in koruna. Two inflation measures are shown: the Czech Statistical Office's national CPI, the headline and the measure the Czech National Bank targets at 2%, which includes owner-occupied housing costs; and the EU's harmonised index (HICP), which leaves them out, for comparison with other EU countries. The interest rate is the CNB's two-week repo rate from the Bank for International Settlements, shown from January 1998, when the CNB adopted inflation targeting, with BIS's daily series used for the months its monthly series has not reached yet. Goods trade is Eurostat's, converted from euros to koruna at each month's European Central Bank reference rate. The 10-year yield is the ECB's long-term interest rate for convergence purposes. Unemployment is Eurostat's survey-based rate, not the Labour Office's registered unemployment share that Czech news often quotes. Exchange rates are World Bank annual averages, as for Poland, since the Federal Reserve publishes no daily koruna rate.
Hungary is an EU member outside the euro area, so every currency figure is in forint. Two inflation measures are shown: the Hungarian Central Statistical Office's national CPI, the headline and the measure the Magyar Nemzeti Bank targets at 3%, computed from the OECD's copy of KSH's index (it can differ by 0.1 from the rate KSH publishes), and the EU's harmonised index (HICP) for comparison with other EU countries. The interest rate is the MNB's base rate from the Bank for International Settlements, from mid-2001, when the MNB adopted inflation targeting; the instrument behind the base rate has changed several times since, and in 2022 and 2023 the MNB steered short-term rates with an 18% one-day deposit tender while the base rate stood at 13%. Goods trade is Eurostat's, converted from euros to forint at each month's European Central Bank reference rate. Foreign direct investment is not shown: the World Bank's figures include large transit flows through special-purpose entities. Exchange rates are World Bank annual averages.
Romania is an EU member outside the euro area and outside the OECD, so every currency figure is in lei and the OECD's 15 to 64 employment and participation rates are not available. Two inflation measures are shown from January 2002: the National Institute of Statistics' national CPI, the headline and the measure the National Bank of Romania targets, from the IMF's copy of INS's index (about a month behind INS), and the EU's harmonised index (HICP), which differs more than in most EU countries. Inflation from 1997 to 2001 ran between 40% and over 150% a year and is not shown. GDP is shown from 2000 and unemployment from January 2009, where Eurostat's earlier quarterly and monthly series for Romania swing or step in ways that look like changes of method. The interest rate is the NBR's monetary policy rate from the Bank for International Settlements, from August 2005, when the NBR adopted inflation targeting. Goods trade is Eurostat's, converted from euros to lei at each month's European Central Bank reference rate. The leu was redenominated in July 2005 (10,000 old lei became 1 leu); figures are in new lei.
The riyal has been pegged at 3.75 to the US dollar since 1986, so the Saudi Central Bank moves its repo rate with the US Federal Reserve and the exchange rate barely moves; the repo rate is shown from January 2000, the start of BIS's official series. GDP and inflation are the General Authority for Statistics' figures as the OECD republishes them; year-on-year GDP growth is computed from the unadjusted series, as GASTAT headlines it. Goods trade is GASTAT's data as the IMF republishes it, in riyals. Unemployment is the International Labour Organization's modelled annual estimate for all residents, until GASTAT's quarterly survey can be read directly; the rate for Saudi nationals alone is considerably higher. Debt and the budget balance are the IMF's World Economic Outlook figures, actual years only. There is no 10-year government bond yield or business confidence index.
Slovakia joined the euro on 1 January 2009 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, Národná banka Slovenska set its own rate, which is not shown. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the Statistical Office's national CPI; its jumps in July 1999, February 2000 and January 2003 came from VAT and regulated price increases. Unemployment is Eurostat's rate, which differs from the labour office's registered rate often quoted in Slovak news. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Croatia joined the euro on 1 January 2023 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, the Croatian National Bank ran its own policy, which is not shown. Figures from before 2023 are converted from kuna at the fixed rate of 7.53450 kuna per euro. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the national CPI the Croatian Bureau of Statistics headlines. Croatia is not an OECD member, so the 10-year yield is Eurostat's convergence-criterion yield and the employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009. Unemployment is Eurostat's rate, which differs from the registered rate often quoted in Croatian news. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Bulgaria joined the euro on 1 January 2026 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, the lev was tied to the euro under a currency board run by the Bulgarian National Bank, which is not shown. Figures from before 2026 are converted from lev at the fixed rate of 1.95583 lev per euro. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the national CPI the National Statistical Institute headlines; it is shown from 2000, because the late 1990s carry the tail of the 1997 hyperinflation. Bulgaria is not an OECD member, so the 10-year yield is Eurostat's convergence-criterion yield (the benchmark bond trades thinly, so the published rate often stays unchanged for months) and the employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009. Unemployment is Eurostat's rate, which differs from the Employment Agency's registered rate often quoted in Bulgarian news. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Lithuania joined the euro on 1 January 2015 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, the litas ran on a currency board run by the Bank of Lithuania, tied to the US dollar until 2002 and to the euro after, which is not shown. Figures from before 2015 are converted from litas at the fixed rate of 3.45280 litas per euro. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the national CPI the State Data Agency also publishes. Lithuania is an OECD member, but the OECD's copies of its 10-year yield and labour-force rates are not available through FRED, so the 10-year yield is Eurostat's convergence-criterion yield and the employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009. The benchmark bond trades thinly, so the published yield often stays unchanged for months, and moves in steps when it does change. Unemployment is Eurostat's rate, which differs from the Employment Service's registered rate. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Estonia joined the euro on 1 January 2011 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, the kroon ran on a currency board run by Eesti Pank, tied to the Deutsche Mark from 1992 and to the euro from 1999, which is not shown. Figures from before 2011 are converted from kroons at the fixed rate of 15.6466 kroons per euro. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the national CPI Statistics Estonia also publishes. Estonia had no 10-year government bond until its first issue in June 2020, so the 10-year yield, Eurostat's convergence-criterion yield, starts then. The employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009, rather than the OECD's copies, which differ from them by up to a percentage point. The OECD's business confidence indicator for Estonia stopped updating in April 2026, so it is not shown. Unemployment is Eurostat's rate, which differs from the Unemployment Insurance Fund's registered rate. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Latvia joined the euro on 1 January 2014 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, Latvijas Banka kept the lats pegged, to the IMF's SDR basket from 1994 and to the euro from 2005, which is not shown. Figures from before 2014 are converted from lats at the fixed rate of 0.702804 lats per euro. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the national CPI the Central Statistical Bureau also publishes. The 10-year yield is Eurostat's convergence-criterion yield, because the OECD's copy is not available through FRED. The employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009. Unemployment is Eurostat's rate, which differs from the State Employment Agency's registered rate. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Slovenia joined the euro on 1 January 2007 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, Banka Slovenije set its own rates for the tolar, which are not shown. Figures from before 2007 are converted from tolars at the fixed rate of 239.640 tolars per euro. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the national CPI the Statistical Office also publishes. The 10-year yield is the OECD's series, which matches Eurostat's convergence-criterion yield. The employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009. Unemployment is Eurostat's rate, which differs from the Employment Service's registered rate. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Luxembourg has used the euro since its launch on 1 January 1999 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate; before it, the Luxembourg franc was at par with the Belgian franc. Figures from before 1999 are converted from francs at the fixed rate of 40.3399 francs per euro. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the national CPI STATEC also publishes. The 10-year yield is the OECD's series, which matches Eurostat's convergence-criterion yield. The employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009. They cover residents only, while many people working in Luxembourg commute from neighbouring countries. Unemployment is Eurostat's rate, which differs from ADEM's registered rate. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
Cyprus joined the euro on 1 January 2008 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, the Central Bank of Cyprus set its own rates, which are not shown. Figures from before 2008 are converted from Cyprus pounds at the fixed rate of 0.585274 pounds per euro. Inflation is the EU's harmonised index (HICP), which can differ by a few tenths from the national CPI CYSTAT also publishes. Cyprus is not an OECD member, so the 10-year yield is Eurostat's convergence-criterion yield and there is no OECD business confidence indicator. The employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009. Unemployment is Eurostat's rate, which differs from the Public Employment Service's registered rate. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's. The deficits of 2012 and 2014 include the cost of supporting the banks during the 2012-13 crisis.
Malta joined the euro on 1 January 2008 and has had no interest rate of its own since, so the rate shown is the European Central Bank's deposit facility rate, from that date; before it, the Central Bank of Malta set its own rates, with the lira pegged to a basket of currencies and, from May 2005, to the euro, which are not shown. Figures from before 2008 are converted from Maltese lira at the fixed rate of 0.429300 lira per euro. Real GDP, and so quarterly growth, starts in 2000, because Eurostat's inflation-adjusted series for Malta begins there. Inflation is the EU's harmonised index (HICP), which can differ from the national retail price index the NSO also publishes. Malta is not an OECD member, so the 10-year yield is Eurostat's convergence-criterion yield and there is no OECD business confidence indicator. The employment and participation rates are Eurostat's seasonally adjusted Labour Force Survey figures, from 2009. Unemployment is Eurostat's rate, which differs from Jobsplus's registered rate. Debt and the budget balance are Eurostat's excessive deficit procedure figures, a year more recent than the IMF's.
The Philippines' figures come from the Philippine Statistics Authority's open database wherever it publishes them: GDP, inflation and the labour force survey. Quarter-on-quarter GDP growth is computed from PSA's seasonally adjusted real GDP, and year-on-year growth from its unadjusted figures, PSA's headline. Inflation is the consumer price index for all income households, 2018=100: PSA's published rate from 2019, and the change in its 2018-based index for 1995 to 2018, a method that reproduces the published rate exactly where both exist. The labour force survey became monthly in 2021, so the labour series start then; they are not seasonally adjusted, and the employment-to-population ratio is computed from PSA's participation and employment rates. The policy rate is the Bank for International Settlements' copy of Bangko Sentral ng Pilipinas' reverse repurchase rate from 2002; earlier values in that series are money-market rates, not a policy rate. PSA reports goods trade in US dollars, so each month is converted to pesos at the BIS monthly average rate. The government balance is the IMF's general government measure.
Taiwan's page is deliberately partial. Its statistics office does not allow automated downloads, and Taiwan is not covered by the World Bank or the Bank for International Settlements, so GDP, real growth, inflation and unemployment are the IMF World Economic Outlook's annual figures, actual years only, with no quarterly or monthly versions. The policy rate is the discount rate from the central bank's own open-data file, held from each decision to the next. Goods trade is Taiwan's customs data as the IMF republishes it in US dollars, converted to New Taiwan dollars at each month's average of the Federal Reserve's daily rate. The current account, foreign direct investment, the bond yield and business confidence are not shown, because no free source covers them.
Vietnam's page is also partial. Its statistics office and central bank publish web pages rather than downloadable data, so inflation and goods trade come from the IMF's republication of the official monthly figures, and GDP, growth and unemployment are annual figures from the World Bank; unemployment is the International Labour Organization's modelled estimate. Trade is converted from US dollars to dong at the Bank for International Settlements' monthly average rate; that series has no August 2025 rate, so that month uses the average of July and September. There is no policy rate, bond yield or business confidence index on the page, because no free downloadable source carries them.
Nigeria's page is partial. The National Bureau of Statistics and the Central Bank of Nigeria publish web pages rather than downloadable data, so inflation and the policy rate come from the IMF's republication of the official monthly figures, and GDP, growth and unemployment are annual figures from the World Bank; unemployment is the International Labour Organization's modelled estimate. The National Bureau of Statistics rebased its consumer price index to 2024 prices in January 2025, so inflation rates either side of that month are not strictly comparable. The IMF publishes the Monetary Policy Rate some months after each decision, and the page shows the month of the latest figure. There is no monthly goods trade series, bond yield or business confidence index on the page, because no free downloadable source carries them; trading partners, where shown, are UN Comtrade's annual figures.
Egypt's page is partial. CAPMAS, Egypt's statistics agency, and the Central Bank of Egypt do not offer downloadable data that automated requests can reach, so inflation and goods trade come from the IMF's republication of the official monthly figures, and GDP, growth and unemployment are annual figures from the World Bank, for fiscal years running July to June; unemployment is the International Labour Organization's modelled estimate. Trade is converted from US dollars to Egyptian pounds at the Bank for International Settlements' monthly average rate. The IMF's figures for February 2016 have exports and imports the wrong way round, and the page shows them corrected. There is no policy rate on the page: the IMF's most recent Central Bank of Egypt rate is more than a year old, too old to show as current. There is no bond yield or business confidence index either, because no free downloadable source carries them.
Peru's page draws mainly on the Central Reserve Bank of Peru's own statistics service. Inflation is the consumer price index for Metropolitan Lima, the headline measure the central bank targets, not the national index the IMF republishes, which differed from it by up to half a point in 2026. Real GDP is quarterly, in soles at 2007 prices and not seasonally adjusted, so quarters are compared with the same quarter a year earlier; GDP at current prices is annual, from the World Bank, from 1991, when the sol replaced the inti. Goods trade is the central bank's monthly trade balance in US dollars, converted to soles at the Bank for International Settlements' monthly average rate. The policy rate is the reference rate from September 2003, when it became the bank's instrument. Unemployment is the International Labour Organization's modelled annual estimate.
Hong Kong's GDP is quarterly, from the Census and Statistics Department's own data service: current prices, and chained (2024) dollars for real GDP, with growth on the previous quarter seasonally adjusted. Inflation is the Composite Consumer Price Index as the IMF republishes it; headline rates swing as one-off government relief measures, such as rates or rent waivers, start and end. The policy rate is the Hong Kong Monetary Authority's base rate from October 1998, which follows the US federal funds rate under the link to the US dollar. Goods trade, most of it re-exports, is the Census and Statistics Department's own monthly series in Hong Kong dollars: total exports (domestic exports and re-exports) and imports, as published. Because the Hong Kong dollar has been linked to the US dollar since 1983, Dollarise changes the level of a series but not its shape. Unemployment is the International Labour Organization's modelled annual estimate.
Iceland's GDP, inflation and unemployment are Statistics Iceland's figures as the OECD republishes them. Inflation is the national consumer price index, which includes owner-occupied housing, not the EU's harmonised index. The policy rate is the Central Bank of Iceland's key rate from June 2014, when it took its current form. Goods trade is the IMF's monthly series in US dollars, converted to krónur at the Bank for International Settlements' monthly average rate; three months in which the IMF's import figure repeats its export figure exactly (March and April 2013, January 2026) are left out rather than shown as a zero balance. The exchange rate and Dollarise start in 1981, when the new króna replaced the old at 100 to 1. There is no 10-year bond yield, because the OECD's series ended in August 2022, and no business confidence index.
Pakistan's page is partial. The State Bank of Pakistan's data service refuses automated requests and the Pakistan Bureau of Statistics publishes no downloadable price data, so inflation and goods trade come from the IMF's republication of the official monthly figures, and GDP and growth are annual World Bank figures for fiscal years running July to June. Goods trade is in rupees as Pakistan reports it, from July 2022: the IMF's series has no figures for November 2021 to June 2022. Unemployment is the International Labour Organization's modelled estimate from 2012; before then the modelled series reads under 1%, far below Pakistan's own surveys. There is no policy rate on the page, because no free downloadable source republishes it, and no bond yield or business confidence index.
Click the ⓘ next to any individual metric, on any country page, Compare, or your own Dashboard, for that number's exact source and any conversion caveats specific to it. If something looks wrong or you think a better source exists for a metric currently on a fallback, get in touch, most fixes to date have come from exactly that kind of report.