Central Bank Claims on the Domestic Real Sector as % of GDP | 2021 World Map

The label “central bank assets to GDP” can easily be mistaken for the size of a central bank’s entire balance sheet relative to the economy. World Bank series GFDD.DI.06 is narrower. Its metadata defines the measure as central-bank claims on the domestic real nonfinancial sector as a share of GDP. Restricting the comparison to observations actually dated 2021 leaves 136 countries and areas. Their median is 5.00%, while the simple unweighted mean is 9.25%.

The distribution is strongly right-skewed. Twenty-six of the 136 observations are below 1%, 42 are from 1% to under 5%, and 40 are from 5% to under 15%. Only 19 fall from 15% to under 30%, and nine are at or above 30%. Japan is the highest 2021 observation at 89.06%, followed by Libya at 54.08% and Italy at 43.03%. A high value does not mean that a central bank is “better,” larger in every sense, or running a stronger monetary policy. It means that the particular domestic claims captured by this indicator are large relative to GDP.

“Central bank assets” here does not mean the entire balance sheet

The World Bank Global Financial Development metadata names GFDD.DI.06 “Central bank assets to GDP (%)”, but the definition narrows what is counted. It refers to claims on the domestic real nonfinancial sector held by the central bank. It should therefore not be read as a ratio that automatically includes every foreign reserve asset, gold holding, foreign security, liquidity facility, or other balance-sheet position.

The underlying monetary statistics come from the IMF’s International Financial Statistics and are published in indicator form through the World Bank Global Financial Development Database. The numerator is the relevant stock of central-bank claims and the denominator is GDP. That denominator matters: the same absolute amount of claims would produce a larger ratio in a smaller economy and a smaller ratio in a much larger economy. The map is therefore a scale-adjusted comparison, not a ranking of absolute central-bank assets.

This distinction is especially important for readers looking for quantitative easing or the latest size of Federal Reserve, European Central Bank, Bank of Japan, or other central-bank balance sheets. Those questions require full balance-sheet data from the relevant institutions or a separate harmonized series. GFDD.DI.06 alone does not measure every asset acquired through asset-purchase programs and should not be presented as a current total-balance-sheet ranking.

Why the comparison uses 136 observations from 2021 instead of all 177 latest values

The verified country extract contains 177 latest non-empty observations, but those observations are not synchronized in time. A total of 136 are dated 2021, while 41 retain latest values from 1997 through 2020. Combining every latest value in one ranking would create a mixed-period comparison in which some apparent country differences could reflect different observation years rather than only financial structure.

For that reason, every ranking, mean, median, distribution count, and colored map in this article uses only rows whose year equals 2021. The other 41 observations remain relevant to overall dataset coverage but are not treated as if they described 2021. Mexico, for example, has a much older retained latest observation and is therefore not placed beside 2021 countries in the ranking. A gray country on the map should never be interpreted as a zero.

The 2021 median is 5.00%, while the unweighted mean is 9.25%

Across the 136 same-year observations, the median is 5.00% and the simple unweighted mean is 9.25%. The mean sits about 4.25 percentage points above the median because a relatively small upper tail pulls it higher. Japan at 89.06%, Libya at 54.08%, and Italy at 43.03% have a large effect on a simple average even though most countries are far below those levels.

The grouped distribution makes this clearer. Twenty-six observations are below 1%, 42 are from 1% to under 5%, 40 are from 5% to under 15%, 19 are from 15% to under 30%, and nine are at least 30%. In other words, 108 of the 136 observations—about 79%—are below 15%. The “world mean” here is not GDP-weighted, population-weighted, or intended to represent the balance sheet of a hypothetical global central bank; it is simply the arithmetic average of the country rows.

Japan, Libya and Italy are the highest 2021 observations

Japan records 89.06% in 2021, the highest value in the same-year set. Libya follows at 54.08%, then Italy at 43.03%, Spain at 37.46%, Bolivia at 35.34%, Portugal at 34.35%, Trinidad and Tobago at 32.18%, Algeria at 31.21%, France at 30.44%, and Slovenia at 27.57%. Japan is roughly 35 percentage points above the second-highest observation, making it a pronounced upper-end outlier in this particular series.

The top group spans East Asia, Europe, North Africa, South America, and the Caribbean. That geographic mix argues against a simple continent-wide explanation. It also shows why the indicator should not be treated as a generic proxy for income, central-bank independence, or monetary-policy quality. The verified dataset supports a statement about relative claim levels, but it does not identify one causal mechanism behind those levels.

Twenty-six 2021 observations are below 1%

At the low end, Costa Rica is about 0.00002%, Bosnia and Herzegovina 0.00143%, Czechia 0.00640%, Norway 0.01227%, and Cambodia 0.02185%. Uzbekistan is 0.04118%, Bulgaria 0.05288%, Cyprus 0.06827%, Paraguay 0.07034%, and Botswana 0.09926%. These values are small relative to GDP, but they are not statements that the central banks themselves have virtually no assets.

Norway illustrates the interpretation problem well. A very low GFDD.DI.06 ratio does not erase the existence of central-bank foreign assets, reserves, or other financial arrangements. It only means that the claims counted in this series are small relative to GDP. Reading the number as “total central-bank assets” would therefore create a misleading conclusion for countries whose central-bank balance sheets are concentrated in items outside this indicator’s numerator.

Large economies occupy very different parts of the same scale

Among selected large economies, Japan is at 89.06%, France 30.44%, the United States 24.60%, Brazil 23.39%, Germany 22.23%, Canada 18.78%, and Australia 16.28%. Lower on the same scale are Indonesia at 8.12%, India 6.77%, Korea at 1.67%, China 1.33%, the United Kingdom 0.48%, and the Russian Federation 0.19%. Every value in this comparison is from 2021.

The contrast is especially sharp in East Asia: Japan sits at the top of the world distribution, while Korea and China are below 2%. Western Europe also contains both high and very low values. France, Italy, Spain, Portugal, and Germany are relatively high, while the United Kingdom and Norway are very low. North America has the United States and Canada above 15%. These differences among neighboring or similarly high-income economies are another reason not to use the ratio as a simple development ranking.

Geographic patterns are descriptive, not causal explanations

Several high-value countries form a visible band around parts of Southern Europe and North Africa: Italy, Spain, Portugal, and France are around or above 30%, while Libya and Algeria also exceed 30%. Yet the broader region contains major exceptions. The United Kingdom, Norway, Czechia, Bulgaria, and Cyprus are all low. A continental average would hide more than it explains.

South America is similarly mixed. Bolivia is at 35.34% and Brazil at 23.39%, while Paraguay is only 0.07%. In East Asia, Japan’s 89.06% is dramatically different from Korea’s 1.67% and China’s 1.33%. Such local discontinuities suggest that researchers need country-specific institutional and balance-sheet context before trying to explain the values. The map identifies where large differences exist; it does not prove why they exist.

The 2021 reference year also deserves caution. It falls during the period of pandemic-related economic and policy disruption, but GFDD.DI.06 is not a direct measure of pandemic support or emergency monetary-policy size. A causal pandemic comparison would require a time series and additional indicators. Similarly, studying quantitative easing would require securities holdings and full balance-sheet measures rather than this narrower claim series.

A higher or lower ratio is not automatically better

A high value means that the claims included in GFDD.DI.06 are large relative to GDP. The ratio does not by itself establish stronger financial stability, better inflation performance, faster economic growth, healthier public finances, or more effective monetary policy. Legal mandates, financial-system structure, crisis arrangements, claim composition, and GDP all differ across countries.

A low value likewise does not mean that a central bank is unimportant or inactive. Policy rates, reserve requirements, foreign-exchange operations, securities holdings, lending facilities, settlement systems, and supervisory responsibilities extend well beyond this one numerator. The safest use of the map is therefore narrow and literal: it compares central-bank claims on the domestic real nonfinancial sector relative to GDP in 2021.

Data source and calculation method

The statistical source is the World Bank Global Financial Development Database series GFDD.DI.06, Central bank assets to GDP (%). World Bank metadata describes the measure as central-bank claims on the domestic real nonfinancial sector as a share of GDP and identifies the IMF International Financial Statistics as the underlying source. The unit is percent. The country extract contains 177 latest non-empty values, but the observation years are mixed.

To preserve same-year comparability, this article calculates all rankings and summary statistics from the 136 rows dated 2021. The median of 5.00%, unweighted mean of 9.25%, range counts, and country tables all come from that identical subset. Missing values are not converted to zero, and older observations are not backfilled into 2021.

The map joins ISO-3 country codes to a low-resolution Natural Earth boundary layer. The statistical sample contains 136 values, while 124 are directly represented by polygons at this scale after standard code matching. Small islands and separately reported areas can have valid statistics without a distinct visible shape. Gray therefore means “not mapped in the 2021 comparison,” not “zero central-bank claims.”

Frequently Asked Questions

Related Articles

These Green Map articles help separate the meaning of a GDP-denominated financial ratio from economic size, real growth, and other ratios that also use GDP as a denominator.

• Global GDP Growth Map – Country Patterns in 2025
• Global GDP per Capita Map – Country Levels in 2025
• Natural Resource Rents as % of GDP – 2021 World Map