InsightDeck / Themes / What did digital banks bring Brazil's unbanked?

Theme · Brazil · Financial inclusion and household debt

What did digital banks bring Brazil's unbanked?Accounts, debt and interest rates, 2011–2026

Published 2026-11-18 Sources Central Bank of Brazil, World Bank, IBGE and others (data retrieved 2026-11-17) Updated 2026-11-18

Key takeaways

  1. 1 The share of Brazilian adults with an account at a bank or other financial institution rose from 55.9% in 2011 to 85.6% in 2024 — from about 82.7 million to about 145.7 million people, an increase of about 63.0 million. Among the lowest-income 40%, it went from 37.5% to 76.4%, roughly double.
  2. 2 At the same time, household debt reached 50.1% of annual disposable income and debt service (principal and interest) 29.0% of disposable income, both the highest since the series began in 2005 (the debt figure includes mortgages; excluding them it is 31.0%). The credit card revolving rate is 451.7% a year (a monthly rate of about 15.3% compounded over a year; a balance can stay on revolving credit for only one month).
  3. 3 Researchers disagree on whether this amounts to "financial inclusion" (people who could not use bank accounts or loans gaining access to them) or to drawing low-income households into debt. This page sets out the numbers and arguments on both sides, with sources.
  4. 4 Public statistics cannot separate how much of the rise in accounts and debt is due to digital banks. This page lays out what happened during the years digital banks spread.

Figures come from public data of the Central Bank of Brazil, the World Bank and the Brazilian Institute of Geography and Statistics (IBGE); each chart lists its source. InsightDeck does not recommend any particular view. Disclosure: the operator holds shares of Nu Holdings (NU). This page does not assess any company's results or share price.

Adults with an account (2024) 85.6% About 145.7 million (estimate) · 55.9% in 2011 World Bank Findex
Lowest-income 40% with an account (2024) 76.4% About 52.0 million (estimate) · 37.5% in 2011 · 81.6% in 2021 World Bank Findex
Borrowed from a financial institution (2024) 45.9% About 78.2 million (estimate) · 33.8% in 2014 · Latin America avg. 27.4% World Bank Findex
Growth of household credit outstanding 5.0× R$4.6 trillion (about US$891 billion) · end-2011 to June 2026 · consumption 3.2× (both nominal) · credit 2.2× in real terms Central Bank of Brazil, IBGE
Household debt (% of annual disposable income) 50.1% August 2026 · 31.0% excluding mortgages Central Bank of Brazil
Debt service (% of disposable income) 29.0% August 2026 · seasonally adjusted Central Bank of Brazil
Credit card revolving rate (annual) 451.7% September 2026 · about 15.3% a month · revolving balances are about 2% of household credit Central Bank of Brazil
Household arrears rate (free-rate credit) 8.1% September 2026 · more than 90 days overdue · inflated by a 2025 accounting change Central Bank of Brazil

People counts are InsightDeck estimates (share × population aged 15 and over). US dollar amounts use the quarter-end rate (US Federal Reserve H.10; end-June 2026: US$1 = R$5.1665). The 451.7% revolving rate is a monthly rate of about 15.3% compounded over a year. In practice a balance cannot stay on revolving credit beyond the next due date (about 30 days), and total interest is capped (section 03).

01

More people have accounts

The share of Brazilian adults with an account at a bank or other financial institution rose from 55.9% in 2011 to 85.6% in 2024, well above the Latin America & Caribbean (excluding high income) average of 66.4%. That is from about 82.7 million to about 145.7 million people, an increase of about 63.0 million (estimate). These were the years when digital (app-only) banks, where an account can be opened entirely on a phone (Nubank, Inter, C6 Bank and others), spread, the instant payment system Pix launched in November 2020, and COVID-19 benefits were paid into digital accounts.

Brazilian adults with an account (estimate)

Aged 15 and over · 2011 and 2024

InsightDeck chart
All adults201182.7M2024145.7M+63.0MLowest-income 40%201122.2M202452.0M+29.8M

Source: InsightDeck calculation from World Bank Global Findex 2025 (a survey of people aged 15 and over; the latest round was conducted in 2024) shares and World Bank World Development Indicators (WDI) population. People counts are InsightDeck estimates: the share multiplied by the World Bank population aged 15 and over (total population minus ages 0–14), rounded to the nearest 100,000. The lowest-income 40% is taken as 40% of the population aged 15 and over.

Adults with an account at a financial institution

Aged 15 and over · five survey rounds, 2011–2024 (Mexico's fourth round was in 2022)

InsightDeck chart
0%25%50%75%100%20112014201720212024Brazil85.6%・about 145.7 millionLatAm avg.66.4%Mexico52.4%Colombia43.4%
BrazilLatin America & Caribbean avg.MexicoColombia

Source: World Bank Global Findex 2025 (a survey of people aged 15 and over; the latest round was conducted in 2024). Indicator: account at a bank or other financial institution (WB_FINDEX_FIACCOUNT_T_D), which includes digital banks such as Nubank. The Latin America average is the World Bank group "Latin America & Caribbean (excluding high income)". Colombia's large drop in 2024 may reflect a change in survey method; we have not been able to confirm the World Bank's explanation.

Show the numbers
Country / region (%)20112014201720212024
Brazil55.968.170.083.685.6
Latin America & Caribbean average (excl. high income)39.551.353.565.066.4
Mexico (fourth round in 2022: 46.2)27.438.734.6—52.4
Colombia30.438.444.955.943.4
Japan96.496.698.298.598.5
World average50.660.967.971.575.3

Brazilian adults with an account, by income group

Share of people aged 15 and over (%) · five survey rounds, 2011–2024 · lowest-income 40% and highest-income 60% within the country

InsightDeck chart
0%25%50%75%100%20112014201720212024Highest 60%91.7%All adults85.6%・about 145.7 millionLowest 40%76.4%・about 52.0 million
Lowest-income 40%All adultsHighest-income 60%

Source: World Bank Global Findex 2025 (a survey of people aged 15 and over; the latest round was conducted in 2024). Income groups are based on household income rank within each country. People counts are InsightDeck estimates: the share multiplied by the World Bank population aged 15 and over (total population minus ages 0–14), rounded to the nearest 100,000. The lowest-income 40% is taken as 40% of the population aged 15 and over. As this is a survey, its estimates are subject to error; these numbers alone cannot tell whether the drop for the lowest-income group from 2021 to 2024 (81.6% → 76.4%) is a real change.

02

More people borrow, too

More Brazilians also hold credit cards and have borrowed from financial institutions. In 2024, 45.9% of adults (about 78.2 million, estimate) had borrowed from a financial institution, about 1.7 times the Latin America & Caribbean (excluding high income) average of 27.4%. But the gap between income groups is wider than for accounts.

Card holders and people who borrowed from a financial institution (2024)

Share of people aged 15 and over. Japan: 2021 data. (InsightDeck is based in Japan, so we use Japan as a reference point.)

InsightDeck chart
Has a credit cardBrazil43.5%Latin America avg.25.2%Mexico14.2%Colombia8.7%Japan (2021)69.7%Borrowed from a financial institution (incl. card use)Brazil45.9%Latin America avg.27.4%Mexico14.8%Colombia12.4%Japan (2021)61.2%

Source: World Bank Global Findex 2025 (a survey of people aged 15 and over; the latest round was conducted in 2024). "Borrowed from a financial institution" means borrowing from a bank or other financial institution in the 12 months before the survey; by the World Bank's definition it includes using a credit card (WB_FINDEX_FIN22A). "Has a credit card" is WB_FINDEX_FIN10. For Japan, borrowing has no 2024 value and the 2024 card value is recorded as 0, so both use 2021 values. The Latin America average is the World Bank group "Latin America & Caribbean (excluding high income)".

How to read this chart

  • "Has a credit card" and "Borrowed from a financial institution" are both shares of people aged 15 and over, on the same scale (0–80%), so bar lengths can be compared directly.
  • By the World Bank's definition, "borrowed from a financial institution" includes credit card use: buying something with a card and paying the next month counts as borrowing. That is the main reason Japan's value is high. The chart shows how widely cards and loans are available and used, not how heavy the debt is. The burden is shown in section 04 (The burden on households) and section 06 (Loans in arrears).
  • The denominator in every country is everyone aged 15 and over. Japan's 69.7% card ownership includes people aged 15–17 and 70 and over, so it is lower than domestic surveys of people aged 20–69 (in the mid-80s per cent in JCB's annual credit card survey).

Brazilian adults who borrowed from a financial institution, by income group

Share of people aged 15 and over (%) · four survey rounds, 2014–2024

InsightDeck chart
0%25%50%75%2014201720212024Highest 60%56.2%All adults45.9%・about 78.2 millionLowest 40%30.4%・about 20.7 million
Lowest-income 40%All adultsHighest-income 60%

Source: World Bank Global Findex 2025 (a survey of people aged 15 and over; the latest round was conducted in 2024) (WB_FINDEX_FIN22A). In 2024 the gap between the lowest-income 40% and the highest-income 60% was 15.3 points for accounts, 25.6 points for card ownership and 25.8 points for borrowing (InsightDeck calculated these from unrounded values, so they can differ by 0.1 from subtracting the table values). People counts are InsightDeck estimates: the share multiplied by the World Bank population aged 15 and over (total population minus ages 0–14), rounded to the nearest 100,000. The lowest-income 40% is taken as 40% of the population aged 15 and over.

Show the numbers
Brazil (%)20112014201720212024
Has a credit card
Lowest-income 40%15.020.515.627.628.1
All adults29.232.027.040.443.5
Highest-income 60%38.739.734.649.053.8
Borrowed from a financial institution
Lowest-income 40%—20.915.927.630.4
All adults—33.826.340.745.9
Highest-income 60%—42.433.249.556.2
03

The cost of borrowing

Interest rates on household loans in Brazil are high: even the average is far above the legal ceilings on loan interest in Japan (15–20% a year). A particularly high rate is that on revolving credit card balances (rotativo), at 451.7% a year. In Brazil, if a cardholder pays less than the full bill but at least the minimum, the unpaid balance automatically rolls into revolving credit for one month — similar to carrying a card balance in the US or UK, but at far higher rates and for one month only. This rate should not be taken at face value (see below). "Free-rate credit" (recursos livres; the central bank's "non-earmarked credit") means loans whose rates are freely agreed between lender and borrower: cards, personal loans, payroll-deducted loans and car loans. It excludes earmarked credit such as mortgages, rural loans and BNDES (national development bank) loans, whose funding or rates are set by policy.

Interest rates on household credit (September 2026)

Annual rate · average for new loans made in the month

InsightDeck chart
Annual rateCard revolving (rotativo)451.7%Card instalments (interest-bearing)193.0%Personal loans (not payroll-deducted)110.9%All card credit100.3%All household free-rate credit62.4%Policy rate (Selic, set by the central bank)13.8%

Source: Central Bank of Brazil time series (SGS). Revolving 22022; card instalments 22023 (the average for instalment credit on which the card issuer charges interest, such as instalment payment of the bill balance; it excludes interest-free store instalments); personal loans (not payroll-deducted) 20742; all card credit 22024; all household free-rate credit 20740; policy rate (Selic, the realised rate for the month, annualised) 4189. Free-rate credit: see the text above.

On paper, 451.7% a year would multiply a debt about 5.5-fold in a year; current rules prevent that. It is a monthly rate of about 15.3% compounded over a year, and there are rules on how long a balance can stay on revolving credit and on total interest. Interest, fees and other charges are capped at the original amount owed, so a borrower repays at most twice the original amount (below).

How interest applies when a Brazilian card bill is not paid in full

Time runs left to right. Rates are September 2026 averages, converted from annual to monthly rates

InsightDeck chart

Paid at least the minimum, left a balance (it rolls into revolving credit automatically)

0%No interest
15.3%/moRevolving 451.7% a year
9.4%/moInstalments (with interest) 193.0% a year

Chose instalments by the due date

0%No interest
9.4%/moInstalments (with interest) 193.0% a year. No revolving period
  • The revolving rate applies to the same balance only for one month, from the due date to the next due date (since April 2017). If the balance is paid in full on the next due date, it ends there; if a balance remains, it moves to instalments on terms better than revolving (usually automatically under the card contract). New purchases go through the same cycle again.
  • The balance rolls into revolving credit automatically when the cardholder pays at least the minimum and chooses nothing else. Paying less than the minimum counts as arrears: on top of the revolving rate, a 2% fine and late interest of 1% a month apply. A bill paid in full carries no interest.
  • Total interest, late fines and interest, and fees are capped at the original amount owed (since January 2024). If revolving credit moves to instalments, the count runs from the start of the revolving period.

Source: rules from National Monetary Council (CMN) Resolution 4.549 (2017, including the 2023 amendment by Resolution 5.112) and Law 14.690/2023 (2024); late fines and interest from the CMN rule of April 2018; rates from Central Bank of Brazil SGS 22022 and 22023. Band lengths are indicative; actual rates and repayment periods vary by card issuer and borrower.

Reading the 451.7%

  • Revolving balances were about R$97.8 billion (about US$18.9 billion) in September 2026, about 2% of all household credit outstanding.
  • Borrowing R$100 at this rate for a full year would, on paper, grow to about R$552. But two rules apply.
  • Revolving only until the next due date: since April 2017, the same balance can stay on revolving credit only until the next due date (about 30 days); what remains can be financed only through instalments or similar credit on terms better than revolving (CMN Resolution 4.549).
  • A cap on total interest: since January 2024, total interest, late fines and interest, and fees on revolving credit and on instalment payment of the bill balance may not exceed the original amount owed (Law 14.690/2023, article 28, implemented by CMN Resolution 5.112). On a R$100 debt, the most a borrower repays is R$200 (excluding IOF tax). It is not a cap on the annual rate, so the annual rate itself has stayed between 412% and 460% since 2024.
  • For reference, in Japan (where InsightDeck is based) interest on cash loans is capped by law (Interest Rate Restriction Act) at 15–20% a year depending on the amount. Fees on revolving payment for card purchases are not subject to that cap, but are usually around 15% a year. The Brazilian figures are average rates actually charged; Japan's are legal ceilings, so the two are not like-for-like.

Card revolving and instalment rates, and the policy rate

Annual rate · March 2011 – September 2026 (monthly)

InsightDeck chart
0%100%200%300%400%500%2012201420162018202020222024202630-day ruleInterest capRevolving451.7%Instalments193.0%Policy rate13.8%
Card revolvingCard instalments (interest-bearing)Policy rate (Selic)

Source: Central Bank of Brazil time series (SGS) (22022, 22023, 4189). After the April 2017 rule took effect, the revolving rate fell to 359% in May 2017 and was back at 428% by August. In April 2018 the CMN set a rule under which the revolving rate for people who paid less than the minimum is the regular revolving rate plus a fine and late interest, and the revolving rate fell to 280% by July 2018. The revolving rate series averages people who paid at least the minimum and those who did not. Rates have since risen again.

04

The burden on households

Debt burdens are at record highs relative to income. Household debt is 50.1% of annual disposable income and debt service 29.0% of disposable income, both the highest since the series began in 2005. But the debt ratio has been roughly flat around 50% since 2022, and excluding mortgages it is 31.0%, slightly below its October 2022 peak (31.6%).

Household debt and debt service as a share of income

January 2005 – August 2026 (monthly) · income is household disposable income after taxes and social contributions

InsightDeck chart
0%10%20%30%40%50%60%200620102014201820222026Debt50.1%Excl. mortgages31.0%Debt service29.0%
Debt (% of income)Debt excluding mortgages (dashed)Monthly debt service (% of income)

Source: Central Bank of Brazil time series (SGS). Debt is household debt to financial institutions divided by household disposable income over the last 12 months (29037; 29038 excluding mortgages). Debt service is principal and interest payments as a share of disposable income, seasonally adjusted (29034). The debt service series starts in March 2005. The peak of the debt ratio excluding mortgages is 31.6% in October 2022.

How to read it

  • A debt service ratio of 29% means roughly 30% of disposable (take-home) income goes to monthly debt payments (an average over all households, including those that do not borrow; for borrowing households alone the ratio is higher).
  • When the policy rate is high, new loans and refinancing also carry higher rates, which can be one factor pushing up debt service. The policy rate (the realised monthly rate, annualised) was 13.8% in September 2026.
05

Credit grew faster than consumption

From end-2011 to June 2026, household credit outstanding grew from about R$922 billion to about R$4.6 trillion (about US$891 billion), roughly fivefold (5.0-fold), while household consumption (annual) grew about 3.2-fold. Prices more than doubled over the same period (about 2.2 times their end-2011 level). In real terms, credit grew about 2.2-fold and consumption about 1.4-fold. Credit outstanding has grown faster than consumption.

Growth since end-2011 (end-2011 = 1; as of June 2026)

Credit and consumption are nominal amounts; prices (IPCA) are an index for comparison

InsightDeck chart
Multiple of end-2011 levelHousehold credit outstanding5.0×of which free-rate credit4.0×Household consumption (annual)3.2×Prices (IPCA)2.2×

Source: credit outstanding from Central Bank of Brazil time series (SGS) 20541 (all household credit) and 20570 (of which free-rate credit); household consumption from IBGE quarterly national accounts (table 1846, household consumption expenditure, nominal), sum of the last four quarters; prices from the IPCA consumer price index (433). Multiples calculated by InsightDeck.

Household credit outstanding as a share of a year's household consumption

March 2007 – June 2026 (quarterly)

InsightDeck chart
0%20%40%60%20082012201620202024All55.3%Free-rate30.5%
All household creditof which free-rate credit

Source: as above. Credit outstanding at the end of each quarter divided by household consumption over the four quarters to that quarter (InsightDeck calculation).

How to read it

  • Credit includes mortgages and car loans, so not all of it goes to day-to-day consumption. For "of which free-rate credit" in the chart, see section 03.
  • The outstanding balance is the total not yet repaid at that point. A rising ratio to consumption means people are borrowing more, taking longer to repay, or both.
06

Loans in arrears

In household free-rate credit, the share more than 90 days overdue was 8.1% in September 2026, the highest value since the series began in 2011. In money terms that is about R$208.9 billion (about $40.3 billion), a rough figure from multiplying the arrears rate by free-rate credit outstanding. However, a 2025 accounting change inflates the figure, so it cannot be described as a record on a like-for-like basis. Figures that Nu Holdings (Nu, Nubank's parent) compiled from central bank data for its earnings presentation show the highest-income group with the lowest arrears rate at every type of lender. But at smaller institutions in August 2026 the middle-income group was higher than the mass-market group, so lower income does not always mean more arrears.

Arrears rate on household free-rate credit

More than 90 days overdue · March 2011 – September 2026 (monthly)

InsightDeck chart
0%2%4%6%8%10%20122014201620182020202220242026Accounting changeArrears8.1%

Source: Central Bank of Brazil time series (SGS) (21112). From January 2025 the accounting standard for financial institutions changed (CMN Resolution 4.966), so overdue loans stay on the books longer. The Central Bank of Brazil estimates that about 70% of the rise in the arrears rate (all financial institutions) up to June 2025 was due to this change (Monetary Policy Report, September 2025). Values from 2025 cannot be simply compared with earlier ones.

Credit card arrears by income group (from Nu's earnings slides)

90+ days overdue · % · two-month rolling average. A comparison chosen by the company; it cannot be read as ranking banks. Groups are income as a multiple of the minimum wage

InsightDeck chart
0%4%8%12%16%Mass market< 3× minimum wage9.2%14.7%13.6%Middle income3–10×7.3%7.9%14.6%High income10×+5.7%4.5%10.2%
NuLarge banks (excl. Nu)Smaller institutions

Source: Nu Holdings Q3 2026 earnings slides, p. 19 (compiled by the company from Central Bank of Brazil data, two-month rolling average). The arrears definition follows the slide ("90+ days") and comes from a different source than the chart above (more than 90 days). Because this is company material, the comparison was chosen by the company. Differences in arrears also reflect lending criteria and customer mix, so this chart cannot be read as a ranking of banks. Large banks are segments S1–S2 excluding Nu; smaller institutions are S3–S5. Same figures as the InsightDeck NU earnings page.

Show the numbers
Income groupNu Jul ’25Nu Aug ’26Large Jul ’25Large Aug ’26Smaller Jul ’25Smaller Aug ’26
Mass market (< 3× minimum wage)9.1%9.2%11.6%14.7%13.9%13.6%
Middle income (3–10×)7.6%7.3%6.4%7.9%11.4%14.6%
High income (10×+)6.2%5.7%2.8%4.5%7.1%10.2%
  • Disclosure: the operator holds shares of Nu Holdings (NU). The bank comparison plays a supporting role on this page. Each company's results are covered on InsightDeck's earnings pages.
07

What researchers say

Researchers disagree on how to assess financial inclusion. InsightDeck does not judge which side is right; it lists the main arguments with sources. Of these, only the BIS bulletin on Pix and Lavinas (2018) deal directly with Brazil.

ViewMain argumentSource
Studies measuring effects on households
Having an accountAccounts make saving and payments easier and help people cope with unexpected expenses, with larger benefits for women and lower-income people. It also finds that effects on national growth and inequality have not yet been rigorously shownDemirgüç-Kunt, Klapper & Singer (2017), World Bank
Mobile moneyEstimates that the spread of mobile money in Kenya lifted about 2% of all households (about 194,000) out of extreme poverty. The estimate has been challengedSuri & Jack (2016) Science; critique: Bateman, Duvendack & Loubere (2019)
MicrocreditSix-country evaluations sum up the effects as "modestly positive, but not transformative"Banerjee, Karlan & Zinman (2015)
Brazil's PixThe central bank's instant payment system is free for person-to-person transfers with low merchant fees, and signed up 67% of adults in about a year (registration, not usage). Says it can lower costs and advance financial inclusionBIS Bulletin No. 52 (2022)
Studies of institutions and power (political economy)
Social policy and debt in BrazilIn Brazil, pensions and benefits became collateral and an entry point for borrowing, and debt spread as a way of supporting livelihoods in place of public social protectionLavinas (2018) Development and Change
Financialisation (everyday life drawn into financial transactions)Financial inclusion can also draw low-income people into financial markets as a new source of profit, with credit scoring based on phone and usage data as a tool (Gabor & Brooks). Evidence that it leads to development or poverty reduction is weak (Mader)Gabor & Brooks (2017); Mader (2018)

How this relates to Hickel

  • We have not found any text in which Hickel discusses Brazilian household debt or digital banks directly. We include his work because it is also read in Japan as a way of looking at the spread of financial inclusion through the relationship between rich countries and the Global South (mainly developing countries).
  • Economic anthropologist Jason Hickel and co-authors argue that rich economies rely on "unequal exchange" with the Global South. They estimate that the resources and labour embodied in the South's exports exceed those in its imports by an amount that, valued at Northern prices, totals $242 trillion (in constant 2010 dollars) over 1990–2015 (Hickel et al. 2022). It is not an amount of money actually paid.
  • In The Divide (2017), he links poverty in developing countries to debt and to the economic policies (structural adjustment) required as conditions of repayment. In Less is More (2020), he criticises an economy built on perpetual growth for continually pushing up consumption through advertising and product replacement.
  • Critics argue that the estimate relies on assuming the North and South have the same productivity and export composition, that without this assumption the value is indeterminate, and that no relationship is found between higher estimated drain and slower growth (Reddy & Urbina Lazardi 2026, not yet peer reviewed).
  • The numbers on this page can be read through his argument, but that is the reader's interpretation, and the reading depends on which numbers one looks at. Numbers showing burden include credit growing faster than consumption, record debt service, and a higher arrears rate (partly inflated by the accounting change). Numbers showing wider access include account ownership among the lowest-income 40% roughly doubling (37.5% → 76.4%, 2011–2024), the share who borrowed from a financial institution rising from 33.8% to 45.9% (2014–2024), and the spread of free transfers via Pix. The rise in accounts itself can be counted on either side, depending on whether it is seen as inclusion or incorporation.

What the data cannot tell us

  • The effect of digital banks alone (the World Bank survey does not distinguish which type of bank an account is with)
  • What borrowed money was used for (daily consumption, covering living costs, refinancing)
  • How much the 2024 interest cap has eased the household burden
08

Sources and data

The numbers on this page were retrieved from each institution's public data on 17 November 2026. Series numbers for each chart are listed in its source note.

Central Bank of Brazil data (SGS) are used under ODbL and World Bank Global Findex under CC BY 4.0. If you spot an error in the numbers, please let us know via the contact on the About page. Disclosure: the operator holds shares of Nu Holdings (NU).